(Advances in Research and Theory 4) Richard J. Davidson, Gary E. Schwartz, David Shapiro (Eds.) - Consciousness and Self-Regulation-Springer (1986)

Vol.
Richard J. Davidson
Gary E. Schwartz
David Shapiro Editors
Consciousness and
Self-Regulation
Consciousness and
Self-Regulation
Advances in Research and Theory
VOLUME4
A Continuation Order Plan is available for this series. A continuation order will bring
delivery of each new volume immediately upon publication. Volumes are billed only
upon actual shipment. For further information please contact the publisher.
Consciousness and
Self-Regulation
Advances in Research and Theory
VOLUME 4
Edited by
RICHARD]. DAVIDSON
University of Wisconsin, Madison
Madison, Wisconsin
GARY E. SCHWARTZ
Yale University
New Haven, Connecticut
and
DAVID SHAPIRO
University of California at Los Angeles
Los Angeles, California
SPRINGER SCIENCE+BUSINESS MEDIA, LLC

Library of Congress Cataloging in Publication Data
Main entry under title:
Consciousness and self-regulation.
Vol. 3, 4 has subtitle: Advances in research and theory.

Vol. 3, 4 edited by Richard J. Davidson, Gary E. Schwartz, and David Shapiro.
Includes bibliographies and indexes.
I. Consciousness. 2. Self-control. I. Schwartz, Gary E., 1944- . II. Shapiro, David,
1924- . lll. Davidson, Richard J.
BF311.C64 1976 !53 76-8907
ISBN 978-1-4757-0631-4 ISBN 978-1-4757-0629-1 (eBook)
DOI 10.1007/978-1-4757-0629-1
© 1986 Springer Science+Business Media New York

Originally published by Plenum Press, New York in 1986
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted

in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher
Contributors
BARRY H. CoHEN, Department of Psychology, State University of New

York, Purchase, New York
BARRY R. DwoRKIN, Pennsylvania State College of Medicine, Milton S.
Hershey Medical Center, Hershey, Pennsylvania
NINA GoLDMAN, Rutgers-The State University, New Brunswick, New
Jersey
CLAUDIA W. HoovER, Center for Behavioral Medicine, Rockville,
Maryland
J. MICHAEL LACROIX, Department of Psychology, Glendon College, York
University, Toronto, Ontario
PAUL M. LEHRER, University of Medicine and Dentistry of New Jersey,
Rutgers Medical School, Busch Campus, Piscataway, New Jersey
}AMES W. LEWIS, Mental Health Research Institute, University of Mich-
igan, Ann Arbor, Michigan
JoHN C. LIEBESKIND, Department of Psychology, University of California,
Los Angeles, Los Angeles, California
LINDA R. NELSON, Department of Psychology, University of California,
Los Angeles, Los Angeles, California
DoNALD A. NoRMAN, Institute for Cognitive Science, University of Cal-
ifornia, San Diego, La Jolla, California
}AMES W. PENNEBAKER, Department of Psychology, Southern Methodist
University, Dallas, Texas
CuRT A. SANDMAN, Fairview Hospital, Costa Mesa, California
TIM SHALLICE, Medical Research Council, Applied Psychology Unit,
Cambridge, England
YEHUDA SHAVIT, Department of Psychology, University of California1 Los
Angeles, Los Angeles, California
v
vi CoNTRIBUTORS
GREGORY W. TERMAN, Department of Psychology, University of Califor-

nia, Los Angeles, Los Angeles, California
RoBERT L. WooLFOLK, Rutgers-The State University, New Brunswick,
New Jersey
Preface
In the Preface to the third volume, we described the evolution of this

Series and the changes that have taken place in the field since the first
volume appeared. The contents of the current volume continue the com-
mitment to a broadly based perspective on research related to con-
sciousness and self-regulation which was embodied in the previous
three volumes. Chapters are included which consider the role of con-
sciousness in cognitive theory and clinical phenomena. Several of the
contributions to this volume are concerned with the nature of self-reg-
ulation and the role of conscious processing in the mediation of self-
regulated behavior. Most of the authors adopt a psychobiological ap-
proach to their subject matter. Our selection of contributors with a bias
toward this approach reflects our own views that the psychobiological
approach is a very fruitful one and that the "architecture" of the nervous
system places important constraints on the types of theories that are
possible in this emerging area.
While the subject matter of the chapters in this volume is quite
diverse, the contributions are united by their emphasis on the impor-
tance of consciousness and/or self-regulation in the understanding of
behavior and experience. We have selected what we believe is repre-
sentative of the best theory and research in the diverse areas which bear
on the theme of this series, maintaining a balance between basic and
clinical research. We are confident that developments in the field will
warrant future volumes in which the outstanding work of the previous
two years are selected for extended treatment and review.
RICHARD J. DAVIDSON
GARY E. ScHWARTZ
DAVID SHAPIRO
vii
Overview
The first chapter "Attention to Action: Willed and Automatic Control

of Behavior" by Donald A. Norman and Tim Shallice underscores the
importance of concepts of consciousness for contemporary cognitive
psychology. These theories explain the importance of attentional inter-
ventions for certain types of action. Willful attention is required for ac-
tion associated with the modification of a plan, the implementation of
a novel alternative action sequence, or the prevention of an habitual act.
Norman and Shallice propose a supervisory attentional system (SAS)
which is activated when actions require "deliberate attention." They
propose that this system is subserved by the frontal cortex and associate
will with this form of attentional process.
The relation between volition and consciousness, a theme intro-
duced in the first chapter, is continued in the second chapter entitled
"The Motor Theory of Voluntary Thinking" by Barry H. Cohen. Cohen
reviews theories that argue for a central role of motor activity in various
forms of conscious processing. He concludes that motor activity (at least
the central efferent commands which specify such activity) is necessary
only for voluntary thinking. The experience of volition is said to arise
from the generation of activity in the motor system. Certain types of
perception and imagery of an automatic nature (which are not experi-
enced as volitional) are said not to require the generation of activity in
the motor system. Cohen concludes his chapter with the specification
of some useful experimental approaches to systematically investigate
the question of the necessity of motor activity in the experience of
volition.
Curt A. Sandman in his chapter, "Cardiac Afferent Influences on
Consciousness," extends the psychophysiological work of the Laceys
on cardiac afferent feedback to the brain. Sandman and his colleagues
have used novel experimental procedures to probe the influence of car-
ix
X OVERVIEW
diac feedback on neurobehavioral processes. For example, he examined

brain evoked potentials elicited by stimulii presented at different phases
of the cardiac cycle. As would be predicted from the Laceys' theory,
responses evoked by stimuli presented at the diastole were more dra-
matically influenced by cardiac phase compared with stimuli presented
at other portions of the cardiac cycle. Hemispheric asymmetries were
also noted in the topography of the responses, suggesting that cardiac
afferent feedback may be lateralized. Sandman has also used several
additional novel experimental procedures to examine other features of
cardiac feedback. He concludes the chapter by considering some of the
implications of rhythmic fluctuations in physiological activity for certain
aspects of consciousness.
The next chapter by James W. Lewis, Linda R. Nelson, Gregory W.
Terman, Yehuda Shavit, and John C. Liebeskind entitled "Intrinsic Con-
trol Mechanisms of Pain Perception" surveys recent findings on different
endogenous mechanisms of pain control. The authors note that stress
produced analgesia can operate through both opioid and non-opioid sys-
tems. Importantly, these investigators have demonstrated that the type
of stressor that releases opioids is both immunosuppressive and pro-
duces learned helplessness. They thus have identified a cluster of effects
resulting from a specific form of stress. These findings underscore the
complex and differentiated nature of stress. Stress-produced analgesia
is not a unitary phenomenon. Different neurochemical, neuroanatomical
and hormonal substrates appear to exist for different forms of analgesia.
Which of these is called into play appears to depend on variations in
the parameters of the stress stimulus. One of the parameters that is most
important is that of control over the stressor. Stressful stimuli that were
least likely to be perceived as controllable are also those that are asso-
ciated with opioid analgesia, learned helplessness and immuno-
suppression.
The clinical implications of variations in different aspects of con-
scious processing is a theme which is taken up next by James W.
Pennebaker and Claudia W. Hoover in their chapter "Inhibition and
Cognition: Toward an Understanding of Trauma and Disease." They
begin their chapter by establishing a relation between early trauma and
reports of illness. Those subjects who experienced early trauma, par-
ticularly of a sexual nature, had more symptoms of illness. Pennebaker
and Hoover also report that those who had sexual trauma were less
likely to confide in others about their trauma compared with subjects
who had been subjected to other traumatic events. The authors then
report the findings of a questionnaire study where they compared sub-
jects who had experienced trauma and confided with those who had
experienced comparable trauma but who had not confided. They found
OvERVIEW xi
that subjects who had not confided were more likely to have illnesses
compared with those who had. The authors consider the failure to con-
fide as an example of inhibition. They consider inhibition to require
mental work and find that it is itself stressful. The data and theorizing
that is presented in this chapter raises several important and fascinating
questions concerning the role of verbalization about and consciousness
of stressful events in the illness process.
In the next chapter, "Mechanisms of Biofeedback Control: On the
Importance of Verbal (Conscious) Processing," J. Michael Lacroix con-
siders the nature of learning in biofeedback. He first reviews Brenner's
notions of how learning in the biofeedback situation occurs. Brenner
argued that when learning to control autonomic responses, subjects first
learn to discriminate and identify interoceptive afferent processes re-
lated to the target response. Lacroix's review of the empirical evidence
finds little support for this suggestion. Training in biofeedback does not
enhance subjects' ability to discriminate changes in the target physio-
logical response. Performance on tests of response discrimination are
usually uncorrelated with performance on tests of response control.
And, training in biofeedback does not usually lead to greater response
specificity as a function of practice. On the basis of these observations,
Lacroix argues that a verbally mediated feedforward process can account
for most forms of learning within the biofeedback situation. It is only
in those relatively rare situations when the "verbal library" cannot access
information relevant to the appropriate behavioral programs that learn-
ing could be expected to proceed along the lines suggested by Brenner.
The next chapter continues the concern with the nature of learning
associated with physiological self-regulation. Barry R. Dworkin, in his
essay entitled "Learning and Long-term Physiological Regulation", pre-
sents a theory of autonomic self-regulation. The theory assumes the
existence of Type II learning of visceral responses. The theory is novel
with respect to contemporary treatments of biofeedback in that it em-
phasizes the interaction between central events and local autoregulatory
processes. Dworkin squarely places these phenomena within their ap-
propriate biological contexts. The theory also stresses the notion that
the nervous system participates in both short-term dynamic adjustments
and long-term or steady state adjustments.
In the final chapter by Paul M. Lehrer, Robert L. Woolfolk, and
Nina Goldman entitled "Progressive Relaxation Then and Now: Does
Change Always Mean Progress?" the concepts of self-regulation are ex-
tended to the clinical domain in a discussion of the efficacy and mech-
anisms of different relaxation techniques. These authors present a re-
view of recent literature on progressive relaxation, a relaxation technique
first developed by Edmund Jacobson in the 1930s. Various procedural
xii OVERVIEW
variations in the administration of relaxation are evaluated, such as live

versus taped instructions, number of sessions required and the number
of muscles which are trained simultaneously. Lehrer et al. then proceed
to a discussion of cognitive and somatic factors in anxiety and its treat-
ment. They build on the theory of cognitive and somatic subcomponents
of anxiety that was put forth by Davidson and Schwartz in 1976. Lehrer
et al. 's review of recent literature supports the Davidson and Schwartz
position that cognitive relaxation techniques have a greater effect on
cognitive symptoms while somatic techniques, like progressive relaxa-
tion, have greater effects on the somatic components of anxiety. The
role of suggestion and biofeedback in relaxation training are also con-
sidered. The authors conclude that anxiety and relaxation are complex,
multifactorial processes. Different procedures, which emphasize differ-
ent features of relaxation, will have reliably different effects.
Contents
1. Attention to Action: Willed and Automatic Control

of Behavior 1
DoNALD A. NoRMAN AND TIM SHALLICE
I. Theory 3
A. Contention Scheduling 5
B. Determination of Activation Values 6
C. The Supervisory Attentional System 6
II. Evidence 8
A. Evidence for a Distinct Supervisory Attentional System:
Neuropsychological Findings 8
B. Attentional Processes Only Modulate Schema Selection
10
C. Attentional Resources Are Primarily Relevant for Action
Selection 13
D. Competition between Tasks 13
E. Will and Deliberate Conscious Control 14
References 16
2. The Motor Theory of Voluntary Thinking 19

BARRY H. COHEN
I. Previous Motor Theories 20

A. The Motor Theory of Consciousness 20
B. The Motor Theory of Thinking 22
C. The Motor Theory of Attention 24
II. Defining the Motor Theory of Voluntary Thinking
26
A. The Experience of Volition 27
xiii
xiv CONTENTS
B. Specific Motor Associations 28

C. General Motor Associations 34
III. The Role of Covert Motor Activity in External Attention
38
A. Voluntary Attention 38
B. Spontaneous Attention 39
IV. Regulating the Stream of Thought 40
A. Ordinary States 40
B. Natural Relaxation 41
C. Progressive Relaxation and Meditation 43
V. Conclusions 45
A. Summary 45
B. Implications for Experimental Research 46
References 52
3. Cardiac Afferent Influences on Consciousness 55

CURT A. SANDMAN
I. Introduction 55
A. Consciousness as Cognition and Emotion (Brain and
Body) 56
II. Cardiovascular and Baroreceptor Physiology 57
III. Cardiovascular Relations to Behavior 63
A. Control of Body, Control of Consciousness 63
B. Effects on Consciousness of Transient Changes in the
Cardiovascular System 67
IV. Influence of Cardiovascular System on the Brain 69
A. The EEG 69
B. Auditory Event-related Potentials 74
C. Evoked Vascular Responses (EVRs) 77
V. Conclusions 80
References 81
4. Intrinsic Control Mechanisms of Pain Perception

87
JAMES w. LEWIS, LINDA R. NELSON, GREGORY w. TERMAN, YEHUDA
SHAVIT, AND JoHN C. LIEBESKIND
I. Introduction 87
II. Evidence for Endogenous Mechanisms of Pain Control
88
III. Activation of Endogenous Analgesia Systems: Stress-
induced Analgesia 89
XV
CoNTENTS
IV. Opioid and Nonopioid Mechanisms of Stress Analgesia

91
V. Role of Endocrine Systems in Stress Analgesia 91
VI. Neurotransmitters Involved in Stress Analgesia 93
VII. The Neuroanatomy of Stress Analgesia 94
VIII. Evidence for Two Independent Opioid Forms of Stress
Analgesia 95
IX. Early Experience Influences Stress Analgesia: The Effects of
Prenatal Alcohol Exposure 97
X. Immunosupressive and Tumor-enhancing Effects of Stress
98
XI. Summary and Conclusions 99
References 100
5. Inhibition and Cognition: Toward an Understanding

of Trauma and Disease 107
}AMES W. PENNEBAKER AND CLAUDIA W. HOOVER
I. Introduction 107
II. Personal Traumas and Health 109
A. Initial Survey 110
B. Psychology Today Sample 112
C. Medical Psychology Class Survey 115
III. The Failure to Confide: The Role of Inhibition 118
A. The Nature of Inhibition 119
B. Inhibition and Physiological Activity 120
C. , Long-Term Inhibition: Personality Considerations
124
D. Failure to Confide as Inhibition 128
IV. Summary and Final Considerations 129
A. Inhibition and Psychotherapy 130
B. Cognitive Organization and Self-disclosure 130
C. Links to Other Theories 131
D. Summary 133
References 133
6. Mechanisms of Biofeedback Control: On the

Importance of Verbal (Conscious) Processing 137
J. MICHAEL LACROIX
I. Introduction 137
II. A Promise Unfulfilled 138
xvi CONTENTS
III. The Verbal (Consciousness?) Interface 147

IV. A Two-process Theory 152
References 159
7. Learning and Long-term Physiological Regulation

163
BARRY R. DWORKIN
References 181
8. Progressive Relaxation Then and Now: Does Change

Always Mean Progress? 183
PAUL M. LEHRER, RoBERT L. WooLFOLK, AND NINA GoLDMAN
I. Introduction 183
II. Pedagogy versus Technology 187
A. Number of Sessions 188
B. One Muscle at a Time versus All at Once 188
c. Live versus Taped Relaxation 189
D. Conditioned Relaxation 190
E. The "Pendulum Effect" 191
F. The Use of Biofeedback 194
III. Differing Views on Cognitive and Somatic Factors in
Emotion and in Treatment 195
IV. Suggestion 202
v. Conclusion 209
References 210
Author Index 217

Subject Index 223
Contents of Previous
Volumes
Volume 1
1 A Model of Consciousness
E. Roy John
2 Self-Consciousness and Intentionality: A Model Based on an

Experimental Analysis of the Brain Mechanisms Involved in the
Jamesian Theory of Motivation and Emotion
Karl H. Pribram
3 Self-Regulation of Stimulus Intensity: Augmenting/Reducing and

the Average Evoked Response
Monte Buchsbaum
4 Neodissociation Theory of Multiple Cognitive Control Systems

Ernest R. Hilgard
5 Hypnotic Susceptibility, EEG-Alpha, and Self-Regulation

David R. Engstrom
6 Toward a Cognitive Theory of Self-Control

Donald Meichenbaum
7 Physiological and Cognitive Processes in the Regulation of Anxiety

Thomas D. Borkovec
8 Dreaming: Experimental Investigation of Representational and

Adaptive Properties
David B. Cohen
9 Biofeedback and the Twilight States of Consciousness

Thomas H. Budzynski
xvii
xviii CoNTENTS OF PREVIous VoLUMES
Volume 2
1 The Human Brain and Conscious Activity

A. R. Luria
2 Imagery and Thinking: Covert Functioning of the Motor System

F. J. McGuigan
3 Regulation of the Stream of Consciousness: Toward a Theory of

Ongoing Thought
Kenneth S. Pope and Jerome L. Singer
4 Self-Deception, Self-Confrontation, and Consciousness

Harold A. Sacken and Ruben C. Gur
5 Visceroception, Awareness, and Behavior

Gyorgy Adam
6 Stimulus-Bound Behavior and Biological Self-Regulation: Feeding,

Obesity, and External Control
Judith Rodin
7 Operant Conditioning of Autonomic Responses: One Perspective

on the Curare Experiments
Larry E. Roberts
8 Acquired Control of Peripheral Vascular Responses

Uwe Schuri
9 On the Nature of Alpha Feedback Training

Matrin T. Orne and Stuart K. Wilson
10 Passive Meditation: Subjective, Clinical, and Electrographic

Comparison with Biofeedback
Charles F. Stroebel and Bernard C. Glueck
CoNTENTS OF PREvious VouJMES xix
Volume 3
1 A Reason for Doubting the Existence of Consciousness

Georges Rey
2 Conscious Contents Provide the Nervous System with Coherent,

Global Information
Bernard J. Baars
3 Event-Related Brain Potentials in the Study of Consciousness

Emanuel Donchin, Gregory McCarthy, Marta Kutas, and
Walter Ritter
4 Anxiety and Fear: Central Processing and Peripheral Physiology

Peter J. Lang, Gregory A. Miller, and Daniel N. Levin
5 Meditation: In Search of a Unique Effect

Robert R. Pagano and Stephen Warrenburg
1 Attention to Action
Willed and Automatic Control of Behavior
DoNALD A. NoRMAN AND TIM SHALLICE
Much effort has been made to understand the role of attention in per-
ception; much less effort has been placed on the role attention plays in
the control of action. Our goal in this chapter is to account for the role
of attention in action, both when performance is automatic and when
it is under deliberate conscious control. We propose a theoretical frame-
work structured around the notion of a set of active schemas, organized
according to the particular action sequences of which they are a part,
awaiting the appropriate set of conditions so that they can become se-
lected to control action. The analysis is therefore centered around ac-
tions, primarily external actions, but the same principles apply to in-
ternal actions-actions that involve only the cognitive processing
mechanisms. One major emphasis in the study of attentional processes
is the distinction between controlled and automatic processing of percep-
tual inputs (e.g., Shiffrin & Schneider, 1977). Our work here can be seen
as complementary to the distinction between controlled and automatic
processes: we examine action rather than perception; we emphasize the
situations in which deliberate, conscious control of activity is desired
rather than those that are automatic.
In this chapter we will be particularly concerned with the different
ways in which an action is experienced. To start, examine the term
automatic: it has at least four different meanings. First, it refers to the
way that certain tasks can be executed without awareness of their per-
formance (as in walking along a short stretch of flat, safe ground). Sec-
DoNALD A. NoRMAN e Institute for Cognitive Science, University of California, San Diego,
La Jolla, California 92093. TIM SHALLICE e Medical Research Council, Applied Psychology
Unit, 15 Chaucer Road, Cambridge, CB2 2EF England. Research support to D. A. Norman
was provided by the Personnel and Training Research Programs, Office of Naval Research
under contract N00014-79-C-0323. The collaboration was made possible by a grant from
the Sloan Foundation to the Program in Cognitive Science at UCSD. Support was also
provided by grant MH-15828 from the National Institute of Mental Health to the Center
for Human Information Processing.
1
2 DoNALD A. NoRMAN AND TIM SHALLICE
ond, it refers to the way an action may be initiated without deliberate

attention or awareness (as in beginning to drink from a glass when in
conversation). Third, it is used in cases such as the orienting response,
in which attention is drawn automatically to something, with no delib-
erate control over the direction of attention. And finally, within con-
temporary cognitive psychology, the term automatic is often defined op-
erationally to refer to situations in which a task is performed without
interfering with other tasks. In this situation, automatic is principally
defined to mean that the task is performed without the need for limited
processing resources _(Shiffrin & Schneider, 1977), although variations
on this theme are prevalent (e.g., Kahneman & Treisman, 1983; Posner,
1978).
It is possible to be aware of performing an action without paying
active, directed attention to it. The most general situation of this type
is in the initiation of routine actions. Phenomenally, this corresponds
to the state that Ach (1905) describes as occurring after practice in re-
action time tasks. Over the first few trials, he said, the response is pre-
ceded by awareness that the action should be made, but later there is
no such awareness unless preparation has been inadequate. In such
well-learned tasks the subject does experience the response as proceed-
ing with "an awareness of determination/' even if it is not immediately
preceded by any experience of intention to act. Awareness of deter-
mination can, however, be absent. One example comes from the study
of slips of action (Norman, 1981; Reason, 1979; Reason & Mycielska,
1982): one may find oneself doing a totally unexpected set of actions,
much to one's own dismay.
In contrast to acts undertaken without active, directed attention
being paid to them are those carried out under deliberate conscious
control. This distinction corresponds closely to Williams James's (1890)
distinction between "ideo-motor" and "willed" acts. To James, "wher-
ever movement follows unhesitatingly and immediately the notion of it
in the mind, we have ideo-motor action. We are then aware of nothing
between the conception and the execution." He contrasted these with
acts which require will, where "an additional conscious element in the
shape of a fiat, mandate, or expressed consent" is involved.
Experientially, a number of different sorts of tasks appear to require
deliberate attentional resources. These tasks fit within the following
categories:
1. They involve planning or decision making
2. They involve components of troubleshooting
3. They are ill-learned or contain novel sequences of actions
4. They are judged to be dangerous or technically difficult
ATIENriON TO ACTION 3
5. They require the overcoming of a strong habitual response or

resisting temptation.
The general principle involved is that these are special situations in
which the uncontrolled application of an action schema is not desired
for fear that it might lead to error.
I. THEORY
Our goal is to account for several phenomena in the control of ac-
tion, including the several varieties of action performance that can be
classified as automatic, the factthat action sequences that normally are
performed automatically can be carried out under deliberate conscious
control when desired, and the way that such deliberate control can be
used both to suppress unwanted actions and to enhance wanted ones.
In addition, we take note both of the fact that accurate, precise timing
is often required for skilled performance and the fact that it is commonly
believed that conscious attention to this aspect of performance can dis-
ru,pt the action. Finally, in normal life numerous activities often overlap
one another, so that preventing conflicts between incompatible actions
is required.
These phenomena pose strong constraints upon a theory of action.
The theory must account for the ability of some action sequences to run
themselves off automatically, without conscious control or attentional
resources, yet to be modulated by deliberate conscious control when
necessary. Accordingly, we suggest that two complementary processes
operate in the selection and control of action. One is sufficient for rela-
tively simple or well learned acts. The other allows for conscious, at-
tentional control to modulate the performance. The basic mechanism,
contention scheduling, which acts through activation and inhibition of sup-
porting and conflicting schemas, is proposed as the mechanism for
avoiding conflicts in.performance. Precise timing is handled by means
of "triggers" that allow suitably activated schemas to be initiated at the
precise time required. The mechanisms for contention scheduling and
triggers follow those developed by McClelland and Rumelhart (1981)
and Rumelhart and Norman (1982).
Start by considering a simple, self-contained, well-learned action
sequence, perhaps the act of typing a word upon the receipt of a signal.
This action sequence can be represented by a set of schemas, which
when triggered by the arrival of the appropriate perceptual event result
in the selection of the proper body, arm, hand, and finger movements.
Whenever the action sequence is effected, its representation by means
r----------------------
i C~omponent Schemas j Ext•rnal &
• 1 Internal
: I Actions
l' __________________ ---:'
SOURCE SCHEMA PSYCHOLOGICAL
PROCESSING STRUCTURES
FIGURE 1. A horizontal thread. For well-learned, habitual tasks an autonomous, self-suf-

ficient strand of processing structures and procedures can usually carry out the required
activities without the need for conscious or attentional control. Selection of component
schemas is determined, in part, by how well the "trigger conditions" of the schema match
the contents of the "trigger data base." Such a sequence can often be characterized by a
(relatively) linear flow of information among the various psychological processing struc-
tures and knowledge schemas involved: a horizontal thread.
of action schemas constitutes a "horizontal thread." The important point.

is that the processing structures which underlie a horizontal thread can
in principle be well specified. The general nature of the processing struc-
ture for a simple action sequence is shown in Figure 1.
When numerous schemas are activated at the same time, some
means must be provided for selection of a particular schema when it is
required. At times, however, there will be conflicts among potentially
relevant schemas, and so some sort of conflict resolution procedure must
be provided. This is a common problem in any information-processing
system in which at any one moment several potential candidates for
operation might require access to the same resources or might result in
incompatible actions. (McDermott & Forgy, 1978, discuss this issue for
production systems and Bellman, 1979, discusses the problem with re-
spect to animal behavior.)
The procedure we propose is constrained by the desire to transmit
properties by means of the single variable of amount of activation, a
concept consistent with current psychological theory. We propose that
the individual schemas of the horizontal threads each have an activation
value that is determined by a combination of factors, some that operate
among schemas, some that result from special processes that operate
upon the schemas.
A schema is selected once its activation level exceeds a threshold.
Once selected, it continues to operate, unless actively switched off, until
it has satisfied its goal or completed its operations, or until it is blocked
when some resource or information is either lacking or is being utilized
by some more highly activated schema. The activation value is important
ATIENTION TO ACTION 5
primarily in the selection process and when the selected schema must
compete either for shared resources or in providing component schemas
with initial activation values.
The scheduling is, therefore, quite simple and direct. No direct at-
tentional control of selection is required (or allowed). Deliberate atten-
tion exerts itself indirectly through its effect on activation values. All the
action, therefore, takes place in the determination of the activation val-
ues of the schemas.
A. Contention Scheduling
To permit simultaneous action of cooperative acts and prevent si-
multaneous action of conflicting ones is a difficult job, for often the
details of how the particular actions are performed determine whether
they conflict with one another. We propose that the scheduling of actions
takes place through what we call contention scheduling, which resolves
competition for selection, preventing competitive use of common or re-
lated structures, and negotiating cooperative, shared use of common
structures or operations when that is possible. There are two basic prin-
ciples of the contention scheduling mechanism: first, the sets of potential
source schemas compete with one another in the determination of their
activation value; second, the selection takes place on the basis of acti-
vation value alone-a schema is selected whenever its activation exceeds
the threshold that can be specific to the schema and could become lower
with use of the schema.
The competition is effected through lateral activation and inhibition
among activated schemas. What degree of lateral inhibition exists be-
tween schemas on the model remains an open issue. Schemas which
require the use of any cqmmon processing structures will clearly need
to inhibit each other. Yet the degree of inhibition cannot be determined
simply a priori. Thus, some aspects of the standard refractory period
phenomena can be plausibly attributed to such inhibition between sche-
mas; explanations based upon conflicts in response selection fit the data
well (Kahneman, 1973). Unfortunately, it is not always clear how to
determine when two tasks use common processing structures. The ex-
perimental literature on refractory periods reveals interference between
tasks involving the two hands. This suggests that responses involving
the two hands may use common processing structures. However, one
cannot assume that the two hands inevitably involve a common pro-
cessing structure, as refractory period effects can disappear if highly
compatible tasks are used (Greenwald & Shulman, 1973). On the model,
as tasks become better learned, the schemas controlling them could be-
come more specialized in their use of processing structures, reducing

potential structural interference and minimizing the need for mutual
inhibition among schemas.
B. Determination of Activation Values

We divide activational influences upon a schema into four types:
influences from contention scheduling, from the satisfaction of trigger
conditions, from the selection of other schemas, and from "vertical
thread" influences. Trigger conditions specify under what conditions a
schema should be initiated, thus allowing for precise environmental
control of performance. How well existing conditions match the trigger
specifications determines the amount of activation contributed by this
factor.
Selection of one schema can lead to the activation of others. Any
given action sequence that has been well learned is represented by an
organized set of schemas, with one-the source schema-serving as the
highest-order control. The term source is chosen to indicate that the other
component schemas of an action sequence can be activated through the
source. We assume that the initial activation values of component sche-
mas are determined by means of their source schema. For example,
when the source schema for a task such as driving an automobile has
been selected, all its component schemas become activated, including
schemas for such acts as steering, stopping, accelerating, slowing, ov-
ertaking, and turning. Each of these component schemas in turn acts
as a source schema, activating its own component schemas (braking,
changing gear, signalling, and so on).
C. The Supervisory Attentional System

The horizontal thread specifies the organization structure for the
desired action sequence. However, a schema may not be available that
can achieve control of the desired behavior, especially when the task is
novel or complex. In these cases, some additional control structure is
required. We propose that an additional system, the Supervisory At-
tentional System (SAS), provides one source of control upon the selec-
tion of schemas, but it operates entirely through the application of extra
activation and inhibition to schemas in order to bias their selection by
the contention-scheduling mechanisms. (A planning mechanism which
performs an analogous function in problem-solving programs has been
simulated by various researchers; see, for example, Boden, 1977). The
HORIZONT Al VE RT ICAL THREA DS

PROCESSING THREADS
PS YCHOlOO.CAl External
PROCESSING & lnltrnal
STRUCTYRES Ac:rions
FIGURE2 . The overall system: Vertical and horizontal threads . When attention to particular
tasks is required, vertical thread activation comes into play . Attention operates upon sche-
mas only through manipulation of activation values, increasing the values for desired
schemas, decreasing (inhibiting) the values for undesired ones . Motivational variables are
assumed to play a similar role in the control of activation, but working over longer time
periods. To emphasize that several tasks are usually active, with the individual components
of each task either being simultaneous or overlapping in time, this figure shows five dif-
ferent horizontal threads. Some means of selecting the individual schemas at appropriate
times while providing some form of conflict resolution becomes necessary. The interactions
among the various horizontal threads needed for this purpose are indicated by the lines
that interconnect schemas from different threads .
overall system is shown in Figure 2. Note that the operation of the SAS
provides only an indirect means of control of action . Attention, which
we will associate with outputs from SAS, controls only activation and
inhibition values, not selection itself. Moreover, it is control overlaid on
the horizontal thread organization. When attentional activation of a
schema ceases, the activational value will decay back to the value that
other types of activating input would produce.
In addition, we assume that motivational factors supplement the
activational influences of the SAS. We take motivation to be a relatively
slow-acting system, working primarily to bias the operation of the hor-
izontal thread structures toward the long-term goals of the organism by
activating source schemas (and through their selection component
schemas).
II. EviDENCE
That horizontal thread control of action may be viewed within a

schema framework is too well known to need reviewing here (see, e.g.,
Pew, 1974, Schmidt, 1975). There are four major aspects of the model
that require assessment:
1. Actions under deliberate conscious control involve a specific
mechanism in addition to those used in automatic actions.
2. Attentional processes can modulate the selection process only
by adding activation or inhibition. Attention to action is neither
sufficient nor necessary to cause the selection of an action
sequence.
3. Attentional processes are primarily relevant to the initiation of
actions, not for their execution.
4. Selection between competing action sequences takes place
through the mechanism of contention scheduling.
Now let us examine the evidence for these aspects of the model.
A. Evidence for a Distinct Supervisory Attentional System:

Neuropsychological Findings
A major feature of our model is that for well-learned action se-
quences two levels of control are possible: deliberate conscious control
and automatic contention scheduling of the horizontal threads. Possibly
the strongest evidence for the existence of both levels comes from neu-
ropsychology. The functions we assume for the supervisory attentional
control-those that require "deliberate attention" -correspond closely
with those ascribed by Luria (1966) to prefrontal regions of the brain,
thought by Luria to be required for the programming, regulation, and
verification of activity. In this view, if the Supervisory Attentional Sys-
tem were damaged the resulting behavior should be similar to that ex-
hibited by patients with prefrontal lesions.
On the model, well-learned cognitive skills and cognitive proce-
dures do not require the higher-level control system. Higher-level con-
trol becomes necessary only if error correction and planning have to be
performed, if the situation is novel, or temptation must be overcome.
It is well known in clinical neuropsychology that lesions confined to
prefrontal structures leave the execution of basic skills such as the use
of objects, speaking, and writing unaffected (see Walsh, 1978, for re-
view). "Well able to work along old routine lines" is a classical char-
acterization of such patients (Goldstein, quoted by Rylander, 1939).

Quantitatively it has, for instance, been shown by McFie (1960) that
performance of WAIS subtests is relatively unaffected by lesions to the
frontal lobes. The model does predict impairments in the performance
of tasks that require error correction or planning, or are in some basic
way novel-just the constellation of deficits that are observed clinically
in the so-called frontal syndrome (see Walsh, 1978).
Evidence for the contrast in performance in the two types of situ-
ations can be obtained from case studies of patients with frontal lobe
lesions. A classic study was that carried out by Lhermitte, Derouesne,
and Signoret (1972). Their two principal patients could perform certain
Verbal and Performance WAIS subtests at normal level (Derouesne, per-
sonal communication). These were tasks which require the use of well-
learned skills in routine fashion. Thus digit span, which uses mainte-
nance rehearsal schemas, was well performed. When much novel pro-
gramming of the external and internal action sequence was required,
performance was extremely poor. Examples were WAIS Block Design
or the reproduction of a complex figure-the Figure of Rey. However,
performance could be greatly improved by providing a program for the
patient; in the case of the Figure of Rey, this involved breaking down
the total design into a series of hierarchically organized subcomponents.
Group studies of neurological patients also provide support. It is
well established that patients with frontal lobe lesions have difficulties
with error correction. The Wisconsin card-sorting test involves multi-
dimensional stimuli and requires the patient to switch from sorting on
one dimension to sorting according to another. In this task, frontal pa-
tients show a strong tendency to perseverate in sorting on the previously
correct dimension, even when they are told they are wrong (Milner,
1964; Nelson, 1976). Planning, too, has been shown to present diffi-
culties for these patients. The simplest example of such a defect is Gad-
zhiev's finding (see Luria, 1966) that frontal patients presented with a
problem tend to miss out the initial assessment of the situation. Shallice
and McCarthy (see Shallice, 1982) showed that patients with left frontal
lesions are significantly more impaired than those with lesions in other
sites in look-ahead puzzles related to the Tower-of-Hanoi; comparison
of performance on this task with that on other tasks suggested that it
was the planning component of the task that was affected. Novel learn-
ing tasks have also been shown to produce specific difficulties for frontal
lobe patients. Petrides and Milner (see Milner, 1982) found that both
patients with left frontal and those with right frontal lesions were sig-
nificantly impaired in learning new arbitrary pairings presented one at
a time in a random sequence.
Prediction about the effect of an impairment to the Supervisory
10 DoNALD A. NoRMAN AND TrM SHALLICE
Attentional System can be approached in another way. On the model,

the failure of this single mechanism can give rise to the apparent con-
tradiction between increased perseveration and increased distractability,
depending on the pattern of trigger-schema relations. What would be
expected if behavior is left under the control of horizontal thread struc-
tures plus contention scheduling? If one schema is more strongly acti-
vated than the others, it will be difficult to prevent it from controlling
behavior. By contrast, when several schemas have similar activation
values, one should obtain another clinical characteristic of frontal pa-
tients: an instability of attention and heightened distractability (see Ry-
lander, 1939; Walsh, 1978). Both types of results are also observed in
animals with prefrontal lesions (see Fuster, 1980, for review).
If the properties of the Supervisory Attentional System seem to
correspond fairly well with neuropsychological evidence, does the same
apply to the properties of contention scheduling? One possible relation
is with mechanisms in the corpus striatum of the basal ganglia, often
thought to be involved in the selection of actions (see Denny-Brown &
Yanagisawa, 1976; Marsden, 1982). The basal ganglia are innervated by
one of the major dopamine projections, and dopamine release is in turn
facilitated by amphetamine. Robbins and Sahakian (1983) have provided
an explanation of the effects of increased doses of amphetamine based
on the work of Lyon and Robbins (1975), in terms closely related to ours.
The account goes like this: Increased amphetamine results in an increase
in the speed with which response sequences are carried out and a de-
crease in the interval between them. At higher levels "competition for
expression via the motor or executive system begins to occur between
different sequences with the result that some sequences are aborted and
their terminal elements are lost. Eventually, the performance of a com-
plete sequence is drastically attenuated and the stereotype occurs." Rob-
bins and Sahakian argue that increased dopamine release potentiates
the activation level of schemas and leads to an increasing number of
schemas being activated above threshold. In our terms, if the potentia-
tion becomes too great, the lateral inhibitory control of contention sched-
uling is broken. Many schemas are selected at the same time, producing
a jamming of almost all objects of behavior. Parkinsonism appears to
provide a complementary condition.
B. Attentional Processes Only Modulate Schema Selection

The motivation for this aspect of the model is that attentional control
is probably too slow and unwieldy to provide the high precision of
AITENTION TO AcnoN 11
accuracy and timing needed to perform skilled acts. Deliberate conscious

control is generally agreed to involve serial processing steps, each step
taking on the order of 100 msec or more. Such control would simply be
too slow to account for skilled human behavior that requires action se-
quences to be initiated just when environmental or internal conditions
call for them; in some situations they must be accurate to the nearest
20 msec. This is consistent with the general view that deliberate control
of skilled performance leads to deterioration of performance. Accord-
ingly, in the model we allow attentional processes only to bias or mod-
ulate the operation and selection of schemas. Precise timing is controlled
by the fit of stimulus input to that required by the set of trigger conditions
for a schema.
Other factors are also involved. Thus, despite one's desire to attend
to one set of signals, if the trigger conditions of another are sufficiently
well met, the other may be selected in contention scheduling despite
the attention directed toward the one: triggering activation can be more
powerful than activation from the Supervisory Attentional Mechanisms.
A classic example of this difficulty is the Stroop phenomenon. Another
set of relevant findings comes from the classical literature on selective
attention in which an attempt is made to keep the subject concentrating
upon a primary task while other signals are presented. Certain classes
of words presented upon a secondary channel can intrude upon or bias
primary task performance, such as a word that fits within the context
of the primary channel, or that has been conditioned to electric shock,
or that has high emotional value. Performance of the other task is im-
paired when the interrupt occurs (e.g., Treisman, 1960, or in the re-
fractory period paradigm, Helson & Steger, 1962). In terms of our model,
these "intrusions" result from data-driven entry of action schemas into
the contention-scheduling mechanism, and their selection there is due
to the strongly activating properties of such triggers.
Further evidence that attention serves a biasing or modulating role
comes from a study by McLean and Shulman (1978) that examined the
role of attention on the speed of performance in a letter-matching task.
Once a subject's attention had been directed toward a particular expec-
tation, performance remained biased toward that expectation even after
the subjects had been told that the expectation was no longer valid. The
bias decayed slowly, lastirig for around one second, thereby acting more
like the decay of activation from a memory structure than of an atten-
tional selection that could be quickly added or taken away. Although
the emphasis in this experiment was on perception rather than action,
their conclusion that attention acts by means of an activation level on
memory units (schemas, in our vocabulary) is support for this aspect of
the model.
Possibly the strongest evidence that conscious attentional control

is not necessary for the initiation or execution of action sequences comes
from the study of slips of action (Norman, 1981; Reason & Mycielska,
1982). In the class of errors known as "capture errors," the person ap-
pears to perform the action without either conscious control or knowl-
edge. Capture errors are easily illustrated by an example: one of Reason's
subjects described how, when passing through his back porch on the
way to get his car out, he stopped to put on his Wellington boots and
gardening jacket as if to work in the garden.
Consider what would happen on the model if a routine task is being
carried out that does not require continuous monitoring and activation
from the Supervisory Attentional System. Its component schemas can
be selected using contention scheduling alone, so the Supervisory At-
tentional System could be directed toward activating some other non-
competing schema (i.e., "thinking about something else") rand the com-
ponent schemas in the routine action would still be satisfactorily selected
by contention scheduling alone. Occasionally, though, a schema that
controls an incorrect action could become more strongly activated in
contention scheduling than the correct schema and capture the effector
systems. The supervisory system, being directed elsewhere, would not
immediately monitor this, and a capture error would result.
Findings from the diary study of Reason (1983) provide support for
an interpretation of this type of error in terms of the model. The data
show that people typically rate themselves as being "preoccupied" and
"distracted" in the situations wherein lapses occur. This would corre-
spond in our model to the case in which no activation is being received
for the "appropriate" schema from the supervisory system: instead, the
supervisory system is activating a different, noncompeting schema. In
Reason's data, both captured and capturing actions are rated as occur-
ring "very often" and being "automatic." Moreover, the captured and
capturing actions were rated as having very similar stimulus character-
istics. These characteristics are all consistent with the model: frequently
performed action sequences are apt to have developed sufficient hori-
zontal thread structure that they could be carried out by contention
scheduling alone-"automatically." The similarity of the captured and
capturing actions is consistent with the suggestion that some data-driven
activation of the capturing schema might take place and that trigger
conditions appropriate for one sequence are likely to be appropriate for
the other as well. All these factors maximize the chance of an incorrect
schema's being more activated in contention scheduling than the correct
one, thus leading to a capture error.
ArrENTION TO ACTioN 13
C. Attentional Resources Are Primarily Relevant for Action

Selection
One theme of the model is that attentional resources are relevant
only at the specific points in an action where schema selection is re-
quired. Thus, control of a hand movement in response to a signal will
usually require attentional resources twice: once to initiate the schemas
that start the motion, once to initiate the schemas that control termi-
nation of the motion (see Keele, 1973). This fits with the results of probe
studies during movement where responses to probes at the start or end
of the movement can be more delayed than those during execution (Pos-
ner & Keele, 1969). (The interpretation of probe studies is not straight-
forward-see McLeod, 1980-but U-shaped functions of the type ob-
tained by Posner and Keele seem unlikely to arise artifactually.) When
a simple movement is made to an external stop, the response time to a
probe during the movement appears to be no greater than if no move-
ment is being made (Posner & Keele, 1969; Ells, 1973). This suggests
that when hand motion can be stopped by an external device the move-
ment can be stopped without initiating an action sequence and without
attentional control.
D. Competition between Tasks

On the model, the degree to which two tasks will interfere with
each other depends upon a number of factors. These include structural
factors critical for the contention scheduling mechanism, the balance of
activation and inhibition in that mechanism, and the degree of learning
which is relevant mainly for the degree of involvement required of the
Supervisory Attentional System.
For most task combinations, precise prediction of the degree of in-
terference depends on too many unknown parameters (see Shallice,
McLeod, & Lewis, 1985). One obvious prediction is that "parallel" dual
task performance should be most easily possible when one or both of
the tasks can be performed without attentional control. This fits the
experimental literature on monitoring (see Duncan, 1980, for review).
When two response streams have to be initiated, the model makes the
standard prediction that parallel performance is more likely if subjects
are skilled and well practiced (see Allport, Antonis, & Reynolds, 1972;
McLeod, 1977; Spelke, Hirst, & Neisser, 1976). Note that even in these
situations performance normally deteriorates somewhat when two tasks
14 DoNALD A. NoRMAN AND TrM SHALLICE
are combined, even though there appear to be no obvious grounds for

structural or attentional interference. We feel this indicates that even
when the individual tasks are well learned at times there will be a need
for schemas that require vertical thread activation for rapid selection.
Thus, as Allport (1980) pointed out, in experiments involving piano play-
ing conducted by Allport, Antonis, and Reynolds (1972), the one subject
who showed no interference "was also the most competent of our pi-
anists." The other subjects all found some technical challenge in the
music such that "moments of emergency occurred" when recovery re-
quired some relatively unpracticed applications of keyboard technique
and therefore, on our model, attentional resources.
E. Will and Deliberate Conscious Control

A major goal of our approach has been to produce an explanation
for the different types of experience one can have of an action. Consider
the types of information the Supervisory Attentional System would re-
quire in order to carry out its complex functions. Representations of the
past and present states of the environment, of goals and intentions, and
of the repertoire of higher-level schemas it could activate would all have
to be available. Yet more would be necessary. The system would need
to know aspects of the operation of a selected schema or, to be more
precise, of those selected schemas which it could potentially activate
(source schemas). It would need to know not only which source schemas
had been selected but also the action sequences they produced and
probably the eliciting triggers as well. Without such information, error
correction would be a hopeless task, but it is a key function of the su-
pervisory system.
How an action is experienced is dependent upon what information
about it is accessed by the Supervisory Attentional System and upon
whether the supervisory system activates source schemas itself and, if
so, how strongly. This, therefore, allows a variety of states of awareness
of actions to exist.
Consider the different meanings of automatic discussed earlier. The
first two meanings which refer to automaticity in the initiation and car-
rying out of an action correspond to the selection and operation, re-
spectively, of a schema without the supervisory system's assessing in-
formation relevant to it. In contrast are those occasions when a trigger
not only activates a schema strongly and directly but also produces an
interrupt in the supervisory system itself. This corresponds to the third,
very different, meaning of automatic, wherein what is automatic is the
attention-demanding characteristics of the stimulus. When the super-
visory system does access some aspect of the triggering or selection of

schema, or where it monitors the action sequence itself while at the
same time providing no attentional activation to assist in schema selec-
tion, we have a correspondence for James' ideomotor acts. Schema se-
lection is elicited solely by triggers, but information about the process
is accessed at the higher level. ·
What happens when the supervisory system does produce atten-
tional activation to modulate schema selection? We propose that will be
this direction of action by deliberate conscious control. This definition
is consistent both with the popular meaning of the term and with the
discussions of will in the earlier psychological literature (e.g., James,
1890; Pillsbury, 1908). Thus, strongly resisting a habitual or tempting
action or strongly forcing performance of an action that one is loathe to
perform seem to be prototypical examples of the application of will. The
former would appear to result from deliberate attentional inhibition of
an action schema, the latter from deliberate activation.
In our view, will varies along a quantitative dimension correspond-
ing to the amount of activation or inhibition required from the super-
visory attentional mechanisms. The assumption that this activation
value lies on a continuum explains why the distinction between willed
and ideomotor actions seems quite clear in considering extreme actions
but becomes blurred in considering those that require very little atten-
tional effort. Thus, introspection fails in determining whether will is
involved in the voluntary lifting of the arm. But there is no need to make
a distinction if this act is simply identified as being near the zero point
of the quantitative scale of attentional activation.
The idea that will corresponds to the output of the Supervisory
Attentional System has certain other useful consequences. Consider the
errors that occur with brief lapses of attention, when there is a failure
to sustain will adequately. One type of error results following a decision
not to do a step within a habitual sequence of actions. To eliminate the
step requires deliberate (willful) inhibition of the relevant schema. If
there is a momentary lapse of attention to the deliberate inhibition, the
step may get done anyway. Closely related is the error that occurred to
one of us, who decided not to take another bite of a delicious but ex-
tremely rich dessert; with only a brief lapse of attention, the cake got
eaten.
Certain aspects of will require elaboration. In some circumstances
an action may seem to require no will at all, yet at other times it will
require extreme demands. Thus, getting out of bed in the morning is
at times an automatic act and at other times requires great exertion of
will. One explanation for this phenomenon is that activation of an action
schema by the attentional mechanisms necessarily involves knowledge
of consequences. When these are negative, they lead to inhibition of the

source schemas which then must be overcome. In some cases, the self-
inhibition can be so intense as to prevent or at least make very difficult
the intended act. Thus, inflicting deliberate injury to oneself (as in prick-
ing one's own finger in order to draw blood) is a difficult act for many
people.
The elicitation of strong activation from the supervisory attentional
mechanism is not necessarily unpleasant. Indeed, many sports and
games seem to be attractive because they do necessitate such strong
activation. In this case concentration is perhaps the more appropriate
experiential equivalent rather than will. In addition, will is not just a
matter of attention to actions. As Roy D' Andrade (personal communi-
cation) has pointed out, a willed act demands not only strong attentional
activation; it also depends on the existence of a "mandated decision,"
independent of one's attending-a conscious knowledge that the par-
ticular end is to be attained. This mandate, in our view, would be re-
quired before the supervisory attentional mechanisms will produce their
desired activation output. However, the critical point for the present
argument is that the phenomenal distinction between willed and ideo-
motor acts flows from the separation of the supervisory attentional
mechanisms from the systems they oversee. The phenomenology of
attention can be understood through a theory of mechanism.
ACKNOWLEDGMENTS
We thank members of the Skills group of the Cognitive Science

Laboratory at UCSD, especially David Rumelhart, Geoffrey Hinton,
Wynne Lee, Jonathan Grudin, and Bernie Baars. We appreciate thought-
ful reviews and comments by Roy D' Andrade, Steve Keele, John Long,
George Mandler, and Peter McLeod.
REFERENCES
ACH, N. (1905). Uber die Willenstiitigkeit und das Denken. Gottingen: Vardenhoek.
ALLPORT, D. A. (1980). Attention and performance. In G. L. Claxton (Ed.), New directions
in cognitive psychology. London: Routledge.
ALLPORT, D. A., ANTONIS, B., & REYNOLDS, P. (1972). On the division of attention: A dis-
proof of the single channel hypothesis. Quarterly Journal of Experimental Psychology, 24,
225-235.
BELLMAN, K. (1979). The conflict behavior of the lizard, Sceloporus Occidentalis, and its im-
plication for the organization of motor behavior. Unpublished doctoral dissertation. Uni-
versity of California, San Diego.
BoDEN, M. (1977). Artificial intelligence and natural man. New York: Basic Books.
AITENTION TO ACTION 17
DENNY-BRoWN, D., & YANAGISAWA, N. (1976). The role of the basal ganglia in the initiation
of movement. In M. D. YAHR (Ed.), The basal ganglia. New York: Raven Press.
DuNCAN, J. (1980). The locus of interference in perception of simultaneous stimuli. Psy-
chological Review, 87, 272-300.
ELLs, J. G. (1973). Analysis of temporal and attentional aspects of movement control. Journal
of Experimental Psychology, 99, 10-21.
FusTER, J. M. (1980). The prefrontal cortex: New York: Raven Press.
GREENWALD, A. G., & SHULMAN, A. G. (1973). On doing two things at once. II. Elimination
of the psychological refractory period. Journal of Experimental Psychology, 101, 70-76.
HELSON, H., & STEGER, J. A. (1962). On the inhibitory effect of a second stimulus following
the primary stimulus to react. Journal of Experimental Psychology, 64, 201-205.
}A.'v!ES, W. (1890). The principles of psychology. New York: Holt.
KAHNEMAN, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall.
KAHNEMAN, D., & TREISMAN, A.M. (1983). Changing views of attention and automaticity.
In R. PARASURAMAN, R. DAVIES, & ]. BEAITY (Eds.), Varieties of attention. New York:
Academic Press.
KEELE, S. W. (1973). Attention and human performance. Pacific Palisades, CA: Goodyear.
LHERMITTE, F., DEROUESNE, }., & SIGNORET, J-L. (1972). Analyse neuropsychologique du
syndrome frontale. Revue Neuropsychologique, 127, 415-440.
LuRIA, A. R. (1966). Higher cortical functions in man. London: Tavistock.
LYON, M., & RoBBINS, T. (1975). The action of central nervous system drugs: A general
theory concerning amphetamine effects. In W. B. EssMANN & L. VALZELLI (Eds.), Cur-
rent developments in psychopharmacology, Vol. 2. New York: Spectrum.
MARSDEN, C. D. (1982). The mysterious motor function of the basal ganglia. Neurology, 32,
514-539.
McCLELLAND, J. L., & RuMELHART, D. E. (1981). An interactive activation model of context
effects in letter perception: Part 1. An account of basic findings. Psychological Review,
88, 375-407.
McDERMOTT, ]., & FoRGY, C. (1978). Production system conflict resolution strategies. In
D. A. WATERMAN & F. HAYES-ROTH (Eds.), Pattern-directed inference systems. New York:
Academic Press.
McFIE, J. (1960). Psychological testing in clinical neurology. Journal of Nervous and Mental
Diseases, 131, 383-393.
McLEAN, J. P., & SHULMAN, G. L. (1978). On the construction and maintenance of expec-
tancies. Quarterly Journal of Experimental Psychology, 30, 441-454.
McLEoD, P. D. (1977). A dual task response modality effect: Support for multiprocessor
models of attention. Quarterly journal of Experimental Psychology, 29, 651-658.
McLEOD, P. D. (1980). What can probe RT tell us about the attentional demands of move-
ment? In G. E. STELMACH & ]. REQUIN (Eds.), Tutorials in motor behavior. Amsterdam:
North-Holland.
MrLNER, B. (1964). Some effects of frontal lobectomy in man. In J. M. WARREN & K. AKERT
(Eds.), The frontal granular cortex and behavior. New York: McGraw-Hill.
MILNER, B. (1982). Some cognitive effects of frontal-lobe lesions in man. Philosophical Trans-
actions of the Royal Society of London, Series B., 298, 211-226.
NELSON, H. (1976). A modified card sorting test sensitive to frontal lobe defects. Cortex,
12, 313-324.
NORMAN, D. A. (1981). Categorization of action slips. Psychological Review, 88, 1-15.
PEw, R. W. (1974). Human perceptual motor performance. In B. H. KANmwrTZ (Ed.),
Human information processing: Tutorials in performance and cognition. Hillsdale, NJ:
Erlbaum.
PILLSBURY, W. B. (1908). Attention. London: Swan Sonnenschein.
PosNER, M. I., & KEELE,' S. W. (1969). Attention demands of movement. In Proceedings of

the 16th International Congress of Applied Psychology. Amsterdam: Swets and Zeitlinger.
PosNER, M. I. (1978). Chronometric explorations of mind. Hillsdale, NJ: Erlbaum.
REASON, J. T. (1979). Actions not as planned. In G. UNDERWOOD& R. STEVENS (Eds.), Aspects
of consciousness. London: Academic Press.
REASON, J. T. (1983). Lapses of attention. In R. PARASURAMAN, R. DAVIES, & J. BEATTY (Eds.),
Varieties of Attention. New York: Academic Press.
REASON, J. T., & MYCIELSKA, K. (1982). Absentminded? The psychology of mental lapses and
everyday errors. Englewood Cliffs, NJ: Prentice-Hall.
RoBBINS, T. W., & SAHAKIAN, B. (1983). Behavioral effects of psychomotor drugs: Clinical
and neuropsychological implications. In I. CREESE (Ed.) Stimulants: Neurochemical, be-
havioral and clinical perspectives. New York: Raven Press.
RuMELHART, D. E., & NoRMAN, D. A. (1982). Simulating a skilled typist: A study of skilled
cognitive-motor performance. Cognitive Science, 6, 1-36.
RYLANDER, G. (1939). Personality changes after operations on the frontal lobes. Acta Psy-
chiatrica Neurologica Scandinavica, Supplement no. 30.
ScHMIDT, R. A. (1975). A schema theory of discrete motor skill learning. Psychological Review,
82, 225-260.
SHALLICE, T. (1982). Specific impairments of planning. Philosophical Transactions of the Royal
Society of London, Series B, 298, 199-209.
SHALLICE, T., McLEoD, P. D., & LEWIS, K. (1985). Isolating cognitive modules with the dual
task paradigm: Are speech perception and production separate processes? Quarterly
Journal of Experimental Psychology (in press).
SHJFFRIN, R. M., & ScHNEIDER, W. (1977). Controlled and automatic human information
processing: II. Perceptual learning, automatic attending, and a general theory. Psy-
SPELKE, E., HIRST, W., & NEJSSER, U. (1976). Skills of divided attention. Cognition, 4, 205-
230.
TREISMAN,. A. M. (1960). Contextual cues in selective listening. Quarterly Journal of Exper-
imental Psychology, 12, 242-248.
WALSH, K. W. (1978). Neuropsychology: A clinical approach. Edinburgh: Churchill Livingstone.
2 The Motor Theory of
Voluntary Thinking
BARRY H. COHEN
The theory proposed in this chapter-the motor theory of voluntary

thinking (MTVT)-is not exactly new; its basic elements can be found
in the motor theories of Bain (1888), Maudsley (1889), Ribot (1889), Pills-
bury (1908), and Washburn (1916) to name just a few. Many of the ideas
expressed below were commonplace at the beginning of this century,
and the interested reader is encouraged to explore the original sources
referenced within this chapter, which frequently contain detailed, an-
ecdotal descriptions and lengthy logical arguments in support of these
ideas. To maintain the clarity of my exposition, I will quote only a few
examples from the early literature; it is not my intention to present a
history of the theoretical formulations concerning the relation between
the motor system and attention or thought (a brief history can be found
in Smith, 1969). Rather, my purpose in writing the present chapter is
to present some of these old ideas in the context of a cohesive theory
that is as clear, plausible, and useful as possible.
It is my opinion that the current lack of popularity of motor theories
derives from the fact that these theories often contained unnecessarily
restrictive or implausible provisions and were often tied to physiological
notions based on the limited knowledge of their times. However, I feel
that to reject these theories in their entirety is to disregard a host of
potentially useful ideas, which appeal to one's common sense and may
stand in close agreement with one's own introspections. The MTVT, as
expressed herein, is far from complete and should certainly be consid-
ered as work in progress. Nonetheless, I feel it is useful to present the
theory at this time, especially considering the current increased interest
in facial muscle action (Ekman, Friesen, & Ellsworth, 1972; Fridlund &
Izard, 1984), in order to encourage the notion that overt or covert
BARRY H. COHEN e Department of Psychology New York University, The development

of this paper and the original research were supported, in part, by NIMH pre-doctoral
fellowship F31-MH07303-02, and post-doctoral fellowship F32-MH08889-01.
19
20 BARRY H. CoHEN
changes in skeletal muscle activity may be more than just indicators of

mental processes-these motoric activities may play an important me-
diational role.
I. PREVIous MoToR THEORIES
Rather than discuss the details of the various motor theories pro-
posed by particular psychologists, I will separate these theories into
broad categories and then discuss issues common to the theories within
each category.
A. The Motor Theory of Consciousness

In its simplest form, the motor theory of consciousness states that
no perception, no image, no thought-in fact, no mental experience-
is possible without the occurrence of some unique pattern of activation
in the motor system; it is the particular motor activity pattern evoked
directly by some external stimulus or associative process that determines
which mental experience will occur.
The arguments in favor of the motor theory of consciousness can
be quite convincing when applied to aspects of perception in which
motor activity plays a critical role, such as discerning the shape of an
object by scanning its boundaries visually and thus making appropriate
eye movements (Sperry, 1952; Weimer, 1977). Accounting for mental
imagery then becomes a simple matter: reproducing the pattern of
motor activity that has been associated with a particular perception, even
if covertly or only centrally, can, in the absence of the appropriate ex-
ternal object, evoke a pattern of cerebral activity similar to the cerebral
activity accompanying the perception and thus produce a "faint" form
of perception perceived as an image (in this case, a visual image). How-
ever, when one considers perceptual distinctions in which motor activity
plays no obvious role, the motor theory of consciousness is far less con-
vincing. For example, the perceptual quality that results from staring at
a red wall is quite different from that which results from staring at a
green wall, yet it is not obvious how the patterns of motor activity as-
sociated with each perception would consistently differ. Of course, if
the motor theory of consciousness cannot explain the perceptual dif-
ferences among colors, it also cannot explain the mental imagery of
colors. In a relatively recent paper arguing in favor of a motor theory
of consciousness, Weimer (1977) deals with the problem of sensory qual-
ities in such a general manner as to offer no concrete explanation at all.
THE MoToR THEORY OF VoLUNTARY THINKING 21
It is undoubtedly such weaknesses that explain the current unpopularity

of the motor theory of consciousness. The motor theory of thinking is
less encompassing than the motor theory of consciousness but may con-
tain similar weaknesses, as will be discussed in the following section.
At this point it is appropriate to discuss an important issue that
affects all types of motor theories: the necessity of peripheral motor
feedback, that is, afferent feedback from the contraction of muscles.
Theories that require peripheral feedback are susceptible to disconfir-
mation from studies in which the skeletal musculature of a human vol-
unteer was paralyzed with curare (e.g., Leuba, Birch, & Appleton, 1968;
Smith, Brown, Toman, & Goodman, 1947). In the Leuba et al. study,
the subject was able to solve complex mental problems just as well when
paralyzed as when not. Although McGuigan (1978) has amply criticized
the curare studies (e.g., most curare studies have not used electrical
recordings to establish that total paralysis had occurred), these studies
have cast a great deal of suspicion on motor theories in general.
Actually, most proponents of motor theories have suggested that
although peripheral feedback may be necessary in the initial stages of
internalizing thought, such feedback becomes unnecessary when re-
peated associations result in "short-circuit" connections within the cen-
tral nervous system. Dunlap (1927) suggested that this short-circuiting
might occur "between cerebellum and cerebrum, with no muscular ac-
tivity ... involved" (p. 265). Although it is not likely that the cerebellum
plays an indispensable role in cognition, the type of speculation offered
by Dunlap does not seem entirely implausible in the light of more recent
research on central efferent monitoring and corollary discharge (Evarts,
1971). Since it is unlikely that peripheral feedback is necessary for many
types of overt motor acts, it seems even less likely that peripheral feed-
back would be necessary for thinking. A good example of the extreme
central position is contained in the motor theory of consciousness pro-
posed by Sperry (1952): "The core of the perceptual process is not itself
a motor pattern. It is more pre-motor or better pre-premotor in nature
owing to the hierarchical plan of neural organization" (p. 309).
Motor theories that are formulated exclusively in terms of central
motor feedback are difficult to disprove experimentally and are there-
fore sometimes viewed with suspicion. But although central motor the-
ories may not be easily disconfirmed, they can represent useful ap-
proaches to organizing our knowledge of cognitive processes and can
suggest valuable experiments. The MTVT, described in subsequent sec-
tions, adopts this viewpoint in assuming that only central motor feed-
back is necessary when motor feedback is necessary at all, and that actual
muscle contraction is never necessary for the motor system to influence
mental experience.
22 BARRY H. CoHEN
B. The Motor Theory of Thinking

The motor theory of thinking, in its simplest conception, is just a
subset of the motor theory of consciousness. The motor theory of think-
ing does not propose that motor activity is required for all mental ex-
periences, but rather that motor activity is required for mental experience
in the absence of external sensation-that is, for images and thoughts.
From a commonsense point of view, the motor theory of thinking seems
much more plausible than the motor theory of consciousness. Most per-
ceptions do not seem to require motor activity necessarily, and certainly
not specific and unique patterns of motor activity. On the other hand,
the occurrence of mental images can seem somewhat mysterious-they
are like perceptions in some respects, yet they have no obvious cause.
The motor theory of thinking provides a reasonable explanation for the
generation of mental images.
According to Washburn (1916), kinesthetic images are simply faint
kinesthetic sensations: kinesthetic "images" are generated through
slight, covert contraction of the appropriate muscles-the afferent feed-
back from these slight contractions is perceived as an image rather than
a perception. It is quite reasonable to assume that one could make such
slight contractions without being at all aware of making them and then
regard the resulting faint kinesthetic sensation as a purely mental ex-
perience. That subjects do, indeed, produce slight muscle contractions
when imagining muscular acts was confirmed by the early work of Ja-
cobson (1932).
Jacobson pioneered the use of electromyography (EMG; the am-
plification and display of the tiny changes in electrical voltage associated
with muscle contraction) to demonstrate that when a subject imagined
bending one arm, slight contractions occurred in the biceps of that arm,
but not the other arm. Moreover, the imaging of rhythmic activities,
such as pumping a tire, were associated with a corresponding rhythm
of slight muscle contraction and relaxation. Shaw (1938) partially rep-
licated these results by exploring a wide range of motoric activities.
Washburn's explanation of kinesthetic imagery cannot, of course,
be applied to other forms of imagery. Kinesthesia is the only modality
in which sensations can be produced directly by voluntary action. Im-
ages in other modalities may be explained, however, through their as-
sociation with kinesthetic sensations. The case of speech imagery (au-
ditory images of one's own voice) or "inner speech" is by far the simplest
to explain. Associations between one's own voice and kinesthetic sen-
sations from one's speech musculature are very specific, consistent, and
frequently repeated. It is a reasonable assumption that eventually the
THE MoToR THEORY oF VoLUNTARY THINKING 23
kinesthetic sensations produced by slight contractions of the speech

musculature can evoke auditory images of one's own voice, even when
one is not aware of any kinesthetic sensations or images. As in the case
of kinesthetic imagery, there is electromyographic evidence that covert
muscle contractions occur in the appropriate area (i.e., the speech mus-
culature) when subjects are asked to imagine speaking; in fact, the evi-
dence is quite extensive (see McGuigan, 1978, for a comprehensive re-
view). The specificity of this covert speech activity was convincingly
demonstrated in an experiment by McGuigan and Winstead (1974), in
which the EMG was recorded from both the tongue and the lips while
subjects read and mentally rehearsed words containing either labial (re-
quiring the lips for pronounciation) or lingual (requiring the tongue for
pronounciation) phonemes. As expected, the lip EMG was significantly
higher than the tongue EMG when subjects mentally processed the labial
words, whereas the reverse was true for lingual words.
The whole notion that kinesthetic feedback can, after frequent as-
sociation, evoke a particular auditory image is rendered plausible by an
experiment conducted by Hefferline and Perera (1963). In this
experiment:
When the subject occasionally emitted an invisibly small thumb twitch (de-
tected electromyographically), he received a tone as a signal to press a key.
After several conditioning sessions, the tone was progressively diminished
to zero. The subject nevertheless continued to press the key whenever he
emitted a thumb twitch, and he reported that he still heard the tone. (p. 835)
It is visual imagery that is the most difficult to explain in terms of

motor activity or kinesthetic sensations, especially the imagery of dif-
ferent colors. Only in the case of visual objects that are extremely elon-
gated in one direction or are moving does the possibility of a consistent
association between eye movements and visual perception become re-
alistic. In fact, in an early study by Jacobson (1930), eye movements (as
indicated by the electrooculogram, or EOG) in the appropriate direction
were found to accompany the visual imagination of elongated objects,
such as the Eiffel Tower. This finding was confirmed by Totten (1935),
who photographed a spot of light reflected off the cornea of her subjects.
In the great majority of cases in which subjects formed a visual image
of elongated objects, eye movements in the appropriate direction were
clearly observable. Instances in which eye movements were not ob-
served generally involved visual images which were "distant" or small.
More recent studies have found an association between eye movements
and visual images involving motion; these studies are discussed in a
later section.
Hebb's (1968) theoretical notions would predict eye movements also
24 BARRY H. CoHEN
in the case of forming a visual image of an object with well-defined

boundaries. Hebb wrote:
It is easy to form a clear image of a triangle or a circle when eye movement
is made freely (not necessarily following the contours of the imagined figure),
harder to do with fixation of gaze while imagining the eye movement, but
impossible if one attempts to imagine the figure as being seen with fixation
on one point. Though such informal evidence cannot carry great weight, it
does agree with the idea that the motor accompaniments of imagery are not
adventitious but essential. (p. 470)
It is not clear whether Hebb would expect eye movements to accompany

an amorphous visual.image, such as one of an empty blue sky. Hebb
also leaves unclear the mechanism involved in imagining eye move-
ments and the conditions under which an individual would be more
likely to imagine making eye movements to form a visual image than
actually to make them.
It is possible, as I have speculated previously (Cohen, 1978), that
the visual imagery of static objects is associated with a characteristic
pattern of muscle tension ordinarily associated with focused visual at-
tention (e.g., squinting, including tension in the brows and around the
eyes). But it seems unlikely that the motor activity accompanying the
visual image of an orange on a table could be distinguished from the
motor activity accompanying the visual image of an apple on a table.
Therefore, it would appear that any motor theory of thinking which
insists that visual images cannot occur without some unique pattern of
motor activity or that the content of the visual images is completely
determined by the accompanying motor activity must remain uncon-
vincing. The motor theory of voluntary thinking, which I describe below,
explicitly avoids this weakness.
C. The Motor Theory of Attention
The motor theory of attention does not assert that motor activity is
necessary for the creating of mental experience but that motor activity
plays a critical role in emphasizing one aspect of perception or thought
over another. Among the earliest expressions of this theory was the
view of Bain (1888) that "in selecting a quality out of a complex effect;
in maintaining the attention upon one of several images that rise to the
view; in a word, in all voluntary control of the thinking trains,-there
is a muscular intervention" (p. 373). A distinction is invariably made
between spontaneous attention and voluntary attention; the former is
drawn automatically by its object with no concomitant sense of effort,
whereas the latter is more abstractly motivated and requires the constant
application of effort and will. It is generally hypothesized that only vol-

untary acts of attention require activation of the motor system.
The motor theory of attention does not require specific patterns of
motor activation to correspond to particular perceptions, images, or
thoughts and thus avoids a major weakness of the motor theories de-
scribed in the previous two sections. However, the motor theory of
attention rests on there being some general relation between motor ac-
tivity and the direction of attention. Pillsbury (1908) delineates three
classes of motor activity related to attention: movements that adapt the
sense organs for keener perception; general movements which depend
on the nature of the stimulus; and "general overflow effects upon the
voluntary muscles which do not depend upon the nature of the stim-
ulus" (p. 12). James (1890/1950) describes motor activation of the first
type as follows:
In attending to either an idea or a sensation belonging to a particular sense-
sphere, the movement is the adjustment of the sense-organ, felt as it occurs.
I cannot think in visual terms, for example, without feeling a fluctuating play
of pressures, convergences, divergences, and accommodations in my eye-
balls. The direction in which the object is conceived to lie determines the
character of these movements. (p. 300)
Fechner described a difference between attention to external objects and
attention to mental images; as quoted by James (1890/1950), Fechner
stated:
I have, when I try to vividly recall a picture of memory or fancy, a feeling
perfectly analogous to that which I experience when I seek to apprehend a
thing keenly by eye or ear; and this analogous feeling is very differently
localized. While in sharpest possible attention to real objects ... the strain
is plainly forwards, ... the case is different in memory or fancy, for here
the feeling withdraws entirely from the external sense-organs, and seems
rather to take refuge in that part of the head which the brain fills; if I wish,
for example, to recall a place or person it will arise before me with vividness,
not according as I strain my attention forwards, but rather in proportion as
I, so to speak, retract it backwards. (p. 435-6)
As an example of the second class of attention-related motor activ-
ities, Pillsbury refers to those subtle, automatic movements and changes
in body posture which can allow a skilled "mind-reader" (or "muscle-
reader," as Pillsbury contended) to discern the direction of an object
upon which a person is concentrating. Generally, the person concen-
trating is unaware of performing these tell-tale motor activities, and the
"mind-reader" may even be unaware of the source of his information.
The third class of motor activity is associated with all forms of at-
tentional effort, regardless of its direction, and may account in large
measure for the. experience of effort. Pillsbury mentions the "wrinkled
brow" as associated with mental effort. James associated tension in the
26 BARRY H. CoHEN
brow and glottis to feelings of approval and disapproval but added: "in
effort of any sort, contractions of the jaw muscles and of those of res-
piration are added to those of the brow and glottis" (p. 301, original
emphasis). Ribot (1889), in describing the muscular reactions accom-
panying intense reflections, concluded that, in addition to more specific
reactions, "in all persons and in every case there are modifications in
the respiratory rhythm" (p. 27).
The theory I describe in the following section-the MTVT-com-
bines elements from both the motor theory of thinking and the motor
theory of attention. Many of the hypotheses I propose occurred to me
independently of the works of the early psychologists cited above. How-
ever, the elegant writings of these pioneer investigators are helpful in
providing rich examples of many of the phenomena to which I refer.
II. DEFINING THE MoToR THEORY oF VoLUNTARY THINKING
The motor theory of voluntary thinking (MTVT) does not propose

that a specific pattern of motor activity is necessary for the occurrence
of each and every mental image or thought. According to the MTVT,
any thought or image may occur as an association to a sensation or
another thought or image. It seems reasonable that the cerebral pro-
cesses involved in the experience of one mental image can stimulate the
cerebral processes underlying another image without the mediation of
any form of motor activity. However, the feedback from motor activity
can also evoke thoughts or images according to principles of association.
These motor associations can be highly specific, as in the case of inner
speech, where the covert reproduction of a specific speech motor pattern
can evoke the auditory image of a particular phoneme; or the associa-
tions can be much less specific, as will be discussed in a subsequent
section. Furthermore, the MTVT does not merely suggest that motor
feedback can evoke mental images and thoughts; it ascribes to the motor
system a very important role in the ordinary thinking process. Patterns
of covert motor activity can guide the direction of thought, sometimes
very specifically, through the associations they generate. Moreover, it
is proposed that motor activity is responsible for the experience of will
or volition in thinking-for mental effort-as well as other important
aspects of thought to be detailed below.
The basic premise of the MTVT is that all acts of will, or effort,
involve activation of the motor system; that central motor activity is the
only cerebral activity under direct voluntary control; and that the ex-
perience of voluntariness arises only from motor activity, even if only
central and not peripherally expressed. Thus, the MTVT parsimoniously
explains that the experience of volition that accompanies certain mental

acts arises from the same source as the experience of volition accom-
panying physical acts.
A. The Experience of Volition

The MTVT explains quite simply why some mental images or
thoughts are experienced as voluntary (e.g., rehearsing an unfamiliar
phone number) and others not (e. g., suddenly experiencing visual mem-
ories upon perceiving a particular fragrance); the former images involve
motor activation, whereas the latter do not. To be specific, rehearsing
a phone number requires a speech motor pattern to be repeated covertly.
However, memories can also occur spontaneously through association
with other cerebral processes, which may arise in the form of perceptions
or may themselves remain out of awareness.
It should be noted that the MTVT does not assert that motor acti-
vation must always lead to an experience of voluntariness. Even overt
motor acts may become so habitual that they can be performed without
any awareness and therefore with no sense of effort. It is possible in
the case of some thought patterns that although they arise through as-
sociation with motor feedback, the covert motor acts with which they
are associated are performed automatically-that is, with no sense of
effort. For instance, the persistent repetition of the auditory image of a
popular song may be the result of a habitual covert motor pattern of
which one is unaware, just as one may be unaware of habitually straight-
ening an article of clothing or walking in the wrong direction while
engrossed in conversation. Thus a thought may appear to be effortless
because no motor activation is involved, or because the motor activity
is of an automatic nature.
Whether a covert motor act associated with mental activity is per-
ceived as voluntary or not would be determined by the same principles
which determine the degree of voluntariness experienced during overt
motor acts. In the preceding chapter, Norman and Shallice describe a
theoretical motor mechanism, the "Supervisory Attentional System,"
the output of which is strongest when motor acts recruit the greatest
exercise of attention, as when the motor act is particularly difficult, dan-
gerous, novel, or unpleasant. Norman and Shallice then propose "that
will be this direction of action by deliberate conscious control" and that
"will corresponds to the output of the Supervisory Attentional System."
The theory they propose is concerned primarily with external actions,
but, as they say, "the same principles apply to internal actions-those
that involve only the cognitive processing mechanisms." It is also worth
28 BARRY H. CoHEN
noting that the concept of the Supervisory Attentional System is bols-

tered by the fact, pointed out by Normal and Shallice, that the functions
proposed for this mechanism "correspond closely with those ascribed
by Luria (1966) to prefrontal regions of the brain, thought by Luria to
be required for the programming, regulation, and verification of
activity."
B. Specific Motor Associations

The chief assertion of the MTVT is that without the motor system
there could be sensation, perception, and even mental images-but no
voluntary thought. The critical aspect of this theory is the delineation
of the mechanisms by which covert motor activity can explain all aspects
of voluntary thinking. One important principle is that the more specific
the associations are between patterns of motor activity and the contents
of mental experience in a particular modality, the more closely mental
experience in that modality can be determined through voluntary motor
control. In the case of kinesthetic imagery and inner speech, the asso-
ciations are simple and direct, as described above in the motor theory
of thinking, and therefore these types of imagery should be easy to
control. To recall how it feels kinesthetically to row a boat, one need
only to activate covertly the appropriate pattern of motor activity. To
imagine how it might feel to perform some new motor activity, one
would only have to activate covertly the pattern of motor activity that
one believes would be necessary, just as one would activate the new
pattern overtly if trying the activity for the first time.
1. Inner Speech
In order to rehearse a telephone number one would simply "speak"

the numbers covertly-that is, activate the appropriate speech motor
patterns, but too slightly to produce audible speech. To take the case
of rehearsing a telephone number a step further, consider that the per-
son is being distracted by loud music. Because the music would be com-
peting for his attention, he would have to increase the amplitude of his
rehearsal by increasing the speech motor activity, perhaps to the point
of making actual lip and tongue movements. Were the music loud
enough he might have to speak the numbers aloud so that the numbers
would capture enough of his awareness to remain in his short-term
memory. McGuigan and Rodier (1968) did indeed find increased speech
motor activity (as measured by EMG) in subjects who read during au-
ditory distraction (e.g., white noise, "reverse" speech) as compared to

reading in silence.
The person rehearsing the telephone number might also be dis-
tracted by his own thoughts or images. For instance, he might be facing
an important deadline and find his awareness flooded with verbal
thoughts about activities which must be performed as soon as possible.
Although it is conceivable that these distracting thoughts could arise
spontaneously without accompanying motor activity, it seems more
likely that these distracting thoughts would themselves be based on
covert speech activity that occurs rapidly and automatically in response
to an awareness of the impending deadline. Increasing the speech motor
activation associated with the telephone number can probably diminish
the distraction of competing thoughts in two ways. First, the telephone
number captures an increasing amount of awareness (this would work
against any distractor, whether music or thought). Second, to the extent
that the distracting thoughts require activation of the speech motor sys-
tem in their own service, attempting to monopolize the speech system
with the telephone number allows less access to the distracting thoughts.
At this point it should be considered that thoughts are not likely to
require full speech motor patterns to reach awareness. Partial patterns
can probably be activated very rapidly, intertwining two or more
thoughts in such a way that they seem virtually simultaneous. However,
it also seems clear that the strong and rapid repetition of one particular
thought can greatly diminish the expression of competing thoughts.
Rehearsing a telephone number is just the type of thinking that is very
likely to require voluntary thinking in every case and, therefore, speech
motor activity, against an inevitable background of distracting thoughts
and images.
2. Visual Imagery
The case of voluntary visual imagery is less straightforward than

the case of inner speech, but its explanation presents no great problem
for the MTVT. Visual imagery represents the case in which the relation
between motor activity and mental content is not highly specific, yet
some degree of voluntary control is still possible. First, it must be made
clear that the MTVT does not posit a one-to-one correspondence be-
tween patterns of motor activity and the contents of visual images. Con-
sequently, it cannot be contended that specific visual images are evoked
by activating the appropriate motor pattern. However, even if visual
images cannot be directly evoked by motor activity, they can to some
extent be suppressed, enhanced, or moved by the appropriate motor
pattern. In the case of visual imagery, the appropriate motor activity
30 BARRY H. CoHEN
very likely involves the extraocular muscles, which control the move-
ments of the eye, and the musculature which surrounds the eye, in-
cluding the muscles which draw down the brow to aid in squinting.
Eye movements seem to play an important role in supplying motion
to visual images. Antrobus, Antrobus, and Singer (1964) found that sub-
jects exhibited significantly more eye movements when forming visual
images involving motion (e.g., a tennis match) than when forming static
images (e.g., an orange on a table). Several experiments have confirmed
that most subjects produce noticeable, rhythmic horizontal eye move-
ments when visually imagining a moving pendulum (Brown, 1968;
Deckert, 1964; Lenox, Lange, & Graham, 1970). Those subjects not pro-
ducing measurable eye movements may yet have been producing the
same central oculomotor activity, but at a level insufficient to move the
eye.
It should be noted at this point that the lack of useful proprioception
from the extraocular muscles, which has been convincingly demon-
strated (e.g., Matin, Pearce, Matin, & Kibler, 1966), does not weaken
this part of the MTVT. Central feedback from motor activity plays an
important role in visual perception and may play an important role in
visual imagery as well. To those who are skeptical of the involvement
of eye movements or oculomotor commands in the motion of visual
images, the following demonstration is recommended:
Imagine that you are standing on the roof of a tall building and
looking across the street at a building that is even twice as tall. On
the roof of the opposite building, just near the edge, is a colorful
beach ball (you must look upward to see it). At this point in the
demonstration, the reader must actually move his eyes upwards as
high as it is physically possible, without inducing pain (obviously,
the reader will have to read the demonstration through once before
actually trying it, unless it is being read to him). Then, while main-
taining your eyes in this upward position as firmly as possible,
imagine that a slight breeze has caused the beach ball to fall off the
edge of the opposing roof. Without allowing your eyes to move
down even slightly from their extreme upward position, imagine
the beach ball falling slowly downward. You must actually "see"
the beach ball as a visual image, as vividly as you can. The beach
ball begins its descent above you, but it is soon level with your
position, as it continues to fall. Remembering not to allow your eyes
to lower even slightly, continue to "see" the beach ball fall down-
wards until it bounces on the street below you.
When this demonstration was read to a group of more than 30
experimental psychologists and graduate students, the majority felt a
THE MoToR THEORY OF VoLuNTARY THINKING 31
strong tendency to move their eyes along with the visual image and felt
that the image was marred if they did not succumb to this tendency.
Many felt they simply could not "see" the beach ball fall without moving
their eyes. I devised this demonstration as a visual analog of a dem-
onstration described by James (189011950) and attributed to Stricker. In
the Stricker demonstration the subject must imagine hearing a word
such as bubble, while keeping his mouth partly open. Some people find
their auditory images disturbed, at least initially, because of the incom-
patibility between the feedback from a partially opened mouth and the
feedback that would occur in pronouncing the word.
If eye movements can be effective in "moving" a visual image, they
can, perhaps, play a role in suppressing unwanted visual images by
"moving" them to the periphery, or even past the boundaries of the
mental visual field. Consistent with this hypothesis is the finding by
Singer and Antrobus (1965) that eye movements were significantly
greater when a subject was trying to suppress rather than maintain an
image, whether the subject's eyes were open or covered with a trans-
lucent surface.
To increase the vividness of visual images, some individuals may
activate the motor pattern that produces squinting, as this is the motor
activity most closely associated with trying to enhance visual perception.
In my own research, I have obtained some evidence which suggests that
subjects who generate greater corrugator (brow) tension relative to lip
or arm tension experience more vivid visual imagery than subjects
whose brow tension is less prominent. After performing two different
cognitive tasks, all subjects rated the vividness of their visual imagery
and inner speech for each task. Only 7 out of the 22 subjects gave higher
ratings generally for their use of visual, as compared to speech, imagery.
These subjects were considered a separate "group" in an ANOVA which
included muscle area as a within-subject factor. The interaction between
imagery group and muscle area was on the borderline of significance,
F(1.63, 32.53) = 3.37, p < .06 (the degrees of freedom were adjusted
according to the method of Geisser and Greenhouse [1958] to account
for the lack of homogeneity of covariance). As can be seen in Figure 1,
the two groups do not differ in total muscle tension during cognitive
performance, but rather in the way that the tension is distributed.
So far I have described how motor activity may modify a visual
image that already exists in awareness. However, there are times when
one voluntarily produces visual images in order, for example, to recall
a previous vacation or to imagine a future one. If, according to the
MTVT, visual images do not directly correspond to patterns of motor
activity, it must be explained how they can be voluntarily evoked. The
explanation is that they are evoked only indirectly by association, but
32 BARRY H. CoHEN
Visual Imagery Use Inner Speech Use

rated higher rated higher
Vi
-~
c
::;)
~56
co
~ 48
;;;
t; 40
~ 32
c
co
a.> 24
:2:
16
FIGURE 1. Mean EMG amplitude in left and right areas of the brow, lips, and forearms
during cognitive performance for two subgroups of subjects differing in reported imagery
use (see text).
that some of these associations are under direct, voluntary control. The
strong association between words and visual images should be apparent
to anyone who has read a novel, especially one in which visual aspects
of the setting played an important role. If one closes one's eyes and
thinks the word elephant, it is difficult not to get some visual impression
of a bulky animal with a trunk, etc. To voluntarily produce a visual
image, then, one need only name the object in question or describe it
with inner speech (in fact, there are extreme visualizers who cannot
think any word, even an abstract one, without immediately experiencing
some kind of image), which in turn requires the activation of speech
motor patterns. Some individuals must close their eyes to experience a
clear image, whereas others tend also to squint. Still others claim that
they never experience clear visual images. The MTVT does not attempt
to explain these differences in visual imagery ability; the theory simply
asserts that all visual images are evoked by association (not necessarily
to inner speech; perhaps kinesthetic images or sensations can evoke
visual images, for instance), more easily for some than for others.
It should be noted that the MTVT does not propose that eye move-
ments are necessary to form vivid, static images or nonelongated images.
In fact, it may be helpful to suppress large eye movements in order to
minimize disruptive movement of the image. This principle may explain
the results of Ehrlichman and Barrett (1983) who recorded fewer eye
movements when subjects solved spatial problems (presumably using
some visual imagery) than when they solved verbal problems, whether
in a dark or lighted environment. These investigators were somewhat
puzzled by their results because they expected imagery disruption to
be a factor only in a lighted environment due to movements of the
external visual field; they did not consider movements of the images per
se. It should be mentioned that Stay (1930), who used very different
spatial problems, obtained the opposite results. Whether imagery move-
ment helps or hinders the solving of a spatial problem may depend on
the details of the particular problem.
An experiment by Marks (1973) suggests yet another role for eye
movements during visual imagery: scanning for detail. In Marks' study,
two groups of subjects were selected on the basis of the subjective ratings
they had given to the vividness and clarity of visual images called for
by a questionnaire. Eye movement rates were measured for both the
"vivid" and "poor" visualizers while they viewed complex visual stimuli
and while they attempted to answer questions concerning the visual
details of what they had seen several seconds before. Marks found that
the vivid visualizers not only recalled more visual details but also ex-
hibited fewer eye movements during recall than did the poor visualizers.
This result might be difficult to explain for a theory that requires eye
movements for all visual imagery, but it is consistent with the MTVT.
The vivid visualizer is such because visual images appear to him more
quickly and completely when he thinks about the image. The poor vis-
ualizer, in an attempt to answer questions about details he could not
easily visualize, may have needed to scan repeatedly his vague and
incomplete image. The vivid visualizer, on the other hand, had all the
details he needed at once before his mind's eye.
Cuthbert and Lang (1984) obtained a result seemingly inconsistent
with the findings of Marks. They found that subjects whose imagery was
below average on vividness as reported on a questionnaire were less
likely than the remaining subjects to exhibit eye movements during recall
of a moving visual display. However, both results are consistent with
a theory which posits specific functions for eye movements during visual
imagery but does not contend that eye movements are necessary to
create a vivid visual image. Subjects with a vivid image in the Cuthbert
and Lang study could be expected to use eye movements to set the image
in motion (the MTVT posits that even the best visualizer cannot vol-
untarily visualize objects in motion without producing eye movements),
whereas subjects with a poor image would not be able to produce a
noticeable effect on their image with eye movements. Quite likely it is
hard to notice a clear pattern of motion when one's image is hazy and
34 BARRY H. COHEN
unstable to start with. Because changes in visual imagery due to eye

movements would not be clear, such movements would not be rein-
forced. Cuthbert and Lang's result replicates a similar finding by Brown
(1968) who used a moving pendulum. In summary, it is suggested that
subjects with vivid visual imagery, as compared to those reporting poor
imagery, would be more likely to produce eye movements when imaging
a moving object but would produce fewer eye movements in searching
for details because the details would be more easily and quickly noticed.
The notion that visual images can be controlled only indirectly is
not a weakness of the MTVT. On the contrary, this principle can explain
some important differences between visual imagery and inner speech.
First, it should be noted that nearly all undergraduate respondents to
a questionnaire reported clear experiences of inner speech, whereas ex-
perience with visual imagery varied widely (Cohen & Jonas, 1978). Sec-
ond, for most individuals, visual imagery is harder to control than inner
speech; it is relatively easy to compose a novel sentence, but for some
individuals it is quite difficult to combine diverse visual elements into
a novel visual image. Third, when associations occur spontaneously as
during daydreams or night dreams, visual images are likely to exhibit
combinations that are more bizarre and less logically structured than
verbal imagery; the visual images are combining according only to the
laws of association, whereas verbal images tend to retain the structure
of language.
3. Other Modalities
Other modalities of imagery (e.g., smells, tastes) can be accounted

for in a manner similar to the explanation of visual imagery. To form
the image of a taste, one might think the appropriate words, while cov-
ertly activating tongue or jaw movements. For an olfactory image, some
form of covert sniffing would likely accompany an internal verbal de-
scription. Auditory images other than the sound of one's own voice
represent an interesting case that has been surprisingly neglected. From
my own introspection, I find that to imagine the sound of a musical
instrument I subvocally mimic the sound of the instrument and "sing"
or "hum" the tune mentally, while my memory fleshes out the sound.
As in the case of visual imagery, these other modalities are under less
direct control than inner speech and rely principally on a combination
of inner speech and the appropriate sensory memories.
C. General Motor Associations

In addition to evoking specific thoughts and images, and their im-
mediate associations, covert motor activity may play a more general role
in the deployment of attention. There is a great deal of research dem-

onstrating that experimentally induced muscle tension can, up to some
optimal tension level, improve performance on a variety of mental tasks
(see Cohen, 1983, Appendix B, for a recent review). There is also sugges-
tive evidence that the mental performance increments associated with
induced tension are due to the increased cortical arousal produced (e.g.,
Pinneo, 1961). That proprioceptive stimuli can be highly effective in
producing cortical arousal was shown directly in cats and monkeys by
Bernhaut, Gellhorn, and Rasmussen (1953). In addition, Gellhorn (1958)
found the reverse effects in cats; muscle relaxation produced by curare
paralysis was accompanied by a marked reduction in cortical arousal
and responsiveness to painful stimulation. Thus it may be supposed
that individuals increase the tension in their muscles globally, or ac-
cording to idiosyncratic patterns, in order to increase their level of arousal
and thus improve their mental performance. This response, however,
can become maladaptive if the individual is already aroused above his
optimal level for the mental task in question.
Covert motor activity may also serve in a general way to focus at-
tention. There may be some motor patterns that are activated whenever
the individual expends mental effort, regardless of the type of imagery
or thought engaged in, or the nature of its content. Each individual may
have his own idiosyncratic tension pattern that he activates whenever
he concentrates (i.e., focuses his attention). However, some motor ele-
ments may have nearly universal association with mental concentration.
In particular, brow contraction appears to be closely related to the pro-
cess of focusing attention.
Rather than being mere by-products of mental effort, these general
tension patterns may serve an important short-term mnemonic function
that aids the individual in maintaining the direction of his attention. If
a certain motor pattern has often been associated with mental effort in
a particular individual, feedback from that motor activity could serve to
remind the individual that he is presently involved in some attentive
act from which he has become distracted. Although the feedback would
not tell the individual where his attention should be, it would cause
him to search his immediate memory. This mechanism could act as a
mental "string around the finger" in the following manner. Suppose an
individual is concentrating on some mental problem while tensing his
muscles in a characteristic pattern. His attention wanders off the prob-
lem, but within seconds he becomes aware of motor feedback from his
"concentration" activity. He then quickly searches his immediate mem-
ory to recall the particular problem he had been working on. This shifting
of attention back and forth could occur very quickly and frequently,
especially if the problem were not intrinsically interesting.
I
.••... I
··········:
•·... \; ./
·.........·. :···~ . .
·······~ ..... ..··.. ~
.....
··.. ~
~
... ·····:.y
....
"
---o~----~or-----~or-----~o------,o------o.-----~oT------o~----~ [
~ ~ N 0 00 ~ ~ N
160
140
~-.
....
·c:E 120
:J : ··.
~ : ·....··
~
:0 100
~
(!)
,;········· ...:...
~.
I •
::;;: 80
.
,w
l!l ...t
"'0> 60 ....r·../ :'i t
l!l :~_,.
E
.<J-~ ~-/~ J,
I
,I::
.
•'
40 . ~
' . .
·.
~ ~-
20 .,,
····/···· \............ ·...·.• ..··... ... .·
.· ...·,_..--·.:..··
PR A E A AEATPRA A A AT PA A A E A E A RT PR A E A E A E A E A RT
PR - please rest
left arm
A - add a symbol lor letter! '***** right arm
E - exch&nge letters
left brow
A rotate the board ==-=== right brow
AT please retain = left lip
- - - - right lip
FIGURE 2. Integrated EMG from left and right areas of the brow, lips, and forearms for one subject during four trials of a mental rotation task
followed, after a brief rest break, by four trials of a letter transformation task.
38 BARRY H. CoHEN
It could be suggested that the individual would soon adapt to the

constant feedback from his concentration activity, which would then be
rendered ineffective. One solution to this problem of adaptation would
be for the individual to increase continuously the intensity of his con-
centration-aetivity during the period of focused attention. This increas-
ing motor activity could be expected to produce a rising tension gradient
in at least some of the individual's muscles. Such tension gradients have
indeed been found during experiments involving continuous attention
(Bartoshuk, 1956; Wallerstein, 1954). In my own research (Cohen, 1983),
I observed that for some subjects brow tension did not decrease appre-
ciably within or between trials but would rise higher with each suc-
ceeding trial until a rest period was given (after each block of four trials).
These gradients were much less pronounced and considerably more
variable in the lip and forearm areas. Trials consisted of a series of aurally
presented instructions describing progressively more difficult mental
rotations or transformations of a string of letters. These instructions were
cumulative within a trial and the subject did not respond until the end
of a trial. Figure 2 presents data for two blocks of trials for one subject
who exhibited prominent brow tension gradients.
III. THE RoLE oF CovERT MoToR AcTIVITY IN ExTERNAL

ATTENTION
A. Voluntary Attention
Mental effort accompanies not only the focusing of attention upon
purely mental experiences such as images or thoughts; a sense of mental
effort can also be experienced in attending to one of several competing
external stimuli. The classic example is the "cocktail party phenome-
non", in which the listener is bombarded with several simultaneous
conversations and must attend to one of them, at least long enough to
understand its meaning. Moreover, the conversation selected need not
be the loudest. An explanation consistent with the MTVT is that the
listener would "shadow" the selected conversation (i.e., repeat it cov-
ertly) to keep his attention oriented toward it. However, covert motor
activity may aid in selective attention by an additional mechanism in
this case, and this additional mechanism may account, in part, for the
sense of effort and volition involved in selectively attending to one of
many external stimuli. This additional mechanism consists of producing
a pattern of covert motor activity to represent symbolically one of the
competing inputs. For example, if one conversation originates from the
THE MoTOR THEORY oF VoLUNTARY THINKING 39
listener's left and the other from his right, to select the leftmost con-
versation the listener may produce a bodily posture subtly oriented to-
ward the left. This posture may be expressed solely in tensional ad-
justments without producing any actual body movements. In the case
of several conversations originating from diverse locations, the tensional
adjustments would likely be less specific and less help£ttl, but none-
theless could aid in the listener's spatial orientation. This mechanism
could be applied analogously for other perceptual modalities. The MTVT
asserts that the experience of effort that sometimes arises during external
attention is due entirely to motor activation, whether in the form of
postural adjustments, sensory organ adjustments, or inner speech. Sym-
bolic postural adjustments may also play a role in directing attention to
one thought or image as opposed to another, but in the case of internal
attention, spatial orientation is less likely to be a useful factor, and inner
speech probably assumes the major burden in determining direction.
B. Spontaneous Attention
James (1980/1950) described two forms of "sensorial.attention": one
that is "active and voluntary" (as described in the previous section) and
one that is "passive, reflex, non-voluntary, effortless" (p. 416). The latter
form of attention, which I prefer to call spontaneous, may be immediate,
drawn to stimuli by virtue of their force or intrinsic interest, or derived,
that is, drawn to stimuli by virtue of their association with objects of
immediate attention. In psychophysiological research, the type of at-
tention described by James as "passive immediate sensorial" has been
evoked experimentally by visual stimuli depicting, for instance, attrac-
tive nudes or repulsive accident scenes; passive derived sensorial at-
tention has been evoked by presenting warning signals for trials in-
volving aversive events, or fast reaction. Both types of passive sensorial
attention have been associated with significant decreases in heart rate,
usually associated with lowered arousal; other arousal indicators si-
multaneously changed in the direction of heightened arousal (e.g.,
Libby, Lacey, & Lacey, 1973; Obrist, Webb, & Sutterer, 1969). This par-
adoxical divergence of arousal indicators has been called directional frac-
tionation (Lacey, 1967).
On the basis of a series of experiments which demonstrated that
heart rate reductions to warning stimuli were accompanied by reduc-
tions in various aspects of somatic activity (e.g., chin tension, respira-
tion, eyeblinks), Obrist, Webb, Sutterer, and Howard (1970) proposed
that both the cardiac and somatic activity reductions are due to a global
inhibition of bodily activity in order to reduce neural noise which might
40 BARRY H. CoHEN
otherwise compete with the perception of the external stimulus. Thus,

it appears that spontaneous external attention may involve increased
activation in specific cortical areas (which may correlate with observed
increases in, for example, electrodermal activity), coupled with an active
inhibition of somatic activity. This phenomenon is easily understood
within the context of the MTVT. Ordinarily, the perception of external
stimuli would stimulate associations, which would lead to a series of
covert motor responses and their accompanying mental images, as this
is how normal thinking is posited to proceed. However, if it is important
to maintain attention to the environment, it would be best to minimize
thinking, and motor inhibition would serve this purpose (in addition to
reducing the distraction of extraneous bodily activity per se). Because
mental effort is not required for this type of attention-both perception
and cortical activation are stimulated directly by the environmental and
cognitive situation-motor activity is not required (and can actually be
a hindrance). However, I suspect that in spontaneous attention, al-
though there is a global reduction in bodily activity, including muscle
tension, there may be increased activity in certain muscle areas, espe-
cially the brow. Tension in selected regions, not sufficient to offset the
global reduction in bodily activity, may be activated automatically to aid
in the focusing of attention. Considering the nature of spontaneous at-
tention, any muscle tension that accompanies this state would probably
remain in a fixed pattern and at an unusually steady level.
The two types of attention, described in the previous two sections
as voluntary and spontaneous, correspond to some degree with the
processes of controlled search and automatic detection, respectively, as
postulated by Shiffrin and Schneider (1977). The former requires mental
effort, often in the form of covert verbal rehearsal, and should therefore
be accompanied by significantly more muscle tension than the latter,
which may benefit from motor inhibition. (Automatic detection, as de-
scribed by Shiffrin and Schneider, more specifically corresponds to
James's definition of passive derived sensorial attention since it must
be learned.) Moreover, the amount of muscle tension generated during
controlled search would likely be a function of the "load" as defined in
the Shiffrin and Schneider paradigm.
IV. REGULATING THE STREAM OF THOUGHT
A. Ordinary States
In Volume Two of this series, Pope and Singer (1978) presented a
paper entitled "Regulation of the Stream of Consciousness" in which
they described the major factors that determine the contents of one's
daily mental experience. Among their excellent descriptions of different

styles of consciousness regulation, Pope and Singer include a hypo-
thetical case of a person who pushes away distracting thoughts while
reading a book about mathematics but indulges in free association while
reading poetry. This example was chosen by Pope and Singer to illus-
trate that "we are able ... to adopt temporary, situation-specific 'plans'
for mental processing" (p. 126). These investigators, however, do not
describe any mechanism by which such plans could be enacted.
The MTVT, not surprisingly, can describe a mechanism which can
account for this important aspect of consciousness regulation. According
to the MTVT, the person pushes away distracting thoughts while read-
ing by increasing the covert speech motor activity which he uses to
represent the words he is reading. He may also increase tension in the
muscles associated with visual perception and the control of eye move-
ments, including the extraocular muscles, and the muscles of the brow
and neck. Finally, he may increase the tension in his skeletal muscu-
lature generally, but especially in those muscles with which he associates
the act of concentrating. On the other hand, to allow the ideas evoked
by his reading to distract him from the reading process, he need only
relax his motor activity somewhat and refrain from engaging in the types
of motor activity which would direct his attention to his primary task.
In such a manner as described above, the MTVT can explain a very
salient experiential dimension of ordinary consciousness-the feeling
of voluntariness with respect to mental activity. These are times during
the course of ordinary daily activities when we seem to be struggling
with our thoughts-trying to push unwanted thoughts from our mind,
or perhaps repeating particular thoughts in the attempt to find some
elusive solution to a problem. There are other times, for instance, sitting
idly on a bus or train, when our thoughts seem to flow quite by them-
selves, and we feel more that we are observing them than creating them.
According to the MTVT, the feeling of effort and will associated with
some instances of thinking is due to the deliberate engagement of the
motor system during such instances in the service of guiding or con-
trolling the direction of thought. When thoughts seem to flow effort-
lessly, it is because the motor system has not been voluntarily enlisted
in the thinking process. A state in which one is absorbed by an exciting
train of thought can be described as spontaneous internal attention. Any
motor activity (overt or covert) occurring during such a state would be
of an automatic nature, not involving the highest cortical control.
B. Natural Relaxation
An obvious prediction from the MTVT is that states of global motor
quiescence should be accompanied by free-flowing, nonvolitional men-
42 BARRY H. CoHEN
tation. This prediction does not apply to states in which muscular re-
laxation is induced pharmacologically, as with curare, because there
would still be the possibility of considerable central motor activity. Nor
does it apply to the case of spontaneous external attention during which
the decrease in motor activation appears to be the result of an active
inhibition coupled with high cortical arousal. It is when reduced motor
activation occurs due to fatigue, and the environment does not draw
attention, that attention is likely to turn inward. However, this state
differs from spontaneous internal attention because cortical arousal is
likely to be lower than normal. It is more a state of reduced attention.
During such drowsy states, the MTVT predicts that images would flow
more according to principles of free association than logical requirements
because of the reduction in motor control. Even when the individual is
not drowsy, a natural reduction in motor activity should allow ideas to
flow more freely by association. This principle can account for the find-
ing of psychoanalysts that lying on the couch tends to encourage free
association.
Antrobus et al. (1964) interrupted subjects during periods of natu-
rally occurring ocular motility or relative quiescence. The subjects rated
their mental content as it was just preceding the interruption. Ocular
quiescence tended to be associated with more reports of daydreamlike
mentation than did ocular motility. If ocular quiescence is indicative of
reduced motor activation, this finding fits well with the predictions of
the MTVT. Singer (1975) reported the results of personal, introspective
experiments involving interruptions of his own stream of thought: "The
impression I gained ... was that under relaxed, slightly drowsy con-
ditions, one's brain is not geared to producing an orderly sequence of
thought. Personalized visual imagery intruded upon me quite regularly"
(p. 44). From the viewpoint of the MTVT, spontaneous bursts of visual
imagery during relaxed states are quite understandable. As Singer notes,
such imagery is likely to involve a memory sequence which flows ac-
cording to well-worn associations. Much less likely in such a state would
be to "hear" an auditory image of one's own voice composing a complex
essay.
States of profound relaxation, such as occur just prior to the onset
of sleep, are often accompanied by increasingly bizarre combinations of
thoughts and images. If, as the MTVT implies, motor activity is required
for organizing images and thought into analytical and logical progres-
sions, then it is not surprising to find images evoking other images
according to primitive principles of association when the motor system
is inactive. It is, perhaps, no coincidence that the state of normal con-
sciousness most closely associated with bizarre, illogical mentation is
THE MOTOR THEORY OF VOLUNTARY THINKING 43
the state accompanied by the most profound facial motor inhibition-

namely, rapid eye movement (REM) sleep.
Creative thinking often involves a stage of free-flowing mentation
during which ideas can combine in new ways that may even seem il-
logical during ordinary states of thinking. Often a creative episode is
preceded by a period of ordinary mental effort during which some prob-
lem is considered along with proposed logical solutions. However, the
novel, creative solution to the problem often occurs when the individual
has stopped thinking about the problem and has relaxed his motor ac-
tivity. It is in such a state that ideas are more likely to flow by free
association.
C. Progressive Relaxation and Meditation

The MTVT allows that thoughts can arise either by association with
sensations or previous thoughts, or by association with the feedback
from motor activity. At this point, I would like to propose tentatively
that without sustained motor activity any train of associations would
quickly fade away, that it is only the maintenance of motor activity that
can prolong a train of thought indefinitely. In this regard, it should be
recalled that, according to the MTVT, thoughts may arise through as-
sociation with spontaneous or habitual motor activity. Together, these
two notions-that some motor activity is necessary for sustained
thought and that spontaneous motor activity is sufficient for thought-
have some interesting implications. One implication is that very deep
relaxation of the entire motor system should result in the cessation of
thought, but not necessarily of consciousness. Another implication is
that the cessation of thought is impossible without the relaxation of all
thought-evoking motor activity.
The above hypotheses explain why most people find it impossible
to stop their flow of thoughts for more than a very few seconds. Some
of the motor activity which is capable of evoking thoughts occurs ha-
bitually and may be motivated by the same relentless forces that main-
tain various nervous habits, such as nail biting. In order to eliminate a
habit, one must first be made aware of performing the habitual act. But
unlike nail biting, habitual covert motor activity may have no readily
observable consequences other than the thoughts it produces. Still, in
principle, it should be possible to gain voluntary control over any par-
ticular pattern of motor activity; the results of EMG biofeedback exper-
imentation lend much support to this principle (Basmajian, 1979). The
possibility of using EMG biofeedback to gain control of specific thought
patterns seems remote at present. However, two global procedures offer
44 BARRY H. CoHEN
some promise towards the goal of reducing or even eliminating habitual

thought. Jacobson (1938) observed that many individuals exhibited a
good deal of residual muscle tension of which they were unaware when
they were lying down and supposedly relaxing. Typically, a person
asked to relax further would "try" to do so and thus increase, rather
than decrease, his tension. Through his experimentation, Jacobson dis-
covered that subjects could relax residual tension in a particular muscle
more completely when they became aware of the proprioceptive sen-
sations arising from tension in that muscle. To improve his subjects'
ability to perceive muscle tension and thus relax it, Jacobson devised
the technique of progressive relaxation. In this technique, a subject sys-
tematically tenses and relaxes all of the major muscles, noting the feeling
of tension in each, and the feeling of relaxation, of "letting go." Subjects
are encouraged to continue letting go even after the point at which they
believe the muscle has been completely relaxed. Particular attention is
paid to the muscles most closely associated with thought, those involved
with speech and control of visual perception. Remarkably, Jacobson
found that well-trained subjects could attain a state of muscular relax-
ation so profound that all images and thoughts would cease, leaving
the subject awake and alert but with his consciousness devoid of any
particular object. Thus, Jacobson believed that his technique could elimi-
nate habitual thinking, at least temporarily, during total relaxation. He
also believed that the effects of progressive relaxation would generalize
to daily life and minimize excessive rumination or worrying.
The state Jacobson described, in which a subject's mind was alert
but "blank," is similar to the state of "pure awareness" or "pure con-
sciousness" that is said to be the goal of many forms of meditation.
Meditative techniques can be divided into two types depending on
whether they involve active or passive forms of concentration (Naranjo
& Ornstein, 1971). This active-passive distinction is similar to the one
drawn by James in discussing attention. In active concentration, mental
effort is used to direct the focus of attention; in passive concentration,
the attention is left free to be drawn by any mental object, though usually
some focal object, such as the act of breathing or a mantra (a Sanskrit
word that is usually easy to articulate but may have no meaning to the
user), is used as a point of departure. It can be shown, within the context
of the MTVT, how passive mantra meditation can lead, by a comple-
mentary route, to the same state of thought-free alertness described by
Jacobson.
According to the technique of transcendental meditation (TM), a
form of passive mantra meditation popularized by Maharishi Mahesh
Yogi (Bloomfield, Cain, & Jaffe, 1975), one is to return one's attention
to the mantra whenever one becomes aware that his attention has wan-
dered, but one is not to use mental effort to maintain the mantra in
awareness. The effect of "trying" to think of the mantra without ex-
pending effort is that one becomes increasingly aware of the effort ex-
erted in producing any thought. According to the MTVT, the effort in-
volved in producing a thought is nothing more than the effort exerted
to create the motor activity that evokes the thought. By becoming in-
creasingly aware of producing very subtle motor activity, the individual
may gain the ability to relax such motor activity, including the motor
activity responsible for habitual thinking. Thus, it may be possible
through both passive mantra meditation and progressive relaxation to
become aware of and relax away the spontaneous covert motor activity
underlying habitual thought.
The alleged benefits of progressive relaxation Gacobson, 1964) and
transcendental meditation (Bloomfield et al. 1975) with respect to re-
duced anxiety and increased energy and efficiency may be due in large
part to the reduction of habitual thinking that may be produced by both
techniques. Habitual thinking may be harmful not only as a consistent
waste of energy but as a mediator of stress reactions.
V. CONCLUSIONS
A. Summary
The central assertion of the MTVT is that the experience of acting
voluntarily arises solely from activity in the motor system, whether the
voluntary act is physical or purely mental. However, not all activity in
the motor system produces an experience of volition; only activity that
requires deliberate conscious control produces this experience. Because
the ordinary perception of one's environment does not appear to require
any volitional act or to give rise to any sense of effort, the MTVT, unlike
the motor theory of consciousness, does not posit the necessity of motor
activity for the emergence of conscious perceptions. Similarly, some
mental images seem to arise spontaneously by association with a per-
ception or another image. The MTVT does not posit that these mental
experiences require motor activity either, although the possibility is
raised that images which seem to arise spontaneously may actually be
associated with habitual or automatic motor activity.
The MTVT does assert that all images and thought produced or
guided in a voluntary manner are associated with motor activity. De-
liberately imagined kinesthetic imagery is evoked by covert motor ac-
tivity directly corresponding to the imagined activity. Similarly, all de-
liberate inner speech is based on the appropriate covert activity in the
46 BARRY H. CoHEN
speech musculature. Visual imagery, however, cannot be controlled so

closely or directly because there is no one-to-one correspondence be-
tween visual perception and motor patterns. Similarly, imagery in the
other modalities is not directly controlled, nor are shifts in external
attention.
The MTVT asserts that all mental acts which involve the experience
of voluntariness must be accompanied by some distinct pattern of motor
activity. But it is conceded that it is not presently possible to state for
all modalities the exact types of motor ac~ivation involved, nor can the
possibility of significant individual differences be ignored. However,
reasonable speculations can be offered. Inner speech appears to play a
major role in evoking and controlling imagery in other modalities. Inner
speech may be accompanied by motor patterns associated with percep-
tion as appropriate for each modality. Thus, visual imagery may be
accompanied by squinting and, if the images require motion or scanning,
eye movements in the appropriate direction. The voluntary imaging of
tastes or smells may be accompanied by covert oral or nasal adjustments.
Attending to some types of imagery or some aspects of the external
environment may be aided by directional adjustments in bodily posture.
In the most general case, overall increases in bodily tension or, just in
certain idiosyncratically preferred muscles, may aid any form of mental
effort. On the other hand, motor activity may be deliberately inhibited
during states of spontaneous attention, in order to reduce internal dis-
tractions. In such states, only a few muscles may be contracted, and
these contractions would be maintained steadily.
Conversely, during the marked reduction of motor activity pro-
duced by fatigue, the organizing effects of motor activity upon thought
would decrease, and the stream of thought would tend to flow more
by free association than according to logically structured, purposeful
plans. Finally, spontaneous thinking, which may seem virtually im-
possible to eliminate voluntarily, may be the result of habitual covert
motor activity. Both progressive relaxation and certain forms of medi-
tation hold the promise of allowing the individual to gain greater aware-
ness of the habitual motor activity that produces incessant thoughts and
thereby to gain greater control of his consciousness.
B. Implications for Experimental Research

1. The Issue of Necessity
In its present form, the MTVT would not be easy to prove or dis-
prove experimentally. Because the MTVT posits that only central motor
THE MoroR THEORY oF VoLUNTARY THINKING 47
activity is actually necessary to produce the experience of volition during

mental activity, attempts to block muscle contraction peripherally are
irrelevant. Even if data were collected showing that for at least a few
subjects no measurable EMG changes could be found during a period
when the subjects were expending great mental effort and not receiving
any paralytic drugs, it could be argued that these subjects were pro-
ducing tension in idiosyncratic muscle areas not measured. Alterna-
tively, it could be argued that these subjects were generating motor
activity sufficient to generate useful central feedback, but not sufficient
to produce measurable changes in muscle contraction.
To block central motor activity by artificial means is not possible at
the present stage of neurophysiology, and considering the difficulty of
defining motor activity as opposed to other forms of neural activity
within the central nervous system, this means of disconfirming the
MTVT may never be possible. However, there are nonpharrnacological
methods for reducing central motor activity that can be considered. For
instance, if biofeedback procedures are employed to ensure that a subject
maintains a state of relaxation in a certain muscle region, it can be as-
sumed that this relaxation is the result of a reduction in central motor
commands directed at that region. This was the approach adopted by
Hardyck and Petrinovich (1970); these investigators found that subjects
who kept their laryngeal region relaxed while reading a difficult essay
exhibited worse comprehension than subjects who kept their forearms
relaxed, or nonrelaxed control subjects. The group that maintained re-
laxation in their forearms was included to control for the distracting
effects of trying to keep the EMG biofeedback signal off (i.e., keeping
the EMG amplitude below some threshold value). The effectiveness of
this control procedure can be questioned. The tendency for forearm
tension to rise while reading is far less than the tendency for laryngeal
tension to rise (this was even confirmed within the Hardyck and Petri-
novich study, for the nonrelaxation group), so keeping forearm tension
at baseline levels is probably the easier and less distracting condition.
Therefore, distraction cannot be ruled out as an explanation for the im-
paired reading comprehension in the laryngeal relaxation group.
The optimal design for using the biofeedback-relaxation strategy is
to select two different mental tasks which are associated with muscle
tension in two different areas; four conditions are employed such that
in two conditions subjects are relaxing "necessary" muscles and in the
remaining two conditions relaxing "irrelevant" muscles. Unfortunately,
demonstrating that the relaxation of necessary muscles impairs mental
task performance significantly more than the relaxation of irrelevant
muscles would still not prove the necessity of motor activity for _mental
performance. It could always be argued that the tendency to produce
48 BARRY H. CoHEN
tension in one muscle region as opposed to others is invariably greater

for a particular task, even though that tension is not necessary or even
helpful. Because the degree of distraction associated with keeping a
muscle relaxed is likely to increase with the tendency to activate the
muscle during a particular task, distraction could never be ruled out as
the sole cause for the relevant performance decrements. One would
strongly prove the association between motor activation tendencies and
particular mental tasks, but not the absolute necessity of motor activa-
tion. However, the stronger and more reliable these motor tendencies
are proven to be, the more useful and important it becomes to delineate
the exact relation between mental content and motor pattern, as will be
discussed further below.
On the other hand, Cole and Young (1975), using the laryngeal
biofeedback-relaxation technique, claimed to have demonstrated that
subvocalization (i.e., covert speech motor activity) is not necessary for
nonsense syllable memorization. Subjects receiving EMG feedback were
compared to a control group which had no relaxation requirement at
all. Although feedback subjects produced significantly more errors than
the control group, Cole and Young attributed this impairment entirely
to the effects of distraction. These investigators argued that the occur-
rence and type of errors in each group was unaffected by the presence
or absence of subvocalization on any given trial. However, a serious
flaw in Cole and Young's experimental design greatly weakens the force
of their arguments. The EMG was recorded only from the laryngeal area,
and this was the only area that subjects were ever required to relax. The
larynx, though, is not the only area that is activated during audible
speech or covert speech motor activity; according to McGuigan (1978)
the tongue and lips, in that order, are the most sensitive indicators of
inner speech. In fact, the bulk of speech information is encoded in the
form of lip, tongue, and mouth movements; lip readers lose rather little
information even without perceiving the speaker's laryngeal activity. It
is quite possible that feedback subjects quickly learned to "mouth" cov-
ertly the nonsense syllables while suppressing laryngeal activity. In-
deed, subjects may actually have compensated for laryngeal suppression
with an increase in covert tongue and lip activity. Cole and Young's
conclusions would have been far more compelling had they ruled out
the above objection by additionally recording the EMG from the tongue
and/or lips and including a group of subjects required to maintain re-
laxation in several speech areas simultaneously.
To prove that a pattern of motor activity is necessary for a particular
mental event to occur, one must show not only that the motor pattern
is exhibited whenever the mental event occurs but also that preventing
the motor pattern's occurrence also prevents the occurrence of the par-
ticular mental event. It seems safe to state that, at present, the necessity
of motor activity has not been proven with respect to any class of mental
events. Moreover, for the reasons discussed above, it does not seem
likely that the necessity of motor activity will be proven in the near
future. However, although proving the necessity of motor activity for
certain mental experiences would be of enormous theoretical interest,
failing to prove necessity does not render worthless the vast wealth of
data demonstrating the association between motor manifestations and
various mental contents.
2. Applications of Research in Covert Motor Activity
There are at least two important reasons for studying the motor
correlates of mental states even if those motor correlates are totally un-
necessary. First, if patterns of motor activity are correlated in a highly
specific and reliable way with certain aspects of mental experience, the
motor patterns may serve as useful indicators of the type of mental
content. These motor patterns, which can be measured through the use
of several simultaneous EMG recordings, can provide clues concerning
a subject's cognitive strategy at times when verbal report may be in-
convenient, distracting, unreliable, or simply unavailable (e.g., aphasic
patients). Second, the motor activity accompanying mental effort,
whether helpful or not, can have psychological and physiological con-
sequences that are of practical importance. In extreme cases, the muscle
tension generated during mental concentration may produce a debili-
tating degree of discomfort.
In regard to the use of EMG patterns as indicators of cognitive strat-
egy, the recent work of Cacioppo and Petty (1981) should be noted.
These researchers monitored lip EMG during several cognitive tasks
which differed in their expected demand on the use of inner speech.
The relative amounts of lip tension generated by the different tasks
conformed well with theoretical predictions. The EMG was also recorded
from the left forearm to rule out the possibility that changes in general
arousal, rather than inner speech, could account for the task differences
in lip tension. Forearm EMG did not differ significantly between con-
ditions, confirming the specificity of the observed changes in lip tension.
The Cacioppo and Petty experiment represents an excellent example of
the way measurements of covert motor activity can be measured and
used as a tool to explore cognitive processes. However, one improve-
ment should be mentioned. By using forearm EMG as a comparison to
control for the effects of general arousal, Cacioppo and Petty implicitly
assumed that the forearm would be as responsive-or at least nearly
so-as the lips to changes in general arousal. This need not be so. It is
50 BARRY H. COHEN
quite possible, as has been argued before (Cohen, 1983), that some mus-
cles, such as the orbicularis oris (which manipulates the lips), or the
corrugator (which pulls down the brow), are far more responsive than
other muscles to changes in general arousal. The use of additional si-
multaneous EMG recordings to measure more comprehensiv ely the ten-
sion patterns that accompany the performance of cognitive tasks would
expand the usefulness and render more conclusive the results obtained
by Cacioppo and Petty.
The design of the Cacioppo and Petty experiment was based on the
great volume of research (a large portion of which was conducted by
McGuigan and his associates) which has established the reliable and
specific relation between the use of inner speech and the covert acti-
vation of the speech musculature. Further research involving the motor
correlates of other types of mental activity is needed to expand the use-
fulness of covert motor activity as an indicator of cognitive strategy. For
example, the motor correlates of visual imagery should be studied more
thoroughly, using EMG recordings from several facial areas, particularly
around the eyes, in addition to the EOG for measuring eye movements.
The visual imagery tasks used in these experiments should be designed
so as to control more carefully various aspects of the images subjects
are expected to form, such as the size of the image and the direction of
any required motion or scanning. Research involving imagery in other
modalities, though less useful in the exploration of higher cognitive
processes, would be of considerable theoretical importance, as the re-
sults could refine and clarify our conception of the MTVT.
Much research effort has recently been directed towards the clas-
sification of facial expressions associated with distinct emotions, and the
use of facial EMG as an indicator of specific affect (Fridlund and Izard,
1983). The bulk of this research has focused on affective states that are
typically regarded as highly emotional; little attention has been paid to
the affective state known as "interest." However, because interest often
accompanies and blends with various emotions, and because for many
individuals interest may be present by itself more frequently than any
other affective state, it would seem worthwhile to explore the motor
correlates of mental states involving relatively pure interest. It should
be noted, however, that these motor correlates may vary according to
the general direction in which the interest is deployed. For instance,
interest stimulated by the environment may be associated with a some-
what different facial expression (and other motor manifestations) from
interest in one's own mental imagery. Moreover, interest involving au-
dition may differ motorically from interest in the visual field. During
the early phases of research into the motor correlates of interest, in-
vestigators should be careful to note the exact conditions by which the
THE MoroR THEORY OF VoLUNTARY THINKING 51
state of interest was evoked and to record from as wide a representation

of muscle areas as possible.
It would be very helpful if the motor correlates of specific types of
attention proved to be universal, exhibiting similar motor patterns from
one individual to the next; assumptions could then be made about in-
dividuals whose motor patterns had not yet been observed in controlled
circumstances. However, it would also be useful if individuals fell into
well-defined categories based on their motor patterns during mental
work; these categories might then be found to represent important and
consistent individual differences with respect to cognitive style. But even
if the motor correlates of mental activity were highly idiosyncratic-
differing considerably between any two individuals-as long as these
patterns were consistent within each individual over time, they could
still be useful. One form of attention, the motor correlates of which could
be expected to be rather idiosyncratic, is attention which is directed to
one of several similar-sounding auditory stimuli distributed widely in
space. Different individuals may employ a different set of motor patterns
to aid their efforts in directing attention to a particular spatial location.
One individual might control his attention in this situation by the po-
sition of his eyes, whereas a second individual might adjust the position
of his jaw. (In the initial stages of this experimentation, many channels
of simultaneous EMG recordings, as well as EOG and visual observation,
would have to be employed to cover the many possible motor strategies
that could arise.) If an individual maintained such a motor strategy con-
sistently throughout an experiment, this information could be useful in
subsequent experimental sessions with the same individual, in which
the individual might be induced, for instance, to switch his attention
rapidly between two inputs the spatial location and separation of which
could be systematically varied. Just to observe the different motoric
strategies of several experimental subjects as they switch the direction
of their attention could be a useful first step in this research program.
The second reason mentioned above for studying the motor activity
accompanying mental work involves the consequences of such motor
activity. Regardless of whether such motor activity serves any useful
function, it can result in high levels of muscle tension. There is evidence
that muscle tension in some bodily areas, particularly the forehead, can
rise continuously during sustained mental performance. More research
is needed to determine whether the tension in these areas can eventually
produce enough discomfort to interfere with mental performance. It is
likely that individuals who are prone to having tension headaches would
show steeper forehead tension gradients during mental work. Other
individuals may show similar gradients in other muscles, perhaps in
the jaw or neck. Even among those who rarely experience pain from
52 BARRY H. CoHEN
muscular tension headaches, tension levels may reach high enough lev-
els to produce distraction and diminish mental concentration. The pos-
sibility ought to be explored that individuals could be trained to increase
their span for concentration by learning to relax certain muscles during
mental performance. This was the hope of Jacobson (1938) in devising
the concept of differential relaxation.
Finally, the possibility of reducing or eliminating extraneous and
even debilitating thoughts through refined techniques of relaxation is
too intriguing to be ignored. The potential benefits of such relaxation
for mental health are enormous. Although numerous studies have been
conducted using EMG biofeedback and abbreviated forms of progressive
relaxation, such studies have not focussed on the relation between covert
motor activity and the intensity of concurrent mental activity. Perhaps
mental devices for relaxation can be validated and refined through the
use of multichannel EMG measurement. Such experimentation could
be capable of combining the benefits of ancient wisdom concerning med-
itation practices with modern technology in order to approach the high-
est goal of self-control: control of consciousness itself.
REFERENCES
ANTRoBus, J. S., ANTROBUS, T. S., & SINGER, J. L. (1964). Eye movements accompanying
daydreaming, visual imagery and thought suppression. Journal of Abnormal and Social
Psychology, 69, 244-252.
BAIN, A. (1888). The emotions and the will. London: Parker.
BARTOSHUK, A. K. (1956). EMG gradients and EEG amplitude during motivated listening.
Canadian Journal of Psychology, 10, 156-164.
BASMAJIAN, J. V. (1979). Muscles alive: Their functions revealed by electromyography. Baltimore:
Williams & Wilkins.
BERNHAUT, M., GELLHORN, E., & RAsMUSSEN, A. T. (1953). Experimental contributions to
the problem of consciousness. Journal of Neurophysiology, 16, 21-35.
BLOOMFIELD, H. H., CAIN, M.P., & }AFFE, D. T. (1975). Transcendental meditation: Discovering
inner energy and overcoming stress. New York: Delacorte.
BROWN, B. B. (1968). Visual recall ability and eye movements. Psychophysiology, 4, 300-306.
CAOOPPO, J. T., & PEITY, R. E. (1981). Electromyographic specificity during covert infor-
mation processing. Psychophysiology, 18, 518-523.
CoHEN, B. H. (1978, September). Role of the motor system in visual imagery. Paper presented
at the annual meeting of the American Psychological Association, Toronto, Canada.
CoHEN, B. H. (1983). Electromyographic patterns associated with cognition and affect. Unpub-
lished doctoral dissertation, New York University.
CoHEN, B. H., & JoNAS, C. (1978, November). A closer look at the visualizer: Habit or vividness?
Paper presented at the second annual convention on the Imagery and Fantasy Process,
Chicago.
COLE, R. A., & YouNG, M. (1975). Effect of subvocalization on memory for speech sounds.
Journal of Experimental Psychology: Human Learning and Memory, 1, 772-779.
CUTHBERT, B. N., & LANG, P. J. (1984). Eye movements in visual imagery. Physiology, 21,
574.
DECKERT, G. H. (1964). Pursuit eye movements in the absence of a moving visual stimulus.
Science, 143, 1192-1193.
DuNLAP, K. (1927). The short-circuiting of conscious responses. Journal of Philosaphy, 24,
263-267.
EHRLICHMAN, H., & BARRETT, J. (1983). 'Random' saccadic eye movements during verbal-
linguistic and visual-imaginal tasks. Acta Psychologica, 53, 9-26.
EKMAN, P., FRIESEN, W., & ELLSWORTH, P. (1972). Emotion in the human face. New York:
Pergamon Press.
EVARTS, E. V. (1971). Feedback and corollary discharge: A merging of the concepts. Neu-
rosciences Research Program Bulletin, 9, 86-112.
FRIDLUND, A. J., & IZARD, C. E. (1984). Electromyographic studies of facial expressions of
emotions and patterns of emotions. In J. T. CACIOPPO & R. E. PETTY (Eds.), Social
psychophysiology: A sourcebook, New York: Guilford Press.
GEISSER, S., & GREENHOUSE, S. W. (1958). An extension of Box's results on the use of the
F distribution in multivariate analysis. Annals of Mathematical Statistics, 29, 885-891.
GELLHORN, E. (1958). The influence of curare on hypothalamic excitability and the elec-
troencephalogram. Electroencephalography and Clinical Neurophysiology, 10, 697-703.
HARDYCK, C. D., & PETROVICH, L. F. (1970). Subvocal speech and comprehension level as
a function of the difficulty of reading material. Journal of Verbal Learning and Verbal
Behavior, 9, 647-652.
HEss, D. 0. (1968). Concerning imagery. Psychological Review, 75, 466-477.
HEFFERLINE, R. F., & PERERA, T. B. (1963). Proprioceptive discrimination of a covert operant
without its observation by the subject. Science, 139, 834-835.
JACOBSON, E. (1930). Electrical measurements of neuromuscular states during mental ac-
tivities. III. Visual imagination and recollection. American Journal of Physiology, 95, 694-
702.
]ACOBSON, E. (1932). Electrophysiology of mental activities. American Journal of Psychology,
44, 677-694.
]AcossoN, E. (1938). Progressive relaxation (2nd ed.). Chicago: University of Chicago Press.
]ACOBSON, E. (1964). Anxiety and tension control: A physiologic approach. Philadelphia:
Lippincott.
]AMES, W. (1890). The principles of psychology. New York: Holt. (Reprinted: New York: Dover,
1950).
LACEY, J. I. (1967). Somatic response patterning and stress: Some revisions of activation
theory. In M. H. APPLEY & R. TRUMBULL (Eds.), Psychological stress: Issues in research.
New York: Appleton-Century-Crofts.
LENox, J. R., LANGE, A. F., & GRAHAM, K. R. (1970). Eye movement amplitude in imagined
pursuit of a pendulum with the eyes closed. Psychophysiology, 6, 773-777.
LEUBA, C., BIRCH, L., & APPLETON, J. (1968). Human problem solving during complete
paralysis of the voluntary musculature. Psychological Reports, 22, 849-855.
LIBBY, W. L., ]R., LACEY, B. C., & LACEY, J. I. (1973). Pupillary and cardiac activity during
visual attention. Psychophysiology, 10, 270-294.
LURIA, A. R. (1966). Higher cortical functions in man. London: Tavistock.
MARKS, D. F. (1973). Visual imagery differences in eye movements in recall of pictures.
Perception and Psychophysics, 14, 407-412.
MATIN, L., PEARCE, D., MATIN, E., & KIBLER, G. (1966). Visual perception of direction in
the dark: Roles of local signs, eye movements and ocular proprioception. Vision Re-
search, 6, 453-469.
MAUDSLEY, H. (1889). The physiology of mind. New York: D. Appleton.
McGUIGAN, F. J. (1978). Cognitive psychophysiology: Principles of covert behavior. New Jersey:
Prentice-Hall.
54 BARRY H. CoHEN
McGuiGAN, F. J., & RoDIER, W. 1., III. (1968). Effects of auditory stimulation on covert oral
behavior during silent reading. Journal of Experimental Psychology, 76, 649-655.
McGuiGAN, F. J., & WINSTEAD, JR., C. L. (1974). Discriminative relationship between covert
oral behavior and the phonemic system in internal information processing. Journal of
Experimental Psychology, 103, 885-890.
NARANJO, C., & ORNSTEIN, R. E. (1971). On the psychology of meditation. New York: Viking.
NoRMAN, D. A., & SHALLICE, T. (1985). Attention to action: Willed and automatic control
of behavior. In R. J. DAVIDSON, G. E. ScHWARTZ, & D. SHAPIRO (Eds.), Consciousness
and self-regulation: Advances in research and theory, Vol. 4. New York: Plenum Press.
0BRIST, P. A., WEBB, R. A., & SurrERER, J. R. (1969). Heart rate and somatic changes during
aversive conditioning and a simple reaction-time task. Psychophysiology, 5, 696-723.
0BRIST, P.A., WEBB, R. A., SUTTERER, J.R., & HowARD, J. L. (1970). The cardiac-somatic
relationship: Some reformulations. Psychophysiology, 6, 569-587.
PILLSBURY, W. B. (1908). Attention. New York: Macmillan.
PINNEO, L. R. (1961). The effects of induced muscle tension during tracking on level of
activation and on performance. Journal of Experimental Psychology, 62, 523-531.
POPE, K. S., & SINGER, J. L. (1978). Regulation of the stream of consciousness: Toward a
theory of ongoing thought. In G. E. ScHWARTZ & D. SHAPIRO (Eds.), Consciousness and
self-regulation: Advances in research and theory. Vol. 2. New York: Plenum Press.
RIBOT, T. A. (1889). The psychology of attention. New York: Humboldt.
SHAW, W. A. (1938). The distribution of muscular action potentials during imaging. Psy-
chological Record, 2, 195-216.
SHIFFRIN, R. M., & ScHNEIDER, W. (1977). Controlled and automatic human information
processing: II. Perceptual learning, automatic attending, and a general theory. Psy-
SINGER, J. L. (1975). The inner world of daydreaming. New York: Harper & Row.
SINGER, J. L., & ANTROBUS, J. S. (1965). Eye movement during fantasies. Archives of General
Psychiatry, 12, 71-76.
SMITH, M. 0. (1969). History of the motor theories of attention. Journal of General Psychology,
80, 243-257.
SMITH, S.M., BROWN, H. 0., ToMAN, J. E. P., & GoODMAN, L. S. (1947). The lack of cerebral
effects of d-tubocurarine. Anesthesiology, 8, 1-14.
SPERRY, R. W. (1952). Neurology and the mind-brain problem. American Scientist, 40, 291-
312.
STOY, E. G. (1930). A preliminary study of ocular attitudes in thinking of spatial relations.
Journal of General Psychology, 4, 379-385.
ToTTEN, E. (1935). Eye movement during visual imagery. Comparative Psychology Mono-
graphs, 11, No. 3.
WALLERSTEIN, H. (1954). An electromyographic study of attentive listening. Canadian Journal
of Psychology, 8, 228-238.
WASHBURN, M. F. (1916). Movement and mental imagery. New York: Houghton Mifflin.
WEIMER, W. B. (1977). A conceptual framework for cognitive psychology: Motor theories
of the mind. In R. SHAW &J. BRANSFORD(Eds.), Perceiving, acting, and knowing. Hillsdale,
NJ: Erlbaum.
Cardiac Afferent
3
Influences on Consciousness
CURT A. SANDMAN
J. INTRODUCTION
The separation of reality from our experience of it has intrigued, con-

fused, and consumed the greatest thinkers of our civilization. Philo-
sophical treatises have been devoted to its analysis, dualistic religions
have deified it, and theoretical physics has become its current harbinger.
Although it is inviting to plunge impulsively into debate of the issue,
instead a framework or paradigm will be erected (or resurrected) for
merging mentalistic with psychobiological constructs. Within this frame-
work the separation of reality from its experience had its most obvious
expression in the posthumous "debates" of Walter Cannon and William
James. The focus of these debates was the necessary ingredients of
experience.
In the analysis developed by Cannon (1929), sensory input triggered
a predetermined pattern of action or experience. The output of the sys-
tem (effector system) was graded in terms of its magnitude or level of
arousal (the "bang" theory of emotion) and was controlled solely by the
central nervous system. Larger responses were assumed to relate to
more intense stimuli (experiences), but qualitative differences in con-
sciousness were not accommodated. This reflexive theory still guides
most modern-day physiological studies of consciousness.
A radically different view was proposed by James (1892). He sug-
gested that the experience of reality was a product of visceral and au-
tonomic communication with the brain. Perception (or the experience
of reality) gained meaning in the context (apperception) in which it was
experienced. The peripheral nervous system, coupled with the envi-
ronment, provided this context. Thus, the viscera, the muscles, and the
CURT A. SANDMAN e Fairview Developmental Center, Costa Mesa, California 92626; De-
partment of Psychiatry and Human Behavior, University of California Irvine Medical Cen-
ter, Orange, California, 92668.
55
56 CURT A. SANDMAN
endocrine system provided information to the brain and thereby par-

ticipated in the determination of behavior, thoughts, and consciousness.
The notion that "we are afraid because we run" derives from this view.
The intuitive appeal of this formulation is evident in our everyday ac-
tivity. Imagine driving down a residential street when all of a sudden
a small child darts in front of your car. It is likely that you immediately
step on the brakes and, assuming you have missed the child, sit there
and ponder your close call. All of a sudden you begin to shake, your
heart beats faster, your palms become moist, and your emotional state
climaxes. This scenario illustrates that often we first act and then, as the
consequences of physiological and psychological action become "con-
scious," we experience an emotion.
A. Consciousness as Cognition and Emotion (Brain and Body)

Stimulated by the work of Lazarus (1966; Lazarus, Speisman, Mord-
koff, & Davidson, 1962, 1966) indicating predictable relationships be-
tween the focus of attention and emotion, evidence secured in my lab-
oratory (Sandman, 1971, 1975) suggested that the views of both Cannon
and James have merit. Physiological and behavioral responses to stress-
ful, neutral, and pleasant stimuli in subjects with distinctive perceptual
or attentional styles were compared. One group, the field-independent
subjects, use visceral or bodily information more readily than others and
are able to make accurate judgments about the environment even though
they may be presented with distracting perceptual information. They
evidence qualitative differences in physiological response profiles and
apparently enjoy richer emotional existence. Further, they provide more
accurate information about their physiological state. The second group,
field-dependent subjects, base their perceptual judgments primarily
upon external information. Since they utilize environmental, rather than
internal, information to assess their emotional or conscious state, they
can be misled by nonveridical or distracting environmental stimulation.
They tend to have minimal emotional complexity and may respond to
stimuli with differing degrees of arousal (Witkin, Dyk, Fattuson, Good-
enough, & Karp, 1962).
In my studies (Sandman, 1971, 1975), these two groups of subjects
responded differently to the emotionally charged stimuli. Specifically,
field-independent subjects displayed different profiles of physiological
responses to the different classes of stimuli. Further, they provided more
accurate self-reports about their physiological state and evidenced re-
markable concordance between how they said they felt, the length of
time they viewed the stimulus, and their physiological profiles. Field-
CARDIAC AFFERENT INFLUENCES oN CoNSCIOUSNEss 57
dependent persons failed to exhibit different patterns of physiological

responses to the affective stimuli and responded with differing degrees
of arousal. Moreover, there was no relationship among their physio-
logical responses, their ratings of the stimuli, and their viewing time.
These data indicated that perceptual or attentional style influenced
emotional experience or consciousness. The field-independent subjects,
with easy access to visceral and autonomic input, responded to envi-
ronmental stress as predicted by James. The field-dependent group,
with limited access to peripheral information and greater reliance upon
environmental cues, evidenced a pattern of response characterized by
arousal as predicted by Cannon. It is conceivable that these two per-
vasive theories of consciousness simply describe two distinctive cog-
nitive or perceptual strategies partially reflected by field dependence or
independence. Of greatest interest, and a theme developed in this chap-
ter, was the subsequent finding that cognitive strategy, perhaps the
"guide" to consciousness, could be modified by learned control of the
peripheral nervous system (McCanne & Sandman, 1976). Lest these
speculations be viewed as fanciful, the following is a brief review of the
mechanisms underlying the interaction of the peripheral nervous system
(specifically the cardiovascular system) with the central nervous system.
JI. CARDIOVASCULAR AND BARORECEPTOR PHYSIOLOGY
The existence of pressure-sensitive receptors (baroreceptors) in the

carotid sinus and aortic arch is well established. Nerves from the carotid
sinus and aortic arch join the vagus and the glossopharyngeal nerves,
terminate in the lower brainstem, and assist in providing homeostatic
control of blood pressure to ensure survival of the organism. The bam-
receptors increase their firing rates during transient blood pressure in-
creases and decrease their rate of discharge as blood pressure falls. Stud-
ies with spontaneous hypertensive rats indicated that the central
nervous system was responsible for setting the operational pressure-
sensitive range for the baroreceptors Gudy & Farnell, 1979). Participation
of the central nervous system in this homeostatic process complements
the role of the cardiovascular system.
However, the baroreceptors have functions in addition to those clas-
sified as homeostatic (see Figure 1). In 1929 Tournade and Malmejac
found that stimulation of the carotid sinus nerve diminished muscle tone
in anesthetized animals. Shortly thereafter, Koch (1932) found that in-
creased pressure in the carotid sinus of a dog led to decreased motor
activity and even prolonged sleep. These two reports were among the
first to suggest that baroreceptor activity influenced higher levels of the
58 CURT A. SANDMAN
SYSTOLE
SYSTOLE DIASTOLE
DECREASED SPINAL CORD EXCITABILITY INCREASED SPINAL CORD EXCITABILITY

PULSE TREMOR OF SKELETAL MUSCLES (;r MOTORr-!EURONI lNCREASED NOREPINEPHRINE IN LOCUS
DEPRESSED SINGLE CELLS IN NUCLEUS CUNEATUS COERULEUS
AfTER SKIN STIMULATION INCREASED EXPLORATION
INHIBITION OF NOREPINEPHRINE IN LOCUS COERULEUS DECREASED BARORECEPTOR DISCHARGE
DIMINISHED MUSCLE TONE CEREBRAL VASODILATION
DECREASED MOTOR ACTIVITY DECREASED BLOOD PRESSURE
lNCREASEO BARORECEPTOR DISCHARGE RELEASE OF CNS INHIBITION
CEREBRAL VASOCONSTRICTION ENHANCED EVENT·REUHEO POTENTIALS
INCREASED BLOOD PRESSURE OF THE BRAIN
PROLONGED SLEEP RELIABLE EVOKED VASCULAR RESPONSE
lNHif:ITION OF CNS {RELEASED BY SEVERING VAGUS DECREASED PERCEPTUAL THRESHOLO

OR GLOSSOPHARYNGEAL NERVES l
ATTENUATED EVENT-RELATED POTENTIALS OF THE
BRAIN
UN.'IELIA8LE EVOKED VASCULAR RESPONSE
INCREASED PERCEPTUAL THRESHOLD
DIASTOLE
FIGURE 1. Relationships of physiological, biochemical, electrophysiological, and behav-

ioral phenomena with phase (systole/diastole) of carotid pulse-pressure gradient.
nervous system than those needed to maintain cardiovascular homeo-

stasis and that these influences were inhibitory rather than excitatory.
In an ingenious study, Bonvallet, Dell, and Hiebel (1954) distended
the carotid sinus of cats and produced an experimental analog of in-
creased pressure. They discovered that increased pressure shifted elec-
trocortical activity from low-voltage fast activity to high-voltage slow
activity. Thus, when the baroreceptors in the wall of the carotid sinus
detected increased pressure, the electrocortical activity was inhibited.
Decreased blood pressure released the cortex from this inhibitory influ-
ence. Indeed, Bonvallet et al. (1954) discovered that if one severed
the vagus and glossopharyngeal nerves the inhibitory influence
dissipated.
More recent experiments have demonstrated significant barorecep-
tor input to areas of the brain remote from those previously associated
with cardiovascular control. In fact, contemporary cardiovascular phys-
iologists have suggested that the "vasomotor center" be expanded to
include a number of areas superior to the brainstem (Joy, 1975). In an
investigation of baroreceptor input to the hypothalamus, Adair and
Manning (1975) illustrated the importance of these supramedullary con-
nections by recording evoked potentials in the posterior hypothalamus
as early as 10 msec following stimulation of the carotid sinus. Moreover,
when these hypothalamic areas were stimulated, a 65% reduction in
single-unit firing occurred in medullary neurons responsive to barore-
ceptor activation. Thus, even before reaching the vasomotor center,
baroreceptor activity was detected by the hypothalamus and subse-
quently channeled to a number of supramedullary structures.
One supramedullary structure which is responsive to vasomotor

activity is the locus coeruleus. A series of dramatic studies by Svensson
(Svensson & Thoren, 1979; Persson & Svensson, 1981) indicated that
peripheral blood volume receptors participated in control of norepi-
nephrine (NE) in the locus coeruleus (LC) and in behavior. Blood was
withdrawn (experimental hemorrhage) or fresh blood loaded into con-
scious rats. After blood loading (increased pressure) behavioral depres-
sion and inhibition of NE and neuronal firing rate was observed in the
LC. Conversely, hemorrhage resulted in increased exploration and ac-
tivation of NE and firing rate in the LC. Earlier, Coleridge, Coleridge,
and Rosenthal (1976) found that distension of the carotid sinus caused
prolonged depression of activity of pyramidal tract cells in the motor
cortex. This depression ranged from a 15% reduction in firing to com-
plete cessation of activity and lasted approximately 85 sec after the dis-
tension ceased. Similarly, human spinal cord excitability has been shown
to vary directly with the cardiac pulse. Forster and Stone (1976) dem-
onstrated that the "physiological tremor" of normal skeletal muscles was
a function of cardiovascular modulation, presumably through y-motor-
neurons. These authors speculated that the rising phase of systolic pres-
sure might alter neuronal excitability by a piezoelectric effect on motor
neuron membrane. They offered an equally intriguing alternative pos-
sibility that neuronal firing rate changes were a function of microcir-
culatory, oxygen-carbon dioxide tension during the cardiac cycle.
Sensory processes also are influenced by baroreceptor activity. Gah-
ery and Vigier (1974) showed that stimulation of baroreceptor afferents
depressed the responses of single cells in the nucleus cuneatus after
stimulation of the skin. These data emphasized that baroreceptors play
an important role in sensory and motor functions as well as in the control
of blood pressure.
Thus, neurophysiological evidence suggested that increased blood
pressure detected by baroreceptors were transmitted through the vagus
and the glossopharyngeal nerves to the area of the brainstem maintain-
ing homeostasis and to other areas which may serve to inhibit cortical,
autonomic (except cardiovascular), and muscular activity. Accordingly,
increased blood pressure was part of an inhibitory or restraining process
rather than an activating process. Conversely, decreased blood pressure
released this inhibition, resulting in lowering of sensory thresholds and
prolongation of stimulus impact.
The importance of the cortical inhibitory influence of the barore-
ceptors may have adaptive significance for the organism. For instance,
the heart and even the baroreceptors (Obrist, Light, McCubbin, Hutche-
son, & Hoffer, 1979) respond to environmental stimulation. These
changes are detected in bulbar structures of the brain. Thus, cardio-
vascular responses to external events generate a direct inhibitory influ-
60 CuRT A. SANDMAN
ence on the central nervous system. Further, cells firing with a cardiac
rhythm have been recorded in the bulbar areas (Humphrey, 1967; Smith
& Pearce, 1961), and coagulation in this region prolonged the effects of
a stimulus. J. I. Lacey (1967) has suggested that the functions of this
area, rich with cardiovascular representation, may be to "control the
duration of an episode of stimulus produced in the brain" (p. 27).
The complexities and problems associated with this explanation
have not gone unrecognized (Lacey & Lacey, 1970). One problem arises
as a result of the fact that the carotid sinus is not purely passive. It has
its own properties, and Peterson (1962) has shown that the stiffness of
the carotid sinus wall, one determinant of baroreceptor sensitivity, is
affected by acetylcholine and norepinephrine. It seems clear that other
nervous system activity may alter the stiffness of the wall and the sen-
sitivity of the baroreceptors. Thus, the inhibition thought to be deter-
mined solely by baroreceptor activity may be determined in part by other
activity of the nervous system affecting the wall of the carotid sinus.
It is not surprising, therefore, that inhibitory effects do not occur
every time blood pressure increases. It is clear that other processes mod-
ify the cardiovascular-central nervous system relationship. Exercise,
for example, will not necessarily lead to inhibitory effects on the organ-
ism. The inhibition is subject to modification by higher levels of the
central nervous system, not only from the level of the wall of the carotid
sinus itself to the area of the brain it ultimately reaches, but also from
the area of the brain back to the effector processes where inhibition is
observed.
Another limitation is the fact that there are many baroreceptors
throughout the body in addition to those in the aortic arch and carotid
sinus. It would be naive to assert that only those in the aortic arch and
carotid sinus bear any relationship to the central nervous system and
behavior. Undoubtedly, there are complex interactions at many levels
of the nervous system, among various baroreceptor systems that are
scattered throughout the body. For instance, recently Thompson and
his colleagues (1979, 1980, 1983) have mapped rapidly responding ve-
nous afferents from femoral and brachial veins to the motor-sensory
cortex of cats. Venous afferents were stimulated with 200 msec current
pulses while recordings were made from the pial surface of the cortex.
Stimulation of forelimb and hindlimb venous afferents resulted in
unique topographic distribution of the cortical evoked response. The
authors suggested that cardiovascular control may be accomplished
either by very rapid response to venous pressure or anticipation of cir-
culatory challenge. Recently (1983), these authors have used perfusion
distension and mechanical stretch of femoral-saphenous vein, within
CARDIAC AFFERENT INFLUENCES ON CoNSCIOUSNESs 61
physiological limits, and recorded cord dorsum potentials. Thus, unique

topographic patterns in the CNS can be measured in response to car-
diovascular changes distal from the carotid artery which are within phys-
iological limits.
The temporal sequence of baroreceptor discharge poses another
problem. Frequency of baroreceptor firing with changes in blood pres-
sure may be slightly out of phase measured from different locations.
Furthermore, little is reported concerning the interaction between tonic
levels of blood pressure and phasic changes in blood pressure even at
the level of the baroreceptors.
Finally, the significance of cerebral blood flow in this inhibitory
process is unclear. When carotid baroreceptors are stimulated, there is
a decrease in arterial blood pressure which may lead to a significant fall
in cerebral blood flow (Purves, 1972). Diminished cerebral blood flow
causes a decrease in oxygen available to the tissues, which might account
for neuronal inhibition. Although the widely accepted view is that cer-
ebral blood flow is autoregulatory and not subject to any significant
neurogenic control (see Purves, 1972, for a review), there are data to
indicate that cerebral vasoconstriction occurs when systemic blood pres-
sure is raised and that this vasoconstriction is related specifically to baro-
receptor activity (Ponte & Purves, 1974). Stimulus-linked response in
localized pulsatile flow has been measured from the thalamus, hypo-
thalamus, and globus pallidus (Birzis & Tachibana, 1964; Te.chibana,
Kuramoto, Inanaga, Ikemi, 1967). Such rapid blood flow changes prob-
ably are not the result of local regulatory mechanisms (i.e., metabolic
end-products such as carbon dioxide) but may be neurally mediated
events which increase or decrease the availability of essential metabolic
substrates (e.g., oxygen, glucose) in anticipation of altered demands.
The recent description of a cerebral evoked vascular response (Sandman,
O'Halloran, and Isenhart, 1984) supports the coupling of neuronal ac-
tivity and cerebral blood flow (Ingvar, 1972).
The evolving relationship among the electrical and vascular activity
of the brain, the phase of the heart, and behavior is illustrated in Figure
2. Each ventricular contraction of the heart propagates a bolus of blood
through the vascular system which is detected as a resonating pulse.
Carotid systolic pressure begins to ascend with ventricular contraction
and reaches a peak value several milliseconds later. The firing rate of
the baroreceptors is related both to transient changes in pressure and
heart rate, but the precise nature of their interaction is unknown. In-
creased pressure also corresponds to slowing or inhibitory patterns of
EEG activity. With decreased pressure the inhibitory pattern is released
resulting in lower voltage, faster activity. Finally, as will be discussed
62 CURT A. SANDMAN
VENTRICULAR
CONTRACTION
ATRIAL ../ R
CONTRACTION
\ T
INTERBEAT ELECTROCARDIOGRAM
INTERNAL
(HEART RATE)
CAROTID PULSE
PRESSURE WAVE
SYSTOLE DIASTOLE
11111111111111111111111 1 I IIIII I IIIII

r--------------, BARORECEPTOR
ENHANCED ATTENTION FACILITATION OF DISCHARGE
TO THE ENVIRONMENT COGNITIVE PROCESSES
ELECTROENCEPHALOGRAM
AUDITORY ERP
SYNCHRONIZED
WITH SYSTOLE
Nl
STIMULATION
p2
(SYSTOLE)
AUDITORY ERP
SYNCHRONIZED
WITH D lAS TOLE
Nl
STIMULATION
(DIASTOLE)
EVOKED VASCULAR
RESPONSE (SYSTOLE)
STIMULATION
(SYSTOLE)
EVOKED VASCULAR
RESPONSE (DIASTOLE)
STIMULATION
(DIASTOLE)
FIGURE 2. Schematized view of the relationships among the cardiovascular system, central
nervous system, and behavior.
CARDIAC AFFERENT INFLUENCES ON CoNSCIOUSNESS 63
in detail below, these patterns are coupled with unique psychological

states and event-related patterns of the EEG.
III. CARDIOVASCULAR RELATIONS TO BEHAVIOR
In a series of tasks ranging from those requiring attention to the

external environment (i.e., detection of flashes) to those in which at-
tention to the environment may interfere with performance of the task
(i.e., mental arithmetic) it was discovered that heart rate and blood pres-
sure were the physiological responses which best differentiated cogni-
tive-perceptual processes (Lacey, 1959, 1967; Lacey, Kagan, Lacey, &
Moss, 1963). Heart rate decreased during tasks requiring environmental
attention, whereas tasks demanding "mental concentration" or "rejec-
tion of the environment" were related with heart-rate acceleration.
Further, as discussed above, studies from our laboratory (Sandman,
1971, 1975; Walker & Sandman, 1977) and others (Hare, 1973; Libby,
Lacey & Lacey, 1973) indicated that heart rate decelerated while subjects
were viewing unpleasant stimuli. This response was surprising since
heart rate acceleration commonly is believed to be an aspect of the overall
sympathetic nervous system response to stressful stimuli. However,
stressful stimuli can invoke a strong attentional demand with obvious
evolutionary significance.
In a refined analysis of this issue Cacioppo and Sandman (1978)
presented subjects with stressful visual stimuli (pictures of accident vic-
tims) and stimuli which were equated with the pictures of accident vic-
tims on the dimension of unpleasantness but required "cognitive effort."
These latter stimuli included arithmetic problems, anagrams, and strings
of digits to be memorized. Consistent with previous studies, heart rate
decelerated during stressful stimuli but increased for the equally stress-
ful stimuli which required mental effort. Thus, it appeared that heart
rate reflected the type and perhaps the level of processing required by
the task but did not reflect its affective components (see Table 1).
A. Control of Body, Control of Consciousness

Among the most exciting implications of the relationships among
the bram, behavior, and the body is the possibility that specific altera-
tions in physiology may be linked to discrete changes in behavior. Sev-
eral studies conducted in our laboratory provided support for this pos-
sibility. The first two studies (McCanne & Sandman, 1974, 1976)
indicated that perceptual threshold and selectivity of externally directed
64 CuRT A. SANDMAN
TABLE 1
Relationships Between Heart Rate and the Brain and Behavior
Lowered heart rate Elevated heart rate
Slower reaction time in cognitive task Faster reaction time in cognitive task
Faster reaction time in attentional task Slower reaction time in attentional task
Decreased perceptual threshold Increased perceptual threshold
Enhanced ERP Attenuated ERP
Lower resistance to persuasion Greater resistance to persuasion
attention could be enhanced when subjects learned to decelerate their

hearts. These somewhat straightforward findings were explored in a
third study (Cacioppo, Sandman, & Walker, 1978).
One of the implications of narrowed external attention, related to
controlled deceleration, may be the sacrifice of reflection and judgment.
This possibility was investigated in a study of the influence of learned
heart rate control on resistence to persuasion. Resistence to persuasion
is, in part, a function of the number of counterarguments subjects can
generate to rebut an issue. Evidence indicates that the greater the num-
ber of reasons (counterarguments) subjects develop to reject a message,
the less likely it is that they will be persuaded by the message (Petty &
Cacioppo, 1977). We reasoned that increases in heart rate would enhance
cognitive processing. Accordingly, increased heart rate should result in
more counterarguments and lowered susceptibility to persuasion. Con-
versely, subjects may become more susceptible to persuasion during
lowered heart rate as their ability to process information is compromised.
Thus, control of heart rate could influence attitudes.
As predicted, during cardiac acceleration subjects generated more
counterarguments and were less willing to endorse the persuasive mes-
sage. However, during reinforced heart rate deceleration, subjects pro-
vided fewer counterarguments and were much more susceptible to per-
suasion. Thus, during high heart rate, cognitive elaboration may be
facilitated and self-convincing arguments can be generated against a
persuasive message. However, during lowered heart rate our cognitive
functioning may give way to environmental attention and the demands
of the environment prevail. These studies and others, most notably
those from David Shapiro's laboratory (see Shapiro & Reeves, 1982),
suggested that influences on the brain and behavior may be exerted by
controlling the body.
The successful clinical application of this "Clockwork Orange" pos-
sibility was reported with a case of a child-molestor (Nolan & Sandman,
1978). This behavioral aberration (pedophilia) has remained immune
CARDIAC AFFERENT INFLUENCES ON CONSCIOUSNESS 65
from most forms of treatment and the consequences both for the patient
and society justify a radical approach. We introduced the method "bio-
syntonic therapy" to consider jointly the synchrony among physiological
systems and the correspondence between thoughts or attitudes and the
physical state of the organism. That is, we were concerned with the
syntony among verbal (mental), physiological, and gross behavioral as-
pects of the human experience. A major assumption of our approach
was that attitudes, emotions, and behavior were intimately related to
physiological state. From this perspective, the work of the therapist was
carefully to identify (diagnose) the physiological state of the individual
in specific situations before attempting to intervene. Once the relation-
ship between mental and physical activity was observed in nonproblem
areas, then the problem areas were searched for evidence of dishar-
mony. The goal was to target physiological systems in the problem area
that appeared disparate with the patterns of responses in the more nor-
mal areas of the client's life. Once diagnosed, the biosyntonic position
implies that if the physiological pattern changes, a change in attitude,
emotions, or overt behavior will follow.
Mr. J., a 32-year-old blue-collar worker, had a history of numerous
sexual experiences with children (nearly all females) dating from his
teens. Fearful of the personal and legal consequences of the discovery
of his experiences with a female child, he consented to the procedures
described below as a last resort. Traditional approaches had been tried
without success.
Preliminary diagnostic information indicated extreme sexual attrac-
tion for prepubescent females, although Mr. J. maintained an adequate
sexual relationship with his wife. An extensive pretreatment physio-
logical assessment was conducted by attaching electrodes for the mea-
surement of heart rate, peripheral vasomotor activity, respiration, and
skin potential. Mr. J. reclined in a comfortable chair and viewed a stan-
dard sequence (Sandman, 1975) of pictures presented on a screen in
front of him. The series included sexually arousing, neutral, and highly
distressing (e.g., mutilated corpses) slides. He rated each of them on a
pleasure-stress scale while a physiological recording was done.
Mr. ] . exhibited differential physiological responses to pleasurable,
neutral, and unpleasurable stimuli only in heart rate. We focused on
heart rate in a discriminate conditioning paradigm used throughout 16
sessions; the conditioning was designed to alter his physiological re-
sponses to female children without disrupting his responses to adult
females.
Initially, four sets (male and female adults and male and female
children) of six stimuli each were chosen. Mr. J. was asked to bring in
pictures of preadolescent girls that he found arousing to varying de-
66 CuRT A. SANDMAN
grees. Prevalent among the most arousing stimuli were pictures taken
from mail-order catalogues of girls modeling underwear. Upon request
he also brought comparable pictures of preadolescent boys, though he
claimed that none of these were highly arousing. He selected adult male
and female stimuli from a group of slides available at the treatment
center. Mr. J. rated the stimuli in each of the four sets in terms of his
perceived sexual arousal to them. Ratings were based on the 9-point
pleasure-stress scale described earlier.
The most striking feature of his physiological response pattern was
the extremely elevated heart rate response to the pictures of the semi-
clothed female children. However, unlike his response to nude adult
females, his verbal report of arousal to the slides of children was not
consistent with his physiological response to them. Neither his verbal
nor his physiological responses suggested arousal to male children or
male adults. Therapy was therefore designed to accomplish both the
reduction of the inappropriate arousal to female children and improve-
ment of the congruity between the patient's verbal and physiological
responses to such children.
Treatment was conducted in three phases. In phase 1, aversive stim-
ulation (4-6 rnA shock) was delivered each time Mr. J.'s heart rate
exceeded a criterion in response to pictures of young girls. After several
sessions, a significant reduction in heart rate was observed. Further, Mr.
J. reported that during his daily activities he felt a sharp pain in his index
finger (where shock was applied) aborting his "automatic arousal" to
young girls he encountered. New, provocative stimuli were introduced
and elements of Mr. J.'s response reemerged.
In phase 2, a more powerful approach was initiated. In addition to
punishment for increased heart rate during exposure to young girls,
positive reinforcement (monetary incentive) was provided for heart rate
increases to pictures of adult women. This approach was successful, but
the potential contamination and lack of generalizability of aversive pro-
cedures dictated phase 3. In this final phase, positive reinforcement
remained contingent upon heart rate acceleration to adult females and
was also presented for inhibition of acceleration to pictures of young
girls. This procedure produced robust discrimination and was continued
for several sessions after which therapy was terminated.
Six months after treatment ended, Mr. J. reported that he had not
experienced any incidents of child molesting. He remarked that when
he was in a situation formerly considered provocative, he still experi-
enced some pain in his index finger. Of more importance, he consciously
avoided compromising situations, indicating that he was able to control
his behavior much more efficiently than before. It was of interest to learn
CARDIAC AFFERENT INFLUENCES oN CoNSCIOUSNESS 67
that Mr. J. attended X-rated moves at a far higher rate than he had before
therapy. Thus, a socially condoned activity had apparently replaced his
psychopathic and deviant behavior.
This report indicates that processes ascribed to the brain, thoughts,
attitudes, and behavior, can be accessed by control of the body. The
three experiments from our laboratory, the case. report, and the growing
body of biofeedback research and treatment indicate that consciousness
has many components, one of which is the interaction between the brain
and the body.
B. Effects on Consciousness of Transient Changes in the

Cardiovascular System
In addition to the purely correlative studies and to those relating
control of physiological systems to behavior, there has been a series of
reports of changes in awareness or perception synchronized with the
naturally occurring rhythms of the body. As reviewed earlier, similar
baroreceptor discharge occurs during changes in heart rate and pulsatile
pressure waves, both of these mechanisms cooperate in providing ho-
meostatic control of blood pressure and may exert similar influences on
behavior. Even though the Laceys (1967, 1970, 1974) and others (Birren,
Cardon, & Phillips, 1963; Callaway & Layne, 1964; Saari & Pappas, 1976)
reported that attention was enhanced when stimuli were synchronized
with each component (i.e., the P wave) of the EKG, other investigators
(Delfini & Campos, 1972; Elliot & Graf, 1972; Thompson & Botwinick,
1970) have not found a significant relationship between these transient
events and behavior. In an attempt to unravel the differences among
these studies and to provide a link with our studies of heart rate, we
(Sandman, McCanne, Kaiser, & Diamond, 1977) constructed a paradigm
(see Figure 3) in which tachistoscopic stimuli were synchronized with
the P, R, and T wave of the EKG or with decelerating, midrange, and
accelerating heart rate.
As summarized in Table I, we discovered that stimuli presented
during the P wave or during decelerating heart rate were perceived more
accurately than stimuli presented during the T wave or during accel-
erating heart rate. In a second study (Sandman, Walker, & Berka, 1982)
subjects were shown either a circle or a square synchronized with car-
diovascular events and instructed to depress a telegraphic key when the
circle appeared. In this case, with a decisional paradigm, we expected
that cognitive processes would be facilitated with accelerating heart rate.
68 CuRT A. SANDMAN
COMPUTER
.-- v. . .-. ;A'1''--

R STIMULATION
- -- -:..':·=-"A-~T
-~ OR
FEEDBACK
EKG
-- --A--- 1'1--
-- -/---6'---~
PULSE PRESSURE WAVE ,............._.........,
-----_:ri~--rF-
4-------¥-- POLYGRAPH
TACHOGRAPH
FIGURE3. Paradigm used in studies of transient changes in heart rate, interbeat interval,
and pulse-pressure phase on the brain and behavior.
The findings indicated that during accelerating heart rate, reaction time
decrea·sed (thus decision making was facilitated) and, conversely, during
decelerating heart rate, reaction time slowed. However, contrary to our
studies of stimulus detection, there was no corresponding influence of
the cardiac cycle on choice reaction time. These differences may be due
to (1) the imperfect coupling of heart rate and cardiac phase influences
on the brain (2) the comparative insensitivity of reaction time compared
with stimulus detection or (3) the robust influence of heart rate compared
with the seemingly delicate impact of events during the cardiac cycle.
In any case, a significant array of evidence has accrued linking be-
havior to the cardiovascular system. In addition to the data reviewed
above, other evidence from our laboratories (Cacioppo & Sandman,
1978; Kaiser & Sandman, 1975; Sandman & Walker, 1985; Walker &
Sandman, 1977), as well as volumes dedicated to this subject (Cacioppo
& Petty, 1982; Obrist, Black, Brener, & DiCara, 1974) attest to the grow-
ing acceptance and importance of these observations. These data are
increasingly difficult to reconcile with classical views ascribing all attri-
butes of consciousness to the brain and require that we reevaluate the
mind-body debate. However, despite the overwhelming evidence of
the influence of the cardiovascular system on behavior very few exper-
imental studies have investigated the impact of the heart on the brain.
CARDIAC AFFERENT INFLUENCES ON CoNSCIOUSNESS 69
IV. INFLUENCE OF CARDIOVASCULAR SYSTEM ON THE BRAIN
A. The EEG
The relationship of the cardiovascular system to the electrical ac-

tivity of the brain in human subjects has been well documented. The
early report of Obrist and Bissell (1955) suggested that changes in pos-
ture compromised cerebral blood flow and was reflected in the EEG.
The studies of Ingvar (1972) extended these principles and demonstrated
that increased perfusion was related to increased arousal of EEG pat-
terns. The early studies of Callaway (Callaway & Buchsbaum, 1965; Cal-
laway & Layne, 1964) demonstrated synchronous relations between the
ventricular contractions of the heart and the "ascending" wave of the
alpha rhythm. More recently a fascinating influence of cardiovascular
disease on the brain has been reported. Sotaniemi (1980) observed 100
consecutive righthanded patients referred for valvular replacement sur-
gery. Preoperatively, neurological and EEG findings were significantly
more frequent in the left hemisphere. However, postoperatively the
right hemisphere was compromised significantly more frequently.
B. Visual Event-Related Potentials

In order to probe the influence of the heart on the brain, we (Walker
& Sandman, 1979) studied event-related potentials (ERPs) of the brain.
The ERP is electrical activity recorded from the scalp which is time-locked
to a specific stimulus. In order to obtain an evoked response, flashes of
light, brief tones, or electrical shock are delivered to subjects while the
electroencephalogram (EEG) is measured. The electrocortical responses
to each stimulus are summed and the activity which is not related to
the stimulus is random and equals approximately zero. The components
of the wave which are evoked by the stimulus are then clearly identi-
fiable against this background. The form of the ERPs recorded in our
laboratory are presented in Figure 4. Different components have been
associated with discrete behaviors. The early aspect of the wave, the
primary response, is related to the dimension of the stimulus and is
usually unchanged by psychological or physiological interventions. The
secondary components (occurring approximately 100 and 200 msec after
the stimulus) are sensitive to changes in attention and perception. Larger
P1 and P2 waves are usually related to improved attentional or percep-
tual processing. The waves occurring approximately 300 msec after the
stimulus are termed endogenous because they are related to complex
cognitive activity and not the physical characteristics of the stimulus.
70 CuRT A. SANDMAN
STIMULUS
+
VISUAL
ERP
AUDITORY
ERP
0 50 100 150 200 250 300 350 400
FIGURE 4. Examples of visual and auditory event-related potentials.
In our first experiment (Walker & Sandman, 1979) brief flashes of

light were presented to the subject during fast, slow, and midrange heart
rates (as in earlier experiments, Figure 3), while ERPs were recorded
from the right and left hemispheres of the brain. This paradigm relates
most directly to changes in early exogeneous components since the sub-
jects were not required to think or make any behavioral response.
The results indicated that the heart exerted dramatic influences on
the ERPs of the right but not the left hemisphere (Figure 5). Furthermore,
the amplitudes of Pl and P2 were larger when flashes of light were
synchronized with low heart rate than with high heart rate. Although
it may be peculiar to the paradigm we used, it was apparent that the
influence of the heart on the brain was lateralized.
As discussed earlier, each ventricular pulse of the heart sends a
bolus of blood rushing through the arteries. The bolus of blood is re-
ceived as a transient shift in pressure and is reflected as a pulse pressure
wave. At the peak of the wave (systolic pressure), pressure is the greatest
and the baroreceptors increase their firing rate. During the lowest point
of the wave (diastolic pressure) the baroreceptors are relatively quies-
cent. Since pulse pressure is intricately tied to the pumping action of
the heart, systolic and diastolic pressure coincide with electrical com-
LEFT HEMISPHERE RIGHT HEMISPHERE
I II I I!!! t " ! I ' t •• !! f! I It If

0 100 200 300 400 0 100 200 300 400
msec
FIGURE 5. Event-related potentials recorded from the left and right hemisphere of the
brain when visual stimulation was synchronized with accelerating and decelerating heart
rate. Arrows indicate points of statistical preparation as determined by stepwise discrim-
inant function analysis. Key: (---) high heart rate; (-·-) midrange heart rate; (-)low
heart rate.
ponents of the electrocardiogram (ECG). From Figure 2 it is apparent

that diastolic pressure coincides with the early components of the EEG
(P wave) and systolic pressure coincides with the T wave. Therefore,
as with the earlier studies (Sandman et al., 1977; 1982) the influence of
transient changes in pulse pressure on the ERP was measured. Since
similar (but not identical) pressure changes are detected by the baro-
receptor during low heart rate and diastolic pressure and during high
heart rate and systolic pressure, similar influences on the brain were
expected in this study as with the study of heart rate. We reasoned that
the effects of pulse pressure changes on the brain were "time-locked"
and displacement of stimulation in time may alter the impact of the heart
on the brain. Therefore, flashes of light were synchronized with changes
in pulses in the brain (occurring approximately 30 msec after carotid
pulse changes) and the finger (occurring approximately 180 msec after
carotid pressure changes; Walker & Sandman, 1982).
ERPs synchronized with the diastolic phase of the pulse pressure
waves recorded from the carotid and cephalic placements were differ-
entiated only in the right hemisphere (Figure 6). Components occurring
200-300 msec after stimulation were enhanced when synchronized with
diastole. However, the major effect of stimulus synchronization with
72 CuRT A. SANDMAN
LEFT HEMISPHERE RIGHT HEMISPHERE

r.
(\.' \ CAROTID
I I
I I
f \
: ',
\_..,. ... '""\
' .. r ... ,,_
-Systolic pressure (peaks)

--- Diastolic pressure (valleys) DIGITAL
0 100 200
FIGURE 6. Event-related potentials to visual stimuli synchronized with systole and diastole
records from the carotid, opthalmic, and digital arteries.
digital placements was observed in the left hemisphere, and augmen-

tation of the ERP occurred during the systolic pulse. The finding that
the ERP was altered when stimuli were presented at different phases
of the cardiac cycle argued for the importance of the temporal relations
between cardiovascular events and the brain. Cardiovascular changes
detected by pressure receptors in the carotid artery, and possibly the
vasculature of the brain, sensitized the brain to environmental stimu-
lation within very narrow time limits. Further, with the paradigm em-
ployed, this sensitization was lateralized in the right hemisphere of the
brain during stimulation synchronized with diastole. This effect shifted
to the left hemisphere when stimuli were triggered by systole in distal
placements.
The heart rate and pulse pressure wave experiments suggested sim-
ilar effects of these interdependent cardiovascular processes on the
CARDIAC AFFERENT INFLUENCES ON CoNsciousNEss 73
LEFTHEMISPHERE RIGHTHEMISPHERE
6 HIGH HEART RATE 6 HIGH HEART RATE
4
,,,
~ ,,
_j
I\ I I
I \ I \
··I 2 1\/\
I I
12.5 .uv I \ ... 2.5 .uv
-~ ',,.,'~- .
. , ___ .. I '
'~
, __ _ I
-2 - .........
-4~ -4
6 LOW HEART RATE 6 LOW HEART RATE

4 4
2.5 .uv
-
' ... ...
-Systolic pressure (peaks),, __ ,
~4 ---Diastolic pressure (valleys)
FIGURE7. Event-related potentials of the brain when visual stimulation was synchronized
with changes in heart rate and pulse pressure phase.
brain. For instance, stimulation during both lowered heart rate and dias-
tole resulted in amplification of early components of the ERP. Further,
the major influence of these cardiovascular events was unique to the
right hemisphere of brain. The remarkable similarity between the find-
ings of the heart rate and pulse pressure experiments prompted a third
experiment (Walker & Sandman, 1982), to examine the effects on the
ERP of the interaction between heart rate and pulse pressure waves.
This experiment was identical to the previous experiments except that
presentation of the stimulus was contingent upon changes of heart rate
and the pulse pressure wave.
As in the previous experiments, the major effects of stimuli syn-
chronized with heart rate and pulse pressure changes were observed in
the right hemisphere of the brain (Figure 7). Only during accelerating
heart rate, in the right hemisphere, was the difference between systole
and diastole reflected in the ERP. The most significant difference among
components during accelerating heart rate was enhancement of Pl dur-
ing stimuli synchronized with diastole. A similar, but not significant,
trend was apparent during low heart rate.
The results of this study indicated that complex relations exist be-
tween heart rate and the pulse pressure wave. These components do
not appear to be simply additive. Indeed, if they were additive, the ERP
would be maximal for stimuli synchronized with diastole and low heart
74 CuRT A. SANDMAN
rate. This was not found. Instead, suppression of the ERP was observed
during high heart rate and systole (although only in the right hemi-
sphere). The major finding (that maximal separation of systole and dias-
tole was observed only during elevated heart rate) suggested that during
elevated heart rate, maximal (or near maximal) activity of the barore-
ceptors is achieved during each beat of the heart (Sandman et al., 1982).
It is possible that this barrage "exhausts" the cells temporarily, resulting
in significant quiescence during diastole. Thus, the greatest contrast
between systole and diastole may be during elevated heart rate.
The lateralized influence of stimuli synchronized with diastole or
systole may not be surprising. For instance, there is evidence that simple
attentional tasks selectively impact the right hemisphere and as the com-
plexity of the task increases the left hemisphere is engaged (Dustman,
Schenkenberg, & Beck, 1976; Jutai, 1984). Similarly, the early compo-
nents of the ERP reflect exogenous or stimulus-related dimensions,
whereas the later components are thought to be excited by increasing
the work load (Donchin, 1979). This complementary relationship also
pertains to the cardiovascular system since acceptance of the environ-
ment (attention) is related to lowered heart rate or diastolic components
of pulse pressure waves and cognitive processing is facilitated by in-
creases in heart rate or during systole. Consistent with these indepen-
dent observations, our findings indicate that during simple tasks the
early components of the ERP are enhanced in the right hemisphere when
stimuli are synchronized with the diastolic phase of the pulse pressure
wave. As the task demands are increased, later components of the ERP
are augmented in the left hemisphere when stimuli are synchronized
with systole.
Rogers, Battit, McPeek, and Todd (1978) provided dramatic phys-
iological evidence of the lateralization of cardiac control. Stellate gan-
glion blocks of 17 patients indicated that the right stellate ganglion had
a much greater influence on heart rate than does the left stellate gan-
glion. Further, since the cerebral blood vessels are innervated by ad-
renergic fibers originating in superior cervical and stellate ganglion
(Purves, 1978), a lateralized influence may be favored. Coupled with
the report of Sotaniemi (1980), already discussed and the findings of
Carmon and Gombus (1970) that blood from the heart reaches the right
carotid faster than the left, these findings provide a plausible (but un-
tested) explanation for the findings of lateralized influence of the heart
on the electrical activity of the brain.
C. Auditory Event-Related Potentials
In a recent study (Sandman, 1984) with a design identical to that

of the previous studies of the visual evoked potential, auditory evoked
STIMULUS
SYSTOLE
SYNCRONIZED
DIASTOLE
SYNCRONIZED
LEFT HEMISPHERE
------
It
RIGHT HEMISPHERE
IOp.V
-100 0 100 200 300 400

m sec
FIGURE 8. Auditory event-related potentials of the brain synchronized with pulse pres-
sure phase.
responses were collected during stimuli synchronized with systolic and

diastolic phases of the cerebral pulse pressure wave. Furthermore, at-
tentional demands were assessed by comparing a passive condition with
a condition in which subjects counted and reported the number of stim-
uli they detected. The EEG was measured from C3 and C4 referenced
to linked mastoids. Stimuli presented during the diastolic phase of the
pulse pressure wave elicited a significantly larger N1 component of the
ERP than presentations during systole (Figure 8}, a finding which was
perfectly consistent with the results for the visual event-related
potential.
The subjects were divided into correct and incorrect responders
based upon the accuracy of their report. Presumably, subjects whore-
ported accurately were more attentive than subjects who offered incor-
rect responses. This analysis confirmed that Nl was augmented in cor-
rect responders and during diastole. However, stimuli synchronized
with systole resulted in a larger N2 component than stimuli synchro-
nized with diastole. Further, a complex interaction indicated that N2
was generally suppressed in the right hemisphere of accurate respond-
ers when stimuli were synchronized with diastole.
As reviewed by Picton and Hillyard (1974) and Picton, Hillyard,
Krausz, and Galambos (1974), the N1 and P2 components of the auditory
76 CURT A. SANDMAN
ERP reflected selectivity of attention. Although Picton and Hillyard did

not differentiate Nl from P2, and indeed suggested that they accrued
from similar neural assemblies, the results of the current study sug-
gested that they may reflect subtly different processes. For instance,
robust changes were seen in Nl but not P2 when stimuli were syn-
chronized with different phases of the pulse pressure wave. Consistent
with their impression was the recent report of Hansen and Hillyard
(1983) indicating a protracted negativity (Nd) beginning at Nl and con-
tinuing for several hundred msec associated with selective attention and
cognitive processing.
Although the suppression of N2 in accurate responders was con-
sistent with the analysis made by Picton et al., (1974), the significance
of cardiac phase on N2 in the left hemisphere in the subjects who re-
ported accurately was not immediately obvious. These investigators sug-
gested that N2 is influenced by the subject's state of alertness, such that
it w!ls attenuated in alert subjects compared with subjects who were
drowsy. As such, suppression of N2 may represent cortical activation.
Thus, attentive subjects evidenced a cortical index of activation for stim-
uli synchronized with diastole. The suppression of the N2 component
in the right hemisphere during diastole complemented the findings of
enhanced early components in the right hemisphere. It is remarkable
that ERP components thought to be related to dimensions of conscious-
ness were altered by internal events in such a consistent manner. These
findings lend further support to the proposal that a continuum of con-
sciousness may be regulated by covariation of the cardiovascular system
and the brain.
These data suggest that awareness of the environment is partially
regulated by the interactions of the brain and the heart. Although it is
probable that many visceral and somatic systems "tune" the brain
through their afferent pathways, the results of the present study provide
direct information of this tuning by the cardiovascular system. One spec-
ulative conclusion implied by these findings is that a modest portion of
environmental awareness or attention is "hard-wired." For instance,
with a task chosen purposefully for its simplicity, a reliable influence
on the ERP was observed. Even when attentional demands were im-
posed, or in subjects who either were attentive or inattentive, the same
effect on the ERP was evident. Thus, a portion of the variance (10-20%
based on the present data) relating to attention (Sandman et al., 1977),
or to the components of the ERP, may be accounted for by peripheral
physiological systems (or any background activity) which may be in-
violably linked to the brain.
Another speculative conclusion derived from these data is that there
are optimal physiological "windows" for enhancing perception or at-
tention and that these windows are cyclic. Certainly, the data from the
present study and others (Sandman et al., 1977; Saari & Pappas, 1976;
Birren et al., 1963; Walker & Sandman, 1979, 1982) indicated that there
are precise periods during the cardiac cycle or the carotid pressure gra-
dient that optimize perception and the impact of stimulation on the
brain. It is conceivable that the window of optimal performance is the
result of either fortuitous or purposeful synchrony among physiological
systems. Although resolution of this cause-and-effect question must be
deferred, the findings clearly indicate that there are cyclic physiological
states which modulate the impact of external stimuli.
D. Evoked Vascular Responses (EVRs)

Data reviewed earlier indicated that changes in the cerebral vas-
culature covary with electrical activity recorded from the scalp. Kliving-
ton and Galambos (1967) estimated that blood contributes 10% to the
conductivity of the cortex. This estimate agrees with the 10-20% aug-
mentation observed in our studies of cardiovascular phase influences
on the ERP. Increasingly sophisticated studies indicate predictable re-
lationships between cerebral flow or metabolic activity and behavior.
For instance, Willison et al. (1980) reported that patients with elevated
venous hematocrit (VH) do poorly on simple tests of alertness. When
the VH was lowered in these same subjects, performance improved.
Since elevated VH relates to increased blood viscosity, decreased cer-
ebral blood flow (CBF) results (confirmed by Xenon-133 inhalation).
With a more invasive procedure, LeDoux et al. (1983) mapped met-
abolically active (increased blood flow) areas of the rat brain. Using a
rapid index (146-iodoantidyrine metabolized in 30-40 seconds compared
with 30-45 minutes for 2-deoxy-D-glucose required with positron emis-
sion tomography), these authors reported changes in brain structures
associated with primary processing of stimuli as well as areas reflecting
conditioned emotional responses. Thus, discrete and significant changes
in brain metabolism occur during processing of environmental events.
The relationship of these changes to the ERP was investigated in our
laboratory (Sandman et al., 1984) with the rheoencephalogram (REG; an
impedance-based measure of blood volume). Phasic changes in blood
volume elicited by auditory stimulation were measured with the REG
and adapted to our ERP procedures. The pulse pressure wave from the
REG was averaged in response to auditory stimuli and compared with
the pulse-pressure wave occurring just prior to the stimulus. These two
values were subtracted on a trial-by-trial basis resulting in a wave rep-
resenting the stimulus-induced response (Figure 9).
78 CuRT A. SANDMAN
PULSE VOLUME WAVES
PRESTIMULUS
STIMULUS
I
,, ' 'I
I ''
I '
I
~-'
Mean Difference
Response 10%Changeo I
- 2 56 0 512 1024 -256 0 512 1024

msec msec
FrcuRE 9. Top: Wave form from individual subject indicating prestimulus and stimulus
waves. Bottom: Averaged difference waves for stimuli synchronized with systole and
diastole.
During diastole four temporally adjacent factors of the REG were

identified by principal component analysis. The first factor (which ac-
counted for the most variance) comprised the sampling epoch between
300 and 540 msec. The second (40 to 100 msec), third (100 to 200 msec)
and fourth (200 to 300 msec) factors completed the description of the
wave form. These factors are consistent with the average of the indi-
vidual subject waves (Figure 10). The factor structure was not inter-
pretable for waves evoked during systole. Thus, the evoked vascular
response (EVR) can be characterized by factor analytic procedures when
synchronized with diastole but not with systole.
This is the first report of evoked changes in a measure of the cerebral
vasculature from conscious human subjects. The EVR was evident only
when stimulation was synchronized with the diastolic phase of the
ophthalmic artery. Its latency defies the time course of previously de-
scribed vascular changes in response to altered metabolic activity. Thus,
VASCULAR EVOKED RESPONSE
STIMULUS STIMULUS
~ CORTEX + ARM
-256 0 512 1024 -256 0 512 1024

msec msec
FIGURE 10. Average difference bioimpedance waveforms in cortex (N = 10) and arm (N
= 8). Only diastole synchronized waves are presented.
this response may be a neurogenically mediated vascular event in prep-

aration for altered metabolic demand.
Enhancement of the EVR during diastole is consistent with our pre-
vious reports and indicates that rapid changes in blood volume occur
according to a similar schedule and of similar proportion to the ERP.
The temporal similarities between response of the cerebral vasculature
and the electrical response of the brain may suggest common generating
mechanisms. For instance, decreased baroreceptor activity (as during
diastole) is associated both with increased cerebral blood flow and re-
lease from neuronal inhibition. The EVR may prove to be a neurally
80 CURT A. SANDMAN
mediated phenomenon that not only yields clinically relevant infor-

mation regarding the integrity of vasomotor systems but also may pro-
vide new information of brain metabolism related to detecting and pro-
cessing external information in real time.
V. CoNcLusioNs
This program of research indicated that: (1) there are relationships

between the cardiovascular system and behavior; (2) control of cardio-
vascular system alters behavior; (3) predictable relationships exist be-
tween the heart and the brain, and (4) one possible mechanism for heart-
brain interaction is cerebrovascular dynamics. These findings provide
preliminary support for the radical neo-Jamesian proposal that con-
sciousness is the interaction among physiological systems. This proposal
is short of the heretical views of Kennedy (1959). Kennedy proposed
that synchronous brain activity was an artifact induced by the mechan-
ical energy from the ventricular contraction of the heart (Bering, 1955).
This "pulse," applied to a gelantinous mass (the brain), in a limited,
closed container (the skull) initiated and sustained brain activity by
means of the cerebrospinal fluid. Consciousness, or the frequency of
oscillation, was related to the "consistency" of the brain and modified
by the cerebral vasculature. Thus, during attention, the brain was en-
gorged with blood, detuning the oscillators and blocking synchronous
activity. This provocative proposal has not been tested thoroughly, but
the studies reviewed above suggest that even transient changes in the
oscillator can alter consciousness.
If, as Kennedy implies, there are rhythms in the body which are
subservient to, or are driven by, other rhythms, one is left with the
question, What is the basic rhythm? Rephrased, the question may be-
come, Is consciousness ultimately regulated by the lowest common de-
nominator? Bentov (1977) proposed five basic rhythms including un-
detectable (or unnoticeable) body rhythms which were kindled,
entrained, or driven by the basic rhythm of the planet Earth. The
rhythms we commonly suspect as the carriers of consciousness, Bentov
believed, derived from these primitive sources. Supportive observations
by Corner (1977) indicated that primitive activity bursts coupled with
the EEG dominated the behavior of lower organisms. Corner believed
these relationships were vestigial in higher nervous systems and were
expressed during early ontogenesis and perhaps during states of low-
ered consciousness such as sleep (when muscle bursts and EEG are
tightly coupled). Thus, when consciousness is lowered, patterns of an-
cient phylogenetic origin can be expressed.
Recent, but limited, observations in our laboratory extended this

thesis. A variation of our ERP procedure is simply to turn the stimulus
off and observe the relationship between the pulse pressure gradient
and the electrical activity of the brain (Sandman et al., 1982). In normal
subjects, there was no evidence of coupling. However, in six depressed
patients a striking relationship was observed between the cardiovascular
system and the brain. A 9Hz wave (alpha rhythm) was perfectly coupled
with the pulse pressure wave. The ascending (positive-going) wave of
the alpha rhythm was synchronized with systole and the decending
wave was related to diastole. A very similar finding was reported re-
cently by Walker and Walker (1983) for normal subjects selected for high
levels of synchronous (alpha) activity.
Even though this analysis suggests that the hierarchy of conscious-
ness may be a function of a family of rhythms, the issue may not be
that obvious. The passive synchronization of physiological rhythms ap-
pears to be pathological, and as the lowest driving rhythm is achieved,
consciousness follows. However, as we have shown, there are temporal
windows in these rhythms which, when properly probed, raise aware-
ness or consciousness. Further, the temporal parameters governing this
process exquisitely channel the influence of the heart between liberating
and suppressing the left and right hemispheres of the brain. Thus, there
are active synchronicities which are adaptive and precise and which
ensure optimal performance. Perhaps the task of phylogenesis and on-
togenesis is decoupling the passive process. Only as the organism be-
comes free of primitive "rhythms" or passive synchronization will
higher or optimal states be achieved. Thus, reality becomes experienced
as filtered through the waves and rhythms of our potential.
REFERENCES
ADAIR, J. R., & MANNING, J. W. (1975). Hypothalamic modulation of baroreceptor afferent
unit activity. American Journal of Physiology, 229, 1357-1364.
BENTOV, I. (1977). Micromotion of the body as a factor in the development of the nervous
system. Appendix to Stalking the wild pendulum. New York: Dutton.
BERING, E. A. (1955). Choroid plexus and arterial pulsation of cerebral fluid. Archives of
Neurology and Psychiatry, 73, 165-172.
BIRREN, J., CARDON, P., & PHILLIPS, S. (1963). Reaction time as a function of the cardiac
cycle in young adults. Science, 140, 195-196.
BIRZIS, L., & TACHIBANA, S. (1964) Local cerebral impedance and blood flow during sleep
and arousal. Experimental Neurology, 9, 269.
BoNVALLET, M., DELL, P., & HIEBEL, G. (1954). Tonus sympathetique et activite electrique
corticale. Electroencephalography and Clinical Neurophysiology, 6, 119.
CACIOPPO, J. T., & PETrY, R. E. (1982). Perspectives in cardiovascular psychophysiology. New
York: Guilford Press.
82 CuRT A. SANDMAN
CACIOPPO, J. T., & SANDMAN, C. A. (1978). Physiological differentiation of sensory and

cognitive tasks as a function of warning, processing demands and reported unpleas-
antness. Biological Psychology, 6, 181-192.
CACIOPPO, J. T., SANDMAN, C. A., & WALKER, B. B. (1978). The effects of operant heart rate
conditioning on cognitive elaboration and attitude change. Psychophysiology, 15, 330-
338.
CALLAWAY, E., & BucHSBAUM, M. (1965). Effects of cardiac and respiratory cycles on av-
eraged visual evoked potentials. EEG and Clinical Neurophysiology, 19, 476-480.
CALLAWAY, E., & LAYNE, R. (1964). Interaction between the visual evoked response and
two spontaneous biological rhythms: The EEG alpha cycle and the cardiac arousal
cycle. Annals of the New York Academy of Sciences, 112, 424-431.
CANNON, W. B. (1929). Bodily changes in pain, hunger, fear and rage: An account of recent
researches into the function of emotional excitment. New York: Appelton.
CARMON, A., & GoMsus, G. M. (1970). A physiological vascular correlate of hand pref-
erence: Possible implications with respect to hemispheric cerebral dominance. Neu-
ropsychologia, 8, 119-128.
COLERIDGE, H. M., CoLERIDGE, J. C. G., & RosENTHAL, F. (1976). Prolonged inactivation of
cortical pyramidal tract neurons in cats by distension of the carotid sinus. Journal of
Physiology, 256, 635-649.
CoRNER, M. A. (1977). Sleep and the beginnings of behavior in the animal kingdom: Studies
of ultradian motility cycles in early life. Progress in Neurobiology, 8, 279-295.
DELFIN!, L., & CAMPos, J. (1972). Signal detection and the "cardiac arousal cycle." Psycho-
physiology, 9, 484-491.
DoNCHIN, E. (1979). Event-related brain potential: A tool in the study of human information
processing. In H. BEGLEITER (Ed.), Evoked brain potentials and behavior. New York:
Plenum Press, pp. 13-88.
DusTMAN, R. E., ScHENKENBERG, T., & BEcK, E. C. (1976). The development of the evoked
response as diagnostic and evaluative procedure. In R. KARRER (Ed.), Developmental
psychophysiology of mental retardation. Springfield, IL: Charles C. Thomas.
ELLIOTI, R., & GRAF, V. (1972). Visual sensitivity as a function of phase of cardiac cycle.
Psychophysiology, 9, 357.
FoRSTER, A., & SToNE, T. W. (1976). Evidence for a cardiovascular modulation of central
neuronal activity in man. Experimental Neurology, 51, 141-149.
GAHERY, Y., & VIGIER, D. (1974). Inhibitory effects in the cuneate nucleus produced by
vaga-aortic afferent fibers. Brain Research, 75, 241-246.
HANSEN, J. C., & HILLYARD, S. A. (1983). Selective attention to multidimensional auditory
stimuli. Journal of Experimental Psychology: Human Perception and Performance, 9, 1-19.
HARE, R. D. (1973). Orienting and defensive responses to visual stimuli. Psychophysiology,
10' 453-464.
HUMPHREY, D. R. (1967). Neuronal activity in the medulla oblongata of cat evoked by
stimulation of the carotid sinus nerve. In P. KEzm (Ed.), Baroreceptors and hypertension.
New York: Pergamon Press.
lNGVAR, D. H. (1972). Patterns of thought recorded in the brain. Totus Homo, 4, 98-103.
]AMES, W. Psychology. (1892). New York: Henry Holt.
joY, M. D. (1975). The vasomotor center and its afferent pathways. Clinical Science and
Molecular Medicine, 48, 253s-256s.
jUDY, W. V., & FARNELL, S. K. (1979). Arterial baroreceptor reflex control of sympathetic
nerve activity in the spontaneously hypertensive rat. Hypertension, 1, 605-614.
}UTA!, J. W. (1984). Cerebral asymmetry and the psychophysiology of attention. International
Journal of Psychophysiology, 1, 219-226.
KAISER, D. N., & SANDMAN, C. A. (1975). Physiological patterns accompanying complex
problem solving during warning and nonwarning conditions. Journal of Comparative

and Physiological Psychology, 89, 357-363.
KENNEDY, J. L. (1959). A possible artifact in electroencephalography. Psychological Review,
66, 347-352.
KLIVINGTON, K. A., & GALAMBOS, R. G. (1967). Resistence shifts accompanying the evoked
cortical response in the cat. Science, 157, 211-213.
KocH, E. (1932). Die Irradiation der pressoreceptorischen Krieslaufreflexe. Klinische Woch-
enschrift, 11, 225-227.
LACEY, B. C., & LACEY, J. I. (1974). Studies of heart rate and other bodily processes in
sensorimotor behavior. In P. A. 0BRJST, A. H. BLACK, J. BRENER, & L. V. DICARA(Eds.),
Cardiovascular psychophysiology. Chicago: Aldine.
LACEY, J. I. (1959). Psychophysiological approaches to the evaluation of psychotherapeutic
process and outcome. In E. A. RUBINSTEIN & M. B. PARLOFF (Eds.), Research in psycho-
therapy. Washington, DC: American Psychological Association.
theory. In M. H. APPLEY & R. TRUMBULL (Eds.), Psychological stress: Issues in research.
New York: Appleton.
LACEY, J. I., KAGAN, J., LACEY, B. C., & Moss, H. (1963). The visceral level: Situational
determinants and behavioral correlates of autonomic response patterns. In P. H.
KNAPP(Ed.), Expression of emotions in man. New York: International Universities Press.
LACEY, J. I., & LACEY, B. C. (1970). Some autonomic-central nervous system interrelation-
ships. In P. BLACK (Ed.), Phsysiological correlates of emotion. New York: Academic Press.
LAZARUS, R. S. (1966). Psychological stress and the coping process. New York: McGraw-Hill.
LAZARUS, R. S., SPEISMAN, J. C., MoRDKOFF, A.M., & DAVIDSON, L.A. (1962). A laboratory
study of psychological stress produced by a motion picture film. Psychological Mono-
graph, 76.
LEDoux, J. E., THOMPSON, M. D., IADECOLA, C., TucKER, L. S., & REIS, D. J. (1983). Local
cerebral blood flow increases during auditory and emotional processing in the con-
scious rat. Science, 221, 576-578.
LIBBY, W. L., LACEY, B. C., & LACEY, J. I. (1973). Pupillary and cardiac activity during visual
attention. Psychophysiology, 10, 270.
McCA.'\fNE, T. R., & SANDMAN, C. A., (1974). Instrumental heart rate responses and visual
perception: A preliminary study. Psychophysiology, 11, 283-287.
McCANNE, T. R., & SANDMAN, C. A. (1976). Operant autonomic conditioning and Rod-
and-Frame Test performance. Journal of Personality and Social Psychology, 34, 821-829.
NoLAN, J. M., & SANDMAN, C. A. (1978). "Biosyntonic" therapy: Modification of an operant
conditioning approach to pedophilia. Journal of Consulting and Clinical Psychology, 46,
1133-1140.
0BRJST, P. A., LIGHT, K. C., McCUBBIN, J. A., HuTCHESON, J. S., & HoFFER, J. L. (1979).
Pulse transit time: Relationship to blood pressure and myocardial performance. Psy-
chophysiology, 16, 292.
0BRJST, P. A., BLACK, A. H., BRENER, J., & DICARA, L. V. (Eds.). (1974). Cardiovascular
psychophysiology. Chicago: Aldine.
0BRJST, W. D., & BISSELL, L. E. (1955). The electroencephalogram of aged patients with
cardiac and cerebral vascular disease. journal of Gerontology, 10, 315-330.
PERSSON, B., & SvENSSON, T. H. (1981). Control of behavior and brain noredrenaline neu-
rons by peripheral blood volume receptors. Journal of Neural Transmission, 53, 73-82.
PETERSON, L. H. (1962). The mechanical properties of the blood vessels and hypertension.
In J. G. CoRT, V. FENCL, Z. HEJL, & J. ]IRKA (Eds.), The pathogenesis of essential hyper-
tension. Prague: State Medical Publishing House.
PETTY, R. E., & CACIOPPO, J. T. (1977). Forewarning, cognitive responding and resistance
to persuasion. Journal of Personality and Social Psychology, 35, 645-655.
84 CURT A. SANDMAN
PieroN, T. W., & HILLYARD, S. A. (1974). Human auditory evoked potentials. II: Effects
of attention. EEG and Clinical Neurophysiology, 36, 191-199.
PieroN, T. W., HILLYARD, S. A., KRAusz, H. I., & GALAMBOS, R. (1974). Human auditory
evoked potentials. 1: Evaluations of components. EEG and Clinical Neurophysiology, 36,
179-190.
PoNTE, J., & PuRVES, M. J. (1974). The neural control of cerebral blood vessels. In The
Physiology of the cerebral circulation. London: Cambridge University Press.
PURVES, M. J. (1972). The neural control of cerebral blood vessels. In The physiology of the
cerebral circulation. London: Cambridge University Press.
PuRVES, J. J. (1978). Do vasomotor nerves significantly regulate cerebral blood flow. Cir-
culation Research, 43, 485-493.
ROGERS, M. C., BATIIT, G., McPEEK, B., & ToDD, D. (1978). Lateralization of sympathetic
control of the human sinus node: ECG changes of stellate ganglion block. Anesthe-
siology, 48, 139-141.
SAARI, M., & PAPPAS, B. (1976). Cardiac cycle and phase and movement and reaction times.
Perceptual and Motor Skills, 42, 767-770.
SANDMAN, C. A. (1971). Psychophysiological parameters of emotion. Unpublished doctoral dis-
sertation. Louisiana State University.
SANDMAN, C. A. (1975). Physiological responses during escape and non-escape from stress
in field independent and field dependent subjects. Biological Psychology, 2, 205-216.
SANDMAN, C. A. (1984). Augmentation of the auditory event-related potential of the brain
during diastole. International Journal of Psychophysiology, 2, 111-119.
SANDMAN, C. A., McCANNE, T. R., KAISER, D. N., & DIAMOND, B. (1977). Heart rate and
cardiac phase influences on visual perception. Journal of Comparative and Physiological
SANDMAN, C. A., O'HALLORAN, J. P., & IsENHART, R. (1984). Is there an evoked vascular
response? Science, 224, 1355-1357.
SANDMAN, C. A., WALKER, B. B. (1985). Cardiovascular relationship to attention and think-
ing. In V. RENTAL, S. A. CARSON, & B. R. DuNN (Eds.), Psychobiology of reading. New
York: Pergamon Press.
SANDMAN, C. A., WALKER, B. B., & BERKA, C. (1982). Influence of afferent cardiovascular
feedback on behavior and the cortical evoked potential. In J. T. CACIOPPO & R. E. PETTY
(Eds.), Perspectives in cardiovascular psychophysiology, pp. 189-222. New York: Guilford
Press.
SHAPIRO, D., & REEVES, J. L. (1982). Modification of physiological and subjective responses
to stress through heart rate biofeedback. In J. T. CACIOPPO & R. E. PETTY (Eds.), Per-
spectives in cardiovascular psychophysiology. New York: Guilford Press, pp. 127-150.
SMITH, R. E., & PEARCE, J. W. (1961). Microelectrode recordings from the region of the
nucleus solitarius in the cat. Canadian Journal of Biochemical Physiology, 39, 933.
SoTANIEMI, K. A. (1980). Brain damage and neurological outcome after open-heart surgery.
Journal of Neurology, Neurosurgery and Psychiatry, 43, 127-135.
SvENSSON, T. H., & THOREN, P. (1979). Brain noradrenergic neurons in the locus coeruleus:
Inhibition by blood volume load through vagal afferents. Brain Research, 172, 174-178.
TACHIBANA, S., KuRAMOTO, S., INANACA, K., & IKEMI, Y. (1967). Local cerebrovascular re-
sponses in man. Confinia Neurologica, 29, 289.
THOMPSON, F. J., LERNER, D. N., FIELDS, K., & BLACKWELDER, A. (1980). Projection of limb
versus afferents to the feline motor-sensory cortex. Journal of the Autonomic Nervous
System, 2, 39-45.
THOMPSON, F. J., YATES, B. J., FRANZEN, 0., & WALD, J. R. (1983). Lumbar spinal cord
responses to limb vein distension. Journal of the Autonomic Nervous System, 9, 531-546.
THOMPSON, F. J., & BARNES, C. D. (1979). Projection of low threshold venous afferents to
the spinal cord. Brain Research, 177, 561-565.
CARDIAC AFFERENT iNFLUENCES ON CONSCIOUSNESS 85
THOMPSON, L. W., & BoTWINICK, J. (1970). Stimulation in different phases of the cardiac
cycle and reaction time. Psychophysiology, 7, 57-65.
TOURNADE, A., & MALMEJAC, J. (1929) Diversite des actions reflexes qui declenche ]'exci-
tation du sinus carotidien et de son nerf. Comptes Rendus des Seances de Ia Societe de
Biologie, 100, 708-711.
WALKER, B. B., & SANDMAN, C. A. (1977). Physiological response patterns in ulcer patients:
Phasic and tonic components of the electrogastrogram. Psychophysiology, 14, 393-400.
WALKER, B. B., & SANDMAN, C. A. (1979). Relationship of heart rate on the visual evoked
potential. Journal of Comparative and Physiological Psychology, 93, 717-729.
WALKER, B. B., & SANDMAN, C. A. (1982). Visual evoked potentials change as heart rate
and carotid pressure change. Psychophysiology, 19, 520-527.
WALKER, B. B. & WALKER, J. M. (1983) Phase relations between carotid pressure and ongoing
electrocortical activity. International Journal of Psychophysiology, 1, 65-73.
WILLISON, J. R., DuBoULAY, G. H., PAUL, E. A., RussELL, R. W. R., THoMAs, D. J., MARSHALL,
J., PEARSON, T. C., SIMON, L. & WETHERLEY-MEIN, C. (1980). Effect of high haematocrit
on alertness. Lancet, 1, 846-848.
WITKIN, H. A., DYK, R. B., FATTUSON, H. F., GOODENOUGH, D. R., & KARP, s. A. (1962).
Psychological differentiation: Studies of development. New York: Wiley.
Intrinsic Control
4
Mechanisms of Pain Perception
w. LEWIS, LINDA R. NELSON,

]AMES
GREGORY w. TERMAN, YEHUDA SHAVIT,
AND }OHN C. LIEBESKIND
J. INTRODUCTION
Through the course of evolution, the brain has become increasingly able
to respond adaptively to the ever-changing internal and external sensory
world. To do so, it must continually monitor the environment through
specialized sensory systems. One might imagine that the brain passively
receives environmental inputs, processes them, and responds accord-
ingly. We are learning instead that some sensory information can be
modulated before it reaches the brain by the activation of centrifugal
paths descending from higher central nervous system stations to lower
ones in the brain, in the spinal cord, and even in the periphery. Thus,
it appears to be important, at least at certain times, that some inputs
never reach the brain or arrive only after considerable modification.
Evidence supporting this dynamic model of the brain controlling
its own input comes from early studies of several sensory systems (for
review see Livingston, 1959). Perhaps the clearest and most dramatic
example of such centrifugal control, however, comes from more recent
studies of pain perception. In the past fifteen years, a great deal of
evidence has accumulated indicating the existence of an endogenous
pain-modulating substrate originating in the medial brain stem. Path-
ways descend from the brain stem to the dorsal horn of the spinal cord
where the transfer of nociceptive information from the periphery to as-
]AMES W. LEWIS e Mental Health Research Institute, University of Michigan, Ann Arbor,
MI 48109. LINDA R. NELSON, GREGORY W. TERMAN, YEHUDA SHAVIT, AND }OHN c. L!EBES-
KIND e Department of Psychology and Brain Research Institute, University of California,
Los Angeles, CA 90024. The research outlined in this chapter is supported by NIH grant
NS 07628 and a gift from the Brotman Foundation. James W. Lewis is supported by NIDA
Postdoctoral Fellowship F32DA05221.
87
88 JAMES w. LEWIS ET AL.
cending sensory tracts destined for the brain can be greatly reduced or
even blocked completely. The major objective of our chapter is to review
this evidence of a centrifugal system of pain suppression. We will focus
on recent findings addressing the important question of what natural
circumstances call this system into play. Stress appears to be one natural
stimulus activating pain suppressive mechanisms, and our studies of
stress-induced analgesia in the rat will be summarized and related to
similar work by others. Finally, we will describe briefly our most recent
work indicating that although analgesia can be viewed as an adaptive
response to stress, in certain circumstances stress can also prove mal-
adaptive in that it suppresses the immune system and enhances tumor
development.
II. EVIDENCE FOR ENDOGENOUS MECHANISMS OF pAIN

CoNTROL
Ample evidence indicates that intrinsic to the brain and spinal cord
are neural systems the normal function of which is the inhibition of
pain: Reynolds (1969) was first to demonstrate that electrical stimulation
of a portion of the medial brain stem, the periaqueductal gray matter
(PAG), produced analgesia in the rat so profound as to permit surgery
without anesthesia. Subsequently, Mayer et al. (1971) and many others
have elaborated upon this original observation. Stimulation-produced
analgesia (SPA), measured in a variety of standard analgesiometric tests,
is as potent as that obtained using moderate to high doses of morphine
(e.g., Dennis, Choiniere, & Melzack, 1980; Giesler & Liebeskind, 1976;
Mayer et al., 1971). This analgesia appears to represent a true pain
suppression and is not secondary to changes in other sensory or motor
systems (Mayer et al., 1971). In fact, SPA has proven to be useful in the
therapeutic treatment of some forms of intractable pain in man (Hoso-
buchi, Adams, & Linchitz, 1977; Richardson & Akil, 1977; Young et al.,
1984). In general, this descending or centrifugal pain suppression system
appears to take origin in the medial diencephalon and PAG, to descend
from there to the nucleus raphe magnus (NRM) and thence to the cord
by way of the dorsolateral funiculus (Fields & Basbaum, 1979; Lewis,
Stapleton, Castiglioni, & Liebeskind, 1982; Watkins & Mayer, 1982). The
ultimate locus of pain inhibition deriving from stimulation of any of these
areas appears to be the dorsal horn of the spinal cord. Here lie the first
central nociceptive neurons, and it is these neurons preferentially
whose activity is blocked by analgesic brain stimulation (e.g., Oliveras
et al., 1974).
Akil, Mayer, and Liebeskind (1972; 1976b) reported that adminis-
INrRINsic CoNrROL MECHANISMS OF PAIN PERCEPTION 89
tration of naloxone, the specific opiate antagonist, significantly reduced

SPA in the rat. This finding suggested that SPA results from the lib-
eration of endogenous opiatelike substances, a hypothesis that gained
considerable support with the discovery of the opioid peptides (e.g.,
Hughes et al., 1975; for recent review see Akil et al., 1984). Several other
lines of evidence also indicate a significant role for opioid peptides in
pain inhibition. SPA manifests tolerance with repeated administration
and cross-tolerance with morphine (Mayer & Hayes, 1975). Opiate bind-
ing sites and opioid peptide-containing cell bodies or fiber terminals are
found in brain areas supporting SPA, areas where analgesia may also
be elicited by microinjections of opioid peptides and opiate drugs (e.g.,
Hokfelt et al., 1977; Watson, Akil, & Barkas, 1979; Yaksh & Rudy, 1978).
Lesions in this descending system block SPA and opiate analgesia (Bas-
baum, Marley, O'Keefe, & Clanton, 1977). In addition, SPA and anal-
gesic doses of morphine or opioid peptides have a similar excitatory
electrophysiological action on cellular activity within this medial brain
stem substrate (e.g., Oleson, Twombly, & Liebeskind, 1978; Urea & Lie-
beskind, 1979).
However, it is becoming clear that not all intrinsic mechanisms of
pain inhibition rely on this opioid path; rather, multiple mechanisms
exist involving separate or partly separate neurochemical and anatomical
substrates. For example, shortly after the original report by Akil et al.
(1976b), several papers appeared describing failures to replicate nalox-
one antagonism of SPA (e.g., Pert & Walter, 1976; Yaksh, Yeung, &
Rudy, 1976). A possible explanation for this discrepancy is suggested
by the work of Cannon, Lewis, Weinberg, and Liebeskind (1982), who
find that neurochemically discrete SPA substrates appear to exist within
the midline PAG. Stimulation within and below ventral PAG provoked
SPA that is sensitive to naloxone blockade, whereas more dorsal PAG
stimulation caused SPA that was unaffected by this drug. Thus, opioid
and nonopioid forms of SPA appear to exist in anatomically discrete
regions of the midbrain.
III. ACTIVATION OF ENDOGENOUS ANALGESIA SYSTEMS: STRESS-

INDUCED ANALGESIA
A great deal of recent effort has been given to the search for natural,
physiological triggers of the brain's intrinsic analgesia systems. Recently,
it has been shown that exposure to a variety of stressors in the rat elicits
clear pain suppression lasting for at least several minutes (e.g., Amir &
Amit, 1978; Bodnar et al., 1978a, 1978b; Chance, White, Krynock, &
Rosecrans, 1977; Chesher & Chan, 1977; Hayes, Bennett, Newlon, &
Mayer, 1978a; Jackson, Maier, & Coon, 1979; Lewis, Cannon, & Lie-
beskind, 1980a; Madden Akil, Patrick, & Barchas, 1977; Watkins and
Mayer, 1982). Such findings have suggested that although pain is an
important warning signal, under certain emergency or stress conditions,
pain suppression may be yet more adaptive. Thus, stress may be an
important natural input to the endogenous analgesia systems of the
brain at times when feeling pain could disrupt appropriate coping be-
haviors (Liebeskind, Giesler, & Urea, 1976).
Akil et al. (1976a) and Hayes, Bennett, Newlon, and Mayer (1976)
were first to report that exposure to stress causes potent analgesia in
rats. These studies differed, however, in one very important regard:
Akil et al. found that naloxone antagonized stress analgesia suggesting
opioid involvement, whereas Hayes, Bennett, Newlon, & Mayer found
that it did not. Subsequently, other investigators demonstrated stress
analgesia in response to a host of stressors, but the question of opioid
involvement remained controversial. To demonstrate clearly a role for
opioid peptides in stress analgesia, several criteria must be satisfied:
stress analgesia should be blocked by the opiate antagonist drug, na-
loxone; it should develop tolerance with repeated administration; and
it should show cross-tolerance with opiate analgesia. When these cri-
teria were applied to stress analgesia, some investigators obtained evi-
dence of opioid involvement (Amir & Amit, 1978; Bodnar et al., 1978a;
Chesher & Chan, 1977; Madden et al., 1977), but others did not (Bodnar,
Kelly, Steiner, & Glusman, 1978c; Chance & Rosecrans, 1979a, 1979b;
Hayes et al., 1978a).
It was at this time that our studies of stress analgesia began. Because
qualitatively and quantitatively different stressors had been used in prior
studies to produce analgesia (e.g., rotation, oscillation, cold water
swims, restraint, conditioned fear, and different parameters of foot-
shock), integration of these findings was difficult. Our goal was to in-
vestigate the phenomenon of stress analgesia systematically and to clar-
ify what role, if any, was played by the opioid peptides. To this end, a
single stressor, inescapable footshock, was selected and the parameters
of its administration were varied. This procedure allowed precise ex-
perimental control over the stressful stimulus and obviated the problems
introduced by using qualitatively different stressors. Also, it should be
noted that a variety of pain sensitivity tests have been employed in the
assessment of stress analgesia (e.g., hot-plate, tail-flick, formalin test),
and the particular test used may importantly affect the results obtained
(e.g., Dennis & Melzack, 1980). We have chosen to use the tail-flick test
(D'Amour & Smith, 1941) and have defined analgesia as an increase in
latency to respond to radiant heat.
INTRINSIC CoNrROL MECHANISMS OF PAIN PERCEPTION 91
IV. 0PIOID AND N ONOPIOID MECHANISMS OF STRESS

ANALGESIA
We found that inescapable footshock of constant intensity can elicit

qualitatively different analgesic responses depending upon the temporal
parameters of its application (Lewis et al., 1980a). Exposure to prolonged,
intermittent footshock (2.5 rnA, on 1 sec every 5 for 20-30 min) causes
analgesia that is blocked by naloxone in doses as low as 0.1 mglkg (Lewis
et al., 1980a, 1980b). By contrast, brief, continuous footshock of the same
intensity applied for 3 minutes causes a roughly equipotent analgesia
that is unaffected by even high doses of naloxone.
These results suggest the existence of at least two forms of stress
analgesia, one that is mediated by opioid peptides and another that is
not. The naloxone-sensitive analgesic response to prolonged, intermit-
tent footshock satisfies two other criteria of opioid involvement. It shows
tolerance after 14 stress exposures (Lewis, Sherman, & Liebeskind,
1981c; see also Madden et al., 1977), and it is reduced (cross-tolerance)
in morphine-tolerant rats (Lewis et al., 1981c). Once again, brief stress
analgesia appears independent of opioid mechanisms in that it shows
no evidence of tolerance or cross-tolerance with morphine (Lewis et al.,
1981c). Yet another indication of the independence of these two forms
of stress analgesia is the lack of cross-tolerance development between
them. Thus, animals rendered tolerant to prolonged (opioid) footshock
stress display normally robust analgesia in response to brief (nonopioid)
footshock, and the analgesic response to prolonged footshock is unaf-
fected in animals previously exposed to 14 brief stress sessions (Terman,
Lewis, & Liebeskind, 1983a). Taken together, these results indicate the
existence of multiple endogenous analgesia systems selectively accessed
by different parameters of the stressful experience.
v. ROLE OF ENDOCRINE SYSTEMS IN STRESS ANALGESIA
The pituitary-adrenal and the sympatho-adrenal axes have long

been recognized for their important roles in the adaptive response to
stress (Selye, 1956). If pain suppression may be thought of as an adaptive
response to certain stress conditions, participation of these hormonal
systems in stress analgesia seems likely. Several lines of evidence sug-
gest that pituitary hormones contribute importantly to the mediation of
stress analgesia. First, the pituitary is a rich source of 13-endorphin (e.g.,
Guillemin et al., 1977), thought to be the most potent analgesic of the
opioid peptides. Stress has been shown to cause the release of pituitary
92 }AMES W. LEWIS ET AL.
opioids (e.g., Guillemin et al., 1977; Millan et al., 1981). Finally, we and
others have reported that stress analgesia (particularly those forms sen-
sitive to naloxone blockade) is attenuated in hypophysectomized ani-
mals (Amir & Amit, 1979; Bodnar et al., 1979; Lewis, Chudler, Cannon,
& Liebeskind, 1981b; Millan, Przewlocki, Holt, & Herz, 1980). Hypo-
physectomy, however, has been found ineffective in reducing some
opioid forms of stress analgesia (Terman et al., 1984; Watkins et al.,
1982b) and most nonopioid forms (Chance, 1980; Lewis et al., 1981b;
Watkins, Cobelli, Newsome, & Mayer, 1982b).
Another peripheral source of opioid peptides is the adrenal med-
ulla. The adrenal medulla contains enkephalin-like peptides that are
stored and coreleased with catecholamines in response to sympathetic
activation or trauma (see Viveros & Wilson, 1983). Although the precise
physiological function of these peptides remains to be determined, we
have suggested that they are importantly involved in the analgesic re-
sponse to certain forms of stress. Opioid, but not nonopioid, stress an-
algesia is markedly reduced by adrenalectomy, adrenal demedullation,
or denervation of the adrenal medulla via celiac ganglionectomy (Lewis,
Tordoff, Sherman, & Liebeskind, 1982c). Because demedullation and
ganglionectomy have as great an effect as removal of the entire adrenal
gland, and because both basal and stressed adrenocortical function was
unimpaired in adrenal denervated animals, yet these rats failed to man-
ifest opioid stress analgesia, we concluded that this form of stress an-
algesia is dependent on adrenal medullary, not cortical, function (cf.,
however, MacLennan et al., 1982). Moreover, it appears to be the en-
kephalin-like peptides, not catecholamines, that are involved in media-
tion of this form of stress analgesia. A dose of reserpine known to deplete
catecholamines while increasing the adrenal content and stimulation-
induced release of enkephalins (Viveros, Diliberto, Hazum, & Chang,
1980) significantly augments opioid stress analgesia (Lewis et al., 1982c).
This increased analgesia appears to reflect increased release of enke-
phalin-like peptides by stress rather than a nonspecific drug effect in
that the analgesia is still virtually eliminated by an opiate antagonist.
These behavioral observations correlate very well with biochemical
data we have obtained in a collaborative study with Dr. 0. H. Viveros.
The amount of opiate-like material in the adrenal medulla was signifi-
cantly reduced by prolonged (opioid), but not brief (nonopioid), foot-
shock. Medullary enkephalin-like peptide content was dramatically in-
creased in reserpine-treated rats, and this new elevated content was also
reduced by exposure to prolonged stress. Finally, rats made tolerant to
prolonged footshock analgesia no longer showed depletion of adrenal
enkephalin-like peptides after stress (Lewis, Tordoff, Liebeskind, & Viv-
eros, 1982b). These converging lines of evidence strongly implicate ad-
INrRINsic CoNTRoL MECHANISMS oF PAIN PERCEPTION 93
renal enkephalin-like peptides in opioid stress analgesia. Several im-

portant questions, however, such as the locus of the opiate receptor
mediating this analgesia, remain to be answered before this hypothesis
can be fully accepted. In this regard, Maixner and Randich (1984) re-
cently showed that an opioid form of footshock stress analgesia appar-
ently similar to the adrenal medulla-dependent form we have studied
(Lewis et al., 1982c) is attenuated by vagotomy. They suggested that
adrenal enkephalins might bind peripherally and thereby activate vagal
afferents to central pain-inhibitory circuits.
VI. NEUROTRANSMITTERS INVOLVED IN STRESS ANALGESIA
That opioid peptides functioning as hormones appear critical for

some forms of stress analgesia in no way precludes the involvement of
central opioids as well. In fact, there is considerable evidence to indicate
that stressors causing analgesia alter one or another index of brain opioid
activity (e.g., Chance, White, Krynock, & Rosecrans, 1978; Lewis, Can-
non, Liebeskind, & Akil, 1981a; Madden et al., 1977; Millan et al., 1981;
Rossier et al., 1977; Rossier, Gruillemin, & Bloom, 1978). Undoubtedly,
a number of other central neurotransmitter substances are also involved
in the mediation of opioid and nonopioid forms of stress analgesia. For
example, our work has demonstrated an important role for acetylcholine
in opioid stress analgesia (Lewis, Cannon, & Liebeskind, 1983a). We
find that opioid, but not nonopioid, stress analgesia is reduced by sco-
polamine, a muscarinic cholinergic antagonist drug. Methylscopolam-
ine, a muscarinic cholinergic antagonist with only peripheral activity,
failed to affect the analgesic response to prolonged footshock. Addi-
tionally, we and others have shown that oxotremorine, a potent mus-
carinic agonist drug, causes analgesia sensitive to opiate antagonist
blockade (e.g., Lewis et al., 1983a; Pedigo & Dewey, 1981), suggesting
that acetylcholine may stimulate the release of opioid peptides involved
in pain inhibition. From these findings, we conclude that a muscarinic
cholinergic synapse exists in the central opioid pathway mediating stress
analgesia.
Very little is known of the neurochemistry of nonopioid analgesia.
Understanding these systems may prove to be clinically relevant. If these
systems can be harnessed and utilized therapeutically, perhaps prob-
lems associated with opioid analgesia (e.g., tolerance and dependence)
can be avoided. We have found that nonopioid stress analgesia is re-
duced by reserpine (Lewis et al., 1982c; Terman, Lewis, & Liebeskind,
1981). Because reserpine is a relatively nonspecific drug affecting all
monoaminergic neurons, attempts were made to define more precisely
94 }AMES w. LEWIS ET AL.
which monoamine is critical. Nonopioid stress analgesia was unaffected

by specific dopamine, norepinephrine, or serotonin receptor antagonists
or depletors, although monoaminergic agonists potentiated this anal-
gesia (Terman, Lewis, & Liebeskind, 1982a; d., however, Tricklebank,
Hutson, & Curzon, 1982). This form of stress analgesia, however, does
appear to depend on the monoamine, histamine, in that it is reduced
by diphenhydramine, an H 1 receptor blocker, but not by cimetidine, an
H 2 antagonist (Terman, Lewis, & Liebeskind, 1982b). Furthermore, it
appears that histamine of neuronal origin is critical. Nonopioid stress
analgesia is reduced by depletion of both neuronal and mast cell his-
tamine stores with o.-fluoromethylhistidine, an inhibitor of the synthetic
enzyme, histidine decarboxylase. On the other hand, it is not altered
by administration of Compound 48/80, a drug depleting histamine of
mast cell origin only (Terman et al., 1982b). That histamine is involved
in central mechanisms of antinociception is supported by the demon-
stration of histamine in brain regions thought to be involved in pain
inhibition (see Schwartz et al., 1981) and by the fact that analgesia follows
microinjections of histamine into these areas (Glick & Crane, 1978).
VII. THE NEUROANATOMY OF STRESS ANALGESIA
The neuroanatomical substrates of stress analgesia appear to over-

lap considerably with those involved in opiate analgesia and SPA. For
example, we and others have reported that both opioid and nonopioid
stress analgesia rely to some extent on the integrity of spinal pathways
(Chance, 1980; Hayes et al., 1978b), particularly the dorsolateral funi-
culus (DLF) (Lewis et al., 1983c; Watkins, Cobelli, & Mayer, 1982b). Thus,
both forms of stress analgesia share a common substrate and rely on
activation of centrifugal paths.
The nucleus raphe magnus (NRM), a cell group whose axonsproject
to the spinal cord through the DLF (see Fields & Basbaum, 1979), also
appears to be involved in stress analgesia as well as opiate analgesia
and SPA (Azami, Llewelyn, & Roberts, 1982; Chance, 1980; Prieto, Can-
non, & Liebeskind, 1983; Zorman, Hentall, Adams, & Fields, 1981). This
nucleus has been implicated in endogenous mechanisms of antinoci-
ception (Fields, 1984) and has been associated with opioid (Prieto et al.,
1983; Zorman et al., 1981) and nonopioid (Satoh, Akaike, Nakazawa, &
Takagi, 1980) systems. In our work, NRM lesions reduced nonopioid
stress analgesia but did not affect the opioid form (Cannon et al., 1983;
d., however, Watkins et al., 1982b). These same lesions attenuated mor-
phine analgesia (see also Azami et al., 1982), indicating that, although
INTRINSIC CoNTROL MECHANISMS oF PAIN PERCEPTION 95
morphine and opioid stress analgesia are similar in many respects, their
anatomical substrates are not entirely coextensive.
Yet another indication that stress can activate central nervous sys-
tem pain-inhibitory pathways derives from the finding that cross-tol-
erance develops between opioid stress analgesia and opioid mediated
SPA (Penner, Terman, & Liebeskind, 1982). Animals made tolerant to
the opioid form of stress were seen to be cross-tolerant to opioid SPA,
although nonopioid SPA was unaffected. Repeated exposure to nono-
pioid stress affected neither opioid nor nonopioid SPA.
VIII. EviDENCE FOR Two INDEPENDENT 0PIOID FoRMs OF

STRESS ANALGESIA
We have recently found (Terman et al., 1984) that 3 minutes of con-

tinuous footshock (2.5 rnA), parameters previously seen to elicit a non-
opioid form of stress analgesia according to several criteria (e.g., Lewis
et al., 1980a, 1980b, 1981c), can provoke naloxone- or naltrexone-sen-
sitive analgesia in some animals. By systematically varying duration of
exposure while holding other footshock parameters constant, we have
now established that 3 minutes is an intermediate duration between
those values eliciting clearly opioid and those eliciting clearly nonopioid
forms of stress analgesia. Thus, robust naltrexone antagonism of con-
tinuous footshock stress analgesia was observed at 1 and 2 minute
footshock durations, whereas an equipotent analgesia elicited by the
same footshock applied continuously for 4 or 5 minutes was completely
insensitive to this drug (Terman et al., 1983b). Further characterization
of these forms of stress analgesia revealed that the analgesic response
to 1 minute of continuous footshock develops tolerance after 14 expo-
sures and cross-tolerance both with morphine and with the opioid form
of stress analgesia elicited by prolonged, intermittent footshock de-
scribed previously. The analgesic response to 4 minutes of continuous
footshock, by contrast, does not develop tolerance, is unaffected by mor-
phine tolerance, and is not cross-tolerant with either the 1 minute con-
tinuous or 20 minute intermittent forms of opioid stress analgesiC!. Thus,
this form of stress analgesia is mediated by nonopioid mechanisms (Ter-
man et al., 1984).
Several studies have investigated the similarities and differences
between the two opioid forms of stress analgesia elicited by 1 minute
continuous and 20 minute intermittent footshock. We find that both
develop tolerance, and each manifests cross-tolerance with the other,
with morphine, and with the naloxone-sensitive form of SPA (Penner
et al., 1982; Terman et al., 1984). On the other hand, whereas analgesia
96 ]AMES W. LEWIS ET AL.
to the 20 minute intermittent footshock is reduced by hypophysectomy

(Lewis et al., 1981b), is dependent upon adrenal medullary enkephalin-
like peptides (Lewis et al., 1982c), and relies on central muscarinic chol-
inergic mechanisms (Lewis et al., 1983a), analgesia to the 1 minute con-
tinuous footshock is independent of these hormonal and cholinergic
systems. Presumably this form of stress analgesia is dependent on
opioids of central, not peripheral, origin (see Tricklebank et al., 1982).
We also find that the opioid or nonopioid nature of continuous
footshock analgesia depends on current intensity (Terman et al., 1984).
Three minutes of continuous footshock at 1.5 or 2.0 rnA elicits naltrex-
one-sensitive analgesia, whereas a more intense shock (3.0 or 3.5 rnA)
delivered for the same time causes analgesia insensitive to this drug.
Therefore, considering either the intensity or duration of continuous
footshock, it appears that the values used in our earlier work (2.5 rnA
for 3 minutes) are slightly higher than optimal for eliciting opioid stress
analgesia and slightly lower than optimal for eliciting stress analgesia
of the nonopioid sort. It is important to add, however, that, whereas
small increments in intensity or duration of continuous footshock cause
a sudden shift from an opioid to a nonopioid mechanism, applying
exactly the same total amount of footshock as that yielding nonopioid
analgesia (4 minutes continuously), but applying it intermittently for 20
minutes, once again evokes an opioid form (although a different opioid
form) of stress analgesia. Thus, no one parameter (footshock intensity,
duration, or temporal pattern) uniquely determines the neurochemical
basis of stress analgesia, although each can be made the critical variable
by holding the other two constant.
Of special interest to the topic of this volume, we have recently
found that consciousness and higher brain structures are importantly
involved in some, but not all, forms of stress analgesia. Continuous
footshock (2.5 rnA for 1 or 4 minutes) can elicit analgesia in decerebrate
rats or rats given a surgical level of pentobarbital anesthesia (Klein, Lo-
vaas, Terman, & Liebeskind, 1983; Klein, Terman, & Liebeskind, 1984;
Terman et al., 1984). Stress analgesia in these animals, measured by the
tail-flick test, is indistinguishable in magnitude, duration, or sensitivity
to naltrexone antagonism from that displayed by sham-operated or un-
anesthetized controls. On the other hand, the analgesic response to 20
minutes of intermittent footshock is completely abolished by anesthesia
or decerebration (Klein et al., 1983, 1984; Terman et al., 1984; see also
Jensen & Smith, 1981). This finding serves once again to dissociate the
two opioid forms of stress analgesia elicited by continuous versus in-
termittent footshock. It is also evident that if some forms of stress an-
algesia can be studied under conditions of surgical anesthesia to certain
practical or ethical advantage, other forms, unfortunately, cannot.
INTRINSIC CoNTROL MECHANISMS oF PAIN PERCEPTION 97
IX. EARLY ExPERIENCE INFLUENCES STREss ANALGESIA: THE

EFFECTS oF PRENATAL ALCOHOL ExPosuRE
An ample literature suggests that perinatal administration of han-

dling stress or adrenal corticosteroids can profoundly influence the
adult's behavioral and endocrinological response to stress (for review
see Ader, 1975). Exposure to alcohol in utero also appears to affect the
adult's sensitivity to stress (Taylor, Branch, Liu, & Kokka, 1982). How
these perinatal manipulations may specifically affect the analgesic re-
sponse to stress has not been thoroughly researched. That stress an-
algesia may be affected, however, is suggested by findings that neo-
natally administered stress can alter brain enkephalin levels (Torda,
1978) and affect analgesic responsiveness to morphine (Bardo, Bhat-
nagar, & Gebhart, 1981). We have been investigating the long-term ef-
fects of prenatal alcohol exposure and have found that such exposure
augments the analgesic response to stress in the adult rat. Specifically,
an opioid, rather than a nonopioid, mediated form of stress analgesia is
enhanced (Nelson et al., 1982). Adult, female rats, prenatally exposed
to alcohol, were tested for analgesia following two forms of footshock:
intermittent (2.5 rnA, 1 sec on per 5 sec for 10 min) and continuous (2.5
rnA, on for 2 min). These footshock parameters elicit naloxone-sensitive
(opioid) and naloxone-insensitive (nonopioid) forms of stress analgesia,
respectively. Compared to controls, rats prenatally exposed to alcohol
showed a markedly potentiated opioid stress analgesia (Nelson et al.,
1982). Nonopioid stress analgesia was unaffected by the early exposure
to alcohol. These results suggest that fetal exposure to alcohol produces
long-term effects on the functioning of endogenous opioid analgesia
systems. Interestingly, repeated daily exposure of adult rats to the opioid
form of footshock stress causes those prenatally exposed to alcohol, but
not controls, to increase their voluntary consumption of ethanol (Nelson
et al., 1983a).
On the basis of the preceding results, it appears that prenatal ex-
posure to alcohol causes pronounced alterations in endogenous opioid
systems. To test this hypothesis further, we recently carried out several
experiments using morphine as a pharmacological probe to assess the
status of opioid systems in rats prenatally exposed to alcohol. We found
that these rats exhibited markedly enhanced responsiveness to the an-
algesic, hypothermic, and pituitary-adrenal activating effects of mor-
phine (Nelson et al., 1983b). Since these behaviors occur at varying doses
of the drug and are presumed to be mediated by different brain loci, it
appears that the perturbations due to prenatal alcohol exposure are
widespread. One possible explanation of these findings may be that
prenatal alcohol exposure causes changes in brain opiate receptors lead-

ing to increased sensitivity to morphine. However, this simple expla-
nation may not suffice. Our opiate receptor binding studies reveal no
differences in opiate receptor number (of either the mu or delta subtype)
between prenatal alcohol-exposed and normal or pair-fed controls in
frontal cortex, midbrain, striatum, thalamus, medulla, or hippocampus
(Nelson et al., 1984). The only differences we found were in hypothal-
amus, but in this region both prenatal alcohol-exposed and offspring of
pair-fed dams were receptor deficient compared to offspring of normally
fed dams.
X. IMMUNOSUPPRESSIVE AND TUMOR-ENHANCING EFFECTS OF

STRESS
One lesson to be learned from our investigations of the phenom-

enon of stress analgesia is that variations in stress parameters (e.g.,
intermittent versus continuous footshock of the same intensity and total
duration) can elicit neurochemically very different responses (opioid ver-
sus nonopioid analgesia). An extensive but often contradictory literature
relates stress to changes in tumor development and immune function
in human beings and laboratory animals (Ader, 1981; Levy, 1982; Peters
& Mason, 1979). We have recently begun to investigate this problem in
rats, hypothesizing that our dissection of stress parameters into those
causing opioid and nonopioid forms of analgesia would help account
for some of the variability encountered by previous workers in this field.
Our earliest results suggest that this approach is proving fruitful.
We now find that footshock stress that releases opioids (indicated
by the naltrexone sensitivity of the resultant analgesia) is immunosup-
pressive and facilitates the development of a mammary tumor in rats.
In female Fischer 344 rats, the activity of natural killer (NK) cells (lym-
phocytes thought to be important in cancer surveillance) is markedly
suppressed following 4 daily exposures to intermittent (opioid), but not
continuous (nonopioid), footshock. This suppression is attenuated in
rats treated with naltrexone prior to each stress session (Shavit et al.,
1983a). Moreover, the same stress regimen that results in suppression
of NK activity also causes a decrease in survival time and percentage of
survival of Fischer rats implanted with a mammary ascites tumor (MAT
13762 B). These effects are blocked in animals implanted with slow-
release naltrexone pellets (Lewis et al., 1983b). Continuous (nonopioid)
footshock stress affects neither NK activity nor tumor development.
The mechanisms by which opioid peptides interact with these sys-
tems are no doubt complex and remain to be elucidated. In further work
INTRINsic CoNTROL MECHANISMS oF PAIN PERCEPTION 99
on this subject we have attempted to determine whether the tumor-

enhancing effects of stress could be mimicked by morphine and whether
the adverse effects of stress would tolerate upon repeated exposure. The
results of this study gave mixed support for the hypothesized role of
opioid peptides in these phenomena. The tumor-enhancing and im-
munosuppressive effects of stress were mimicked by administration of
morphine (albeit in very high doses) (Lewis et al., 1984; Shavit et al.,
1984a, 1984b). However, in the case of tumor enhancement, no evidence
of tolerance was obtained, nor were the effects of stress reduced in
morphine-tolerant rats (cross-tolerance) (Lewis et al., 1984). Neverthe-
less, taken together, these results suggest a significant role for opioids
in the interactions between stress, the immune system, and cancer.
XI. SuMMARY AND CoNCLUSIONS
These studies provide clear evidence that exposure to stress is an

adequate stimulus for activation of intrinsic pain-suppressive systems.
Our initial results (Lewis et al., 1980a) demonstrated that a single stres-
sor, inescapable footshock, can selectively activate either an opioid or a
nonopioid mechanism of analgesia; thus, stress analgesia is not a unitary
phenomenon. Neurochemically, neuroanatomically, and hormonally
different analgesia substrates appear to exist; and which of these is called
into play appears to depend on variations in the temporal parameters
and intensity of the footshock stress stimulus.
Certainly, there is much we need to learn about when endogenous
pain-inhibitory systems are active and what natural stimuli or environ-
mental circumstances occasion their use. One important factor appears
to be ability of the organism to exert control over the stressful stimulus.
In the work of Maier and colleagues, inescapable (i.e., uncontrollable),
but not escapable, stress appears to activate an opioid analgesia system
(Maier et al., 1980). Although both the 20 minute intermittent and the
3 minute continuous footshock procedures employed in our studies are
inescapable, only the former activates an opioid system. It may be that
only the 20 minute paradigm is perceived as "inescapable" by the rats.
This notion is supported by our finding, in collaboration with Maier,
that exposure to 20 minute intermittent, but not 3 minute continuous,
footshock causes behavioral deficits termed "learned helplessness," pre-
viously shown to result from experience with uncontrollable stress
(Maier et al., 1983). The importance of controllability in determining the
ultimate effects of stress is also apparent in studies of stress-induced
immunosuppression and tumor enhancement. Others have previously
shown that exposure to inescapable, but not escapable, stress facilitates
tumor development (Visintainer, Volpicelli, & Seligman, 1982; Sklar &

Anisman, 1979) and suppresses T-cell function (Laudenslager et al.,
1983). We have recently obtained parallel results measuring NK activity
(Shavit et al., 1983b).
Although inescapable footshock at best merely models natural stress
conditions, it is proving useful as a highly precise, reliable, and con-
trollable probe for elucidating anatomical substrates and neurochemical
mechanisms of antinociception. More naturalistic work by others sug-
gesting that sexual behavior (Crowley, Rodriguez-Sierra, & Komisaruk,
1977), fighting (Miczek, Thompson, & Shuster, 1982), and food depri-
vation (Bodnar et al., 1978b; McGivern et al., 1979) can be adequate stim-
uli for triggering opioid or nonopioid analgesia systems apparently sim-
ilar to those we are studying encourages the belief that a rapprochement
between these more mechanistic and naturalistic kinds of studies can
ultimately be made. Finally, the experience we have gained in these
studies about the neurochemically and hormonally diverse mechanisms
activated by different stress parameters is beginning to provide some
measure of predictive control in our attempts to establish reliable rela-
tionships between stress and both tumor development and immune
function.
REFERENCES
ADER, R. (1975). Early experience and hormones: Emotional behavior and adrenocortical
function. In B. E. ELEFTHERIOU & R. SPROTT (Eds.), Hormonal correlates of behavior. New
York: Plenum Press, pp. 7-33.
ADER, R. (1981). Psychoneuroimmunology. New York: Academic Press.
AKIL, H., MAYER, D. J., & LrEBESKIND, J. C. (1972). Comparaison chez le rat entre l'analgesie
induite par stimulation de Ia substance grise peri-aqueducale et l'analgesie morphi-
nique. Comptes Rendus de l'Academie des Sciences (Paris), 274, 3603-3605.
AKIL, H., MADDEN, J., PATRICK, R. L., & BARCHAS, J. D. (1976a). Stress-induced increase
in ·endogenous opiate peptides; concurrent analgesia and its partial reversal by na-
loxone. In H. W. KosTERLITZ (Ed.), Opiates and endogenous opioid peptides. Amsterdam:
Elsevier, pp. 63-70.
AKIL, H., MAYER, D.]., & LIEBESKIND, J. C. (1976b). Antagonism of stimulation-produced
analgesia by naloxone, a narcotic antagonist. Science, 191 961-962.
AKIL, H., WATSON, S.]., YouNG, E., LEwrs, M. E., KHAGITURIAN, H., & WALKER, J. M.
(1984). Endogenous opioids: Biology and function. Annual Review of Neuroscience, 7,
223-255.
AMrR, S., & AMIT, Z. (1978). Endogenous opioid ligands may mediate stress-induced
changes in the affective properties of pain related behavior in rats. Life Sciences, 23,
1143-1152.
AMIR, S., AMrT, Z. (1979). The pituitary gland mediates acute and chronic pain respon-
siveness in stressed and non-stressed rats. Life Sciences, 24, 439-448.
AZAMI, J., LLEWELYN, M. B., & RoBERTS, H. H. T. (1982). The contribution of nucleus
INTRINSIC CONTROL MECHANISMS OF PAIN PERCEPTION 101
reticularis paragigantocellularis and nucleus raphe magnus to the analgesia produced

by systemically administered morphine, investigated with the microinjection tech-
nique. Pain, 12, 229-246.
BARDO, M. T., BHATNAGAR, R. K., & GEBHART, G. F. (1981). Opiate receptor ontogeny and
morphine-induced effects: Influence of chronic footshock stress in preweanling rats.
Developmental Brain Research, 1, 487-495.
BASBAUM, A. I., MARLEY, N.J., O'KEEFE, J., & CLANTON, C. H. (1977). Reversal of morphine
and stimulus produced analgesia by subtotal spinal cord lesions. Pain, 3, 43-56.
BODNAR, R. ]., GwsMAN, M., BRUTUS, M., SPIAGGIA, A., & KELLY, D. D. (1979). Analgesia
induced by cold-water stress: Attenuation following hypophysectomy. Physiology and
Behavior, 23, 53-62.
BODNAR, R. J., KELLY, D. D., SPIAGGIA, A., EHRENBERG, C., & GwsMAN, M. (1978a). Dose-
dependent reductions by naloxone of analgesia induced by cold-water stress. Phar-
macology, Biochemistry, and Behavior, 8, 667-672.
BoDNAR, R. ]., KELLY, D. D., SPIAGGIA, A., & GwsMAN, M. (1978b). Biphasic alterations of
nociceptive thresholds induced by food deprivation. Physiological Psychology, 6, 391-
395.
BoDNAR, R. ]., KELLY, D. D., STEINER, S. S., & GLUSMAN, M. (1978c). Stress-produced an-
algesia and morphine-produced analgesia: Lack of cross-tolerance. Pharmacology, Bio-
chemistry, and Behavior, 8, 661-666.
CANNON, J. T., LEWIS, J. W., WEINBERG, V. E., & LIEBESKIND, ]. C. (1983). Evidence for the
independence of brain stem mechanisms mediating analgesia induced by morphine
and two forms of stress. Brain Research, 269, 231-236.
CANNON,]. T., PRIETo, G.]., LEE, A., & LIEBESKIND, ]. C. (1982). Evidence for opioid and
nonopioid forms of stimulation-produced analgesia in the rat. Brain Research, 243, 315-
321.
CHANCE, W. T. (1980). Autoanalgesia: Opiate and non-opiate mechanisms. Neuroscience
and Biobehavior Review, 4, 55-67.
CHANCE, W. T., & RosEcRANs, ]. A. (1979a). Lack of cross-tolerance between morphine
and autoanalgesia. Pharmacology, Biochemistry, and Behavior, 11, 639-642.
CHANCE, W. T., & RosECRANS, J. A. (1979b). Lack of effect of naloxone on autoanalgesia.
Pharmacology, Biochemistry, and Behavior, 11, 643-646.
CHANCE, W. T., WHITE, A. C., KRYNOCK, G. M., & RosECRANS, J. A. (1977). Autoanalgesia:
Behaviorally activated antinociception. European Journal of Pharmacology, 44, 283-284.
CHANCE, W. T., WHITE, A. C., KRYNOCK, G. M., & RosECRANS, J. A. (1978). Conditional
fear-induced decreases in the binding of eH)-N-Leu-enkephalin to rat brain. Brain
Research, 141, 371-374.
CHESHER, G. B., & CHAN, B. (1977). Footshock induced analgesia in mice: Its reversal by
naloxone and cross-tolerance with morphine. Life Sciences, 21, 1569-1574.
CROWLEY, W. R., RoDRIGUEZ-SIERRA,]. F., & KoMISARUK, B. R. (1977). Analgesia induced
by vaginal stimulation in rats is apparently independent of a morphine-sensitive pro-
cess. Psychopharmacology, 54, 223-225.
D'AMOUR, F. E., & SMITH, D. L. (1941). A method for determining loss of pain sensation.
Journal of Pharmacology and Experimental Therapeutics, 72, 74-79.
DENNIS, S. G., CHOINIERE, M., & MELZACK, R. (1980). Stimulation-produced analgesia in
rats: Assessment by two pain tests and correlation with self-stimulation. Experimental
Neurology, 68, 295-309.
DENNIS, S. G., & MELZACK, R. (1980). Pain modulation by 5-hydroxytryptaminergic agents
and morphine as measured by three pain tests. Experimental Neurology, 69, 260-270.
Fmws, H. L. (1984). Brainstem mechanisms of pain modulation. In L. KRUGER & ]. C.
LIEBESKIND (Eds.), Neural mechanisms of pain. Advances in Pain Research and Therapy, Vol.
6. New York: Raven Press, pp. 252.
FIELDS, H. L., & BASBAUM, A. I. (1979). Anatomy and physiology of a descending pain
control system. In J. J. BoNICA, J. C. LIEBESKIND, & D. G. ALBE-FESSARD (Eds.), Advances
in pain research and therapy, Vol. 3. New York: Raven Press, pp. 427-440.
GEISLER, G. J., & LIEBESKIND, J. C. (1976). Inhibition of visceral pain by electrical stimulation
of the periaqueductal gray matter. Pain, 2, 43-48.
GLICK, S., & CRANE, L. A. (1978). Opiate-like and abstinence-like effects of intracerebral
histamine administration in rats. Nature, 273, 547-549.
GUILLEMIN, R., VARGO, T., RosslER,]., MINicK, S., LING, N., RlVIER, C., VALE, W., & BLOOM,
F. (1977). 13-endorphin and adrenocorticotropin are secreted concomitantly by the
pituitary gland. Science, 197, 1367-1369.
HAYES, R. L., BENNETT, G. J., NEWLON, P. G., & MAYER, D. ]. (1976). Analgesic effects of
certain noxious and stressful manipulations in the rat. Society for Neuroscience Abstracts,
2, 939.
HAYES, R. L., BENNEIT, G. J., NEWLON, P. G., & MA,YER, D. J. (1978). Behavioral and phys-
iological studies of non-narcotic analgesia in the rat elicited by certain environmental
stimuli. Brain Research, 155, 69-90.
HAYES, R. L., PRicE, D. D., BENNETT, G. ]., WILcox, G. L., & MAYER, D. J. (1978b). Dif-
ferential effects of spinal cord lesions on narcotic and non-narcotic suppression of
nociceptive reflexes: Further evidence for the physiological multiplicity of pain mod-
ulation. Brain Research, 155, 91-102.
H6KFELT, T., LJUNGDAHL, A., TERENIUS, L., ELDE, R., & NILSON, G. (1977). Immunohisto-
chemical analysis of peptide pathways possibly related to pain and analgesia: Enke-
phalin and substance P. Proceedings of the National Academy of Sciences, 74, 3081-3085.
HOSOBUCHI, Y., ADAMS, J. E., & LINCHITZ, R. (1977). Pain relief by electrical stimulation of
the central gray matter in humans and its reversal by naloxone. Science, 197, 183-186.
HuGHES,]., SMITH, T. W., KosTERLITZ, H. W., FoTHERGILL, L.A., MoRGAN, B. A., & MoRRis,
H. R. (1976). Identification of two related pentapeptides from the brain with potent
opiate agonist activity. Nature, 258, 577-579.
}ACKSON, R. L., MAIER, S. F., & CooN, D. J. (1979). Long-term analgesic effects of ines-
capable shock and learned helplessness. Science, 206, 91-93.
}ENSEN, T. S. & SMITH, D. F. (1981). The role of consciousness in stress-induced analgesia.
Journal of Neural Transmission, 52, 55-60.
KLEIN, M. V., LovAAS, K. M., TERMAN, G. W., & LIEBESKIND, J. C. (1983). The effects of
decerebration and spinal transection on three discrete forms of stress-induced anal-
gesia. Society for Neuroscience Abstracts, 9, 795.
KLEIN, M. V., TERMAN, G. W., & LIEBESKIND, J. C. (1984). Effects of pentobarbital on three
forms of stress-induced analgesia. Society for Neuroscience Abstracts, 10, 1105.
LAUDENSLAGER, M. L., RYAN, S. M., DRUGAN, R. C., HYSON, R. L., & MAIER, s. F. (1982).
Coping and immunosuppression: Inescapable but not escapable shock suppresses
lymphocyte proliferation. Science, 221, 568-570.
LEVY, S. M. Biological mediators of behavior and disease: Neoplasia. New York: Elsevier
Biomedical.
LEWIS,]. W., CANNON, J. T., & LIEBESKIND, J. C. (1980a). Opioid and nonopoid mechanisms
of stress analgesia. Science, 208, 623-625.
LEWIS,]. W., CANNON,]. T., & LIEBESKIND, J. C. (1983a). Involvement of central muscarinic
cholinergic mechanisms in opioid stress analgesia. Brain Research, 270, 289-293.
LEWIS, J. W., CANNON, J. T., STAPLETON, J. M., & LIEBESKIND, J. C. (1980b). Stress activates
endogenous pain-inhibitory systems: Opioid and nonopioid mechanisms. Proceedings
of the Western Pharmacology Society, 23, 85-88.
LEWIS, J. W., CANNON,]. T., LIEBESKIND, ]. C., & AKIL, H. (1981a). Alterations in brain 13-
endorphin immunoreactivity following acute and chronic stress. Pain, Supplement 1,
S263.
INrRINsic CoNTRoL MECHANISMS oF PAIN PERCEPTION 103
LEWis, J. W., CHUDLER, E. H., CANNON, J. T., & LJEBESKIND, J. C. (1981b). Hypophysectomy
differentially affects morphine and stress analgesia. Proceedings of the Western Phar-
macology Society, 24, 323-326.
LEWIS, J. W., SHAvrr, Y., TERMAN, G. W., GALE, R. P., & LIEBESKIND, J. C. (1983-84). Stress
and morphine affect survival of rats challenged with a mammary ascites tumor (MAT
137628.) Natural Immunity and Cell Growth Regulation, 3, 43-50.
LEWIS, J. w., SHAVIT, Y., TERMAN, G. w., NELSON, L. R., GALE, R. P., & LIEBESKIND, J. c.
(1983b). Apparent involvement of opioid peptides in stress-induced enhancement of
tumor growth. Peptides, 4, 635-638.
LEWIS, J. W., SHERMAN, J. E., & LJEBESKIND, J. C. (1981c). Opioid and nonopioid stress
analgesia: Assessment of tolerance and cross-tolerance-with morphine. Journal of Neu-
roscience, 1, 358-363.
LEWIS, J. W., STAPLETON, J. M., CASTIGLIONJ, A. J., & LIEBESKIND, J. C. (1982a). Stimulation-
produced analgesia and intrinsic mechanisms of pain suppression. In G. FINK & L. J.
WHALLEY (Eds.), Neuropeptides-Basic and clinical aspects. Edinburgh: Churchill Living-
stone, pp. 41-49.
LEWIS, J. W., TERMAN, G. W., WATKINS, L. R., MAYER, D.]., & LIEBESKIND, J. C. (1983c).
Opioid and nonopioid mechanisms of footshock-induced analgesia: Role of the spinal
dorsolateral funiculus. Brain Research, 269, 231-236.
LEWIS, J. W., ToRDOFF, M. G., LIEBESKIND, J. C., & VIVERos, 0. H. (1982b). Evidence for
adrenal medullary opioid involvement in stress analgesia. Society for Neuroscience Ab-
stracts, 8, 778.
LEWIS, J. W., ToRDOFF, M.G., SHERMAN, J. E., & LIEBESKIND, J. C. (1982c). Adrenal medullary
enkephalin-like peptides may mediate opioid stress analgesia. Science, 217, 557-559.
LIEBESKIND, J. C., GIESLER, G. J., }R., & UReA, G. (1976). Evidence pertaining to an endog-
enous mechanism of pain inhibition in the central nervous system. In Y. ZoTTERMAN
(Ed.), Sensory functions of the skin in primates. Oxford: Pergamon Press, pp. 561-573.
LIVINGSTON, R. B. (1959). Central control of receptors and sensory transmission systems.
In J. FIELD (Ed.), Handbook of physiology, Section 1: Neurophysiology, Vol. 1. Washington,
DC: American Physiological Society, pp. 741-760.
MACLENNAN, A. J., DRUGAN, R. C., HYSON, R. L., MAIER, s. F., MADDEN, J., & BARCHAS,
J. D. (1982). Corticosterone: A critical factor in an opioid form of stress-induced an-
algesia. Science, 215, 1530-1532.
MADDEN, J., AKIL, H., PATRICK, R. L., & BARCHAS, J. D. (1977). Stress-induced parallel
changes in central opioid levels and pain responsiveness in the rat. Nature, 265, 358-
360.
MAIER, S. F., DAviES, S., GRAu, J. W., JAcKSoN, R. L., MoRRISON, D. H., MovE, T., MADDEN,
J., & BARCHAS, J. D. (1980). Opiate antagonists and the long-term analgesic reaction
induced by inescapable shock in rats. Journal of Comparative and Physiological Psychology,
94, 1172-1183.
MAIER, S. F., SHERMAN, J. E., LEWIS, J. W., TERMAN, G. W., & LIEBESKIND, J. C. (1983). The
opioid/nonopioid nature of stress-induced analgesia and learned helplessness. Journal
of Experimental Psychology: Animal Behavior Processes, 9, 80-90.
MAIXNER, W., & R.A.NoicH, A. (1984). Role of the right vagal nerve trunk in antinociception.
Brain Research, 298, 374-377.
MAYER, D. J., & HAYES, R. L. (1975). Stimulation-produced analgesia: Development of
tolerance and cross-tolerance to morphine. Science, 188, 941-943.
MAYER, D. J., WoLFLE, T. L., AKIL, H., CARDER, B., & LIEBESKIND, J. C. (1971). Analgesia
from electrical stimulation in the brainstem of the rat. Science, 174, 1351-1354.
McGIVERN, R. F., BERKA, C., BERNTSON, G. G., WALKER, J. M., & SANDMAN, C. A. (1979).
Effect of naloxone on analgesia induced by food deprivation. Life Sciences, 25, 885-
888.
104 }AMES W. LEWIS ET AL.
MlcZEK, K. A., THoMPSON, M. L., & SHUSTER, L. (1982). Opioid-like analgesia in defeated
mice. Science, 215, 1520-1522.
MlLLAN, M. J., PRZEWLOCKI, R., & HERZ, A. (1980). A non-j3-endorphinergic adenohypo-
physeal mechanism is essential for an analgetic response to stress. Pain, 8, 343-353.
MILLAN, M. J. TsANG, Y. F., PRZEWLOCKI, R., HoLLT, V., & HERZ, A. (1981). The influence
of foot-shock stress upon brain, pituitary, and spinal cord pools of immunoreactive
dynorphin in rats. Neuroscience Letters, 24, 75-79.
NELSON, L. R., LEWIS, J. W., LIEBESKIND, J. C., BRANCH, B. J., & TAYLOR, A. N. (1982). Fetal
exposure to ethanol potentiates opioid stress analgesia in adult rats. Society for Neu-
roscience Abstracts, 8, 596.
NELSON, L. R., LEWIS, J. W., LIEBESKIND, J. C., BRANCH, B. J., & TAYLOR, A. N. (1983a).
Stress-induced changes in ethanol consumption in adult rats exposed to ethanol in
utero. Proceedings of the Western Pharmacology Society, 26, 205-209.
NELsoN, L. R., LEWIS, J.W., LIEBESKIND, J. C., KoKKA, N., RANDOLPH, D., BRANCH, B. J., &
TAYLOR, A. N. (1983b). Enhanced responsiveness to morphine in adult rats following
fetal ethanol exposure. Society for Neuroscience Abstracts, 9, 1242.
NELSON, L. R., TAYLOR, A. N., BRANCH, B. J., LIEBESKIND, J. C., & LEWIS, J. W. (1984). Fetal
exposure to ethanol affects sensitivity to morphine but not brain opiate receptor bind-
ing in rats. Society for Neuroscience Abstracts, 10, 964.
OLESON, T. D., TwoMBLY, D. A., & LIEBESKIND, J. C. (1978). Effects of pain-attenuating
brain stimulation and morphine on electrical activity in the raphe nuclei of the awake
rat. Pain, 4, 211-230.
OLIVERAS, J. L., BESSON, J. M., & LIEBESKIND, J. C. (1974). Behavioral and electrophysiological
evidence of pain inhibition from midbrain stimulation in the cat. Experimental Brain
Research, 20, 32-44.
PEDIGO, N. W., & DEWEY, W. L. (1981). Acetylcholine induced antinociception; comparisons
to opiate analgesia. In G. PEPEU & H. LADINSKY (Eds.), Cholinergic mechanisms: Phylo-
genetic aspects, central and peripheral synapses, and clinical significance. Advances in Behav-
ioral Biology, Vol. 25. New York: Plenum Press, pp. 795-807.
PENNER, E. R., TERMAN, G. W., & LIEBESKIND, J. C. (1982). Cross-tolerance between opioid
mediated stimulation-produced and stress-induced analgesia. Society for Neuroscience
Abstracts, 8, 619.
PERT, A., & WALTER, M. (1976). Comparison between naloxone reversal of morphine and
electrical stimulation induced analgesia in the rat mesencephalon. Life Sciences, 19,
1023-1032.
PETERS, L. J., & MASON, K. A. (1979). Influence of stress on experimental cancer. In B. A.
STOLL (Ed.), Mind and cancer prognosis. New York: Wiley, pp. 103-124.
PRIETO, G. J., CANNON, J. T., & LIEBESKIND, J. C. (1983). N raphe magnus lesions disrupt
stimulation-produced analgesia from ventral but not dorsal midbrain areas in the rat.
Brain Research, 261, 53-57.
REYNOLDS, D. V. (1969). Surgery in the rat during electrical analgesia induced by focal brain
stimulation. Science, 164, 444-445.
RicHARDSON, D. D., & AKIL, H. (1977). Pain reduction by electrical brain stimulation in
man. Part 2: Chronic self-administration in the periventricular gray matter. Journal of
Neurosurgery, 47, 184-194.
RosslER, J., FRENCH, E. D., RiviER, C., LING, N., GuiLLEMIN, R., & BLOoM, F. E. (1977). Foot-
shock induced stress increases 13-endorphin levels in blood but not brain. Nature, 270,
618-620.
RosslER, J., GuiLLEMIN, R., & BLOOM, F. E. (1978). Foot-shock induced stress decreases leu5 -
enkephalin immunoreactivity in rat hypothalamus. European Journal of Pharmacology,
48' 465-466.
INTRINSIC CONTROL MECHANISMS OF PAIN PERCEPTION 105
SATOH, M., AKAIKE, A., NAKAZAWA, T., & TAKAGI, H. (1980). Evidence for involvement of
separate mechanisms in the production of analgesia by electrical stimulation of the
nucleus reticularis paragigantocellularis and nucleus raphe magnus in the rat. Brain
Research, 194, 525-529.
ScHWARTZ, J. C., BARBIN, G., DucHEMIN, A. M., GARBARG, M., PoLLARD, H., & QuACH, T.
T. (1981). Functional role of histamine in the brain. In G. C. PALMER (Ed.), Neuro-
pharmacology of central nervous system and behavioral disorders. New York: Academic
Press, pp. 539-570.
SELYE, H. (1956). The stress of life. New York: McGraw-Hill.
SHAVIT, Y., LEWIS, J. W., TERMAN, G. W., GALE, R. P., & LIEBESKIND, J. C. (1983a). Endog-
enous opioids may mediate the effects of stress on tumor growth and immune func-
tion. Proceedings of the Western Pharmacology Society, 26, 53-56.
SHAVIT, Y., LEWIS,]. W., TERMAN, G. W., GALE, R. P., & LIEBESKIND, J. C. (1984a). Opioid
peptides mediate the suppressive effect of stress on natural killer cell cytotoxicity.
Science, 223, 188-190.
SHAVIT, Y., RYAN, S. M., LEWis,]. C., LAUDENSLAGER, M. L., TERMAN, G. W., MAIER, S. F.,
GALE, R. P., & LIEBESKIND, J. C. (1983b). Inescapable but not escapable stress alters
immune function. Physiologist, 26, A-64.
SHAvrr, Y., TERMAN, G. W., MARTIN, F. C., GALE, R. P., & LIEBESKIND, J. C. (1984b). Nal-
trexone-sensitive suppression of the immune system's natural killer cells by morphine.
Society for Neuroscience Abstracts, 10, 726.
SKLAR, L. S., & ANISMAN, H. (1979). Stress and coping factors influence tumor growth.
Science, 205, 513-515.
TAYLOR, A. N., BRANcH, B.]., LIU, S. H., & KoKKA, N. (1982). Long-term effects of fetal
ethanol exposure on pituitary-adrenal response to stress. Pharmacology, Biochemistry,
and Behavior, 16, 585-589.
TERMAN, G. W., LEWIS, J. W., & LIEBESKIND, J. C. (1981). Monoaminergic mechanisms of
stress analgesia. Society for Neuroscience Abstracts, 7, 879.
TERMAN, G. W., LEWIS, J. W., & LIEBESKIND, J. C. (1982a). Role of the biogenic amines in
stress analgesia. Proceedings of the Western Pharmacology Society, 25, 7-10.
TERMAN, G. W., LEwis, J. W., & LIEBESKIND, J. C. (1982b). Evidence for the involvement
of histamine in stress analgesia. Society for Neuroscience Abstracts, 8, 619.
TERMAN, G. W., LEWIS, J. W., & LIEBESKIND, J. C. (1983a). Opioid and nonopioid mecha-
nisms of stress analgesia: Lack of cross-tolerance between stressors. Brain Research,
206, 147-150.
TERMAN, G. W., LEWIS, J. W., & LIEBESKIND, J. C. (1983b). The sensitivity of opioid-mediated
stress analgesia to narcotic antagonists. Proceedings of the Western Pharmacology Society,
26, 49-52.
TERMAN, G. W., SHAVIT, Y., LEWIS, J. W., CANNON, J. T., & LIEBESKIND, J. C. (1984). Intrinsic
mechanisms of pain inhibition and their activation by stress. Science, 226, 1270-1277.
ToRDA, C. (1978). Effects of recurrent postnatal pain-related stressful events on opiate
receptor-endogenous ligand system. Psychoneuroendocrinology, 3, 85-91.
TRICKLEBANK, M. D., HUTSON, P. H., & CuRZON, G. (1982). Analgesia induced by brief
footshock is inhibited by 5-hydroxytryptamine but unaffected by antagonists of 5-
hydroxytryptarnine or by naloxone. Neuropharmacology, 21, 51-56.
UReA, G., & LIEBESKIND, J. C. (1979). Electrophysiological indices of opiate action in awake
and anesthetized rats. Brain Research, 161, 162-166.
VISINTAINER, M. A., VOLPICELLI, J. R., & SELIGMAN, E. P. (1982). Tumor rejection in rats
after inescapable or escapable shock. Science, 216, 437-439.
VIVEROS, 0. H., DILIBERTO, E. J., ]R., HAZUM, E., & CHANG, K.-J. (1980). Enkephalins as
possible adrenomedullary hormones: Storage, secretion, and regulation of synthesis.
106 ]AMES w. LEWIS ET AL.
In E. CosTA & M. TRABUCCHI (Eds.), Neural peptides and neuronal communication. New
York: Raven Press, pp. 191-201.
VIVEROS, 0. H., & WILSON, S. P. (1983). The adrenal chromaffin cell as a model to study
the co-secretion of enkephalins and catecholamines. Journal of the Autonomous Nervous
System, 7, 41-58.
WATKINS, L. R., CoBELLI, D. A., & MAYER, D. J. (1982b). Opiate vs. non-opiate footshock
analgesia (FSIA): Descending and intraspinal components. Brain Research, 245, 97-
106.
WATKINS, L. R., CosELLI, D. A., NEwsoME, H. H., &MAYER, D. J. (1982c). Footshockinduced
analgesia is dependent neither on pituitary nor sympathetic activation. Brain Research,
245, 81-96.
WATKINS, L. R., & MAYER, D. J. (1982). The organization of endogenous opiate and non-
opiate pain control systems. Science, 216, 1185-1192.
WATSON, S. J., AKIL, H., & BARCHAS, J.D. (1979). Immunohistochemical and biochemical
studies of the enkephalins, ~-endorphin and related peptides. In E. UsmN, W. E.
BuNNEY, & N. S. KLINE (Eds.), Endorphins in mental health research. New York: Oxford
University Press, pp. 30-44.
YAKSH, T. L, & RUDY, T. A. (1978). Narcotic analgetics: CNS sites and mechanisms of action
as revealed by intracerebral injection techniques. Pain, 4, 299-359.
YAKSH, T. L., YEUNG, J. C., & RUDY, T. A. (1976). An inability to antagonize with naloxone
the elevated thresholds resulting from electrical stimulation of the mesencephalic cen-
tral gray. Life Sciences, 18, 1193-1198.
YouNG, R. F., FELDMAN, R. A., KROENING, R., FULTON, W., & MoRRIS, J. (1984). Electrical
stimulation of the brain in the treatment of chronic pain in man. In L. KRuGER & J.
C. LIEBESKIND (Eds.), Neural mechanisms of pain. Advances in Pain Research and Therapy,
Vol. 6. New York: Raven Press, pp. 289-303.
ZORMAN, G., HENTALL, I. D., ADAMS, J. E., & FIELDS, H. L. (1981). Naloxone-reversible
analgesia produced by microstimulation in the rat medulla. Brain Research, 219, 137-
148.
5 Inhibition and Cognition
Toward an Understanding of Trauma and Disease
]AMES W. PENNEBAKER AND CLAUDIA W. HoovER
l. INTRODUCTION
When we are upset or confused about an event, it often helps to write

or to talk about it with someone. The mere act of organizing and cate-
gorizing the event makes us feel better and allows us to devote our
attention to other things. Certain events, such as getting a parking ticket
or having an airline lose our bags, can be assimilated and explained
relatively easily. Others, however, such as being assaulted or being re-
jected by one's spouse, require far more cognitive work in order to or-
ganize, understand, and ultimately to forget. One problem that occa-
sionally occurs is that we are sometimes unable to discuss certain
personally traumatic events for fear of embarrassment or punishment.
In such cases, we must actively inhibit the confiding process. As will
be seen, the inhibition process is itself stressful. Hence, the combined
physiological effects of experiencing the trauma, attempting to assimilate
the trauma, and inhibition can sum to produce long-term stress and
susceptibility to disease. The model that we propose has evolved from
rather diverse literatures and serendipitous observations. Because our
research in this area is in a nascent stage, our model should be viewed
as a tentative conceptual framework that will undoubtedly be revised
over the next few years.
Our interest in traumatic events was piqued by a series of results
that we uncovered while examining the nature of the reporting of phys-
ical symptoms. As part of a large survey administered to over 400 college
A version of this paper was presented at the annual meeting for the Society for Psycho-
physiological Research, Pacific Grove, CA, 1983.
}AMES W. PENNEBAKER e Department of Psychology, Southern Methodist University, Dal-

las, Texas 75275. CLAUDIA W. HooVER e Center for Behavioral Medicine, Du Pont As-
sociates, Rockville, Maryland 20852.
107
108 }AMES W. PENNEBAKER AND CLAUDIA W. HoovER
students, we had included an item asking, "Prior to the age of 17, did
you have a traumatic sexual experience (e.g., rape, being molested)?"
Among high-symptom reporters, 14.4% responded affirmatively
whereas only 2.5% of the low-symptom reporters did so (Pennebaker,
1982). Later surveys revealed that reports of a sexual trauma were highly
correlated with physician visits, drug use, and increased incidence of
cardiovascular, digestive, kidney, lung, skin, and other disorders. Fi-
nally, as will be discussed in greater detail below, those reporting sexual
traumas claimed that the traumatic event had occurred several years
earlier and that, by and large, they had not confided in others about
the event.
A second issue that has long puzzled us (as well as many psycho-
therapists) is why psychotherapy, self-disclosure, or even writing a diary
makes people feel better and makes them less prone to illness? A par-
ticularly intriguing example comes from the seminal work of Harold G.
Wolff (Wolf & Goodell, 1968) wherein Wolff and his colleagues perform
a number of stress tests on patients with psychosomatic diseases in order
to isolate specific psychological precursors to physiological change.
Wolff often presents clinical data indicating that once the patient un-
derstands the psychological causes of his or her health problems, the
symptoms gradually or, on occasion, abruptly subside.
A third enigma concerns the nature of physiological responses to
deception and inhibition in laboratory or real-world polygraph settings.
There is little doubt that inducing a subject to lie in the guilty knowledge
test (GKT) produces an impressive physiological response. Further, if
the subject must respond no to a series of words (one of which he is
lying about), after the lie word, further no responses produce much
smaller physiological responses than no responses preceding the lie
word. By the same token, when we lie in the real world, we often can
feel our own body changing in response to the lie. Perhaps most sig-
nificant is that our interviews with professional polygraphers suggest a
dramatic change in physiological responses after a confession. For ex-
ample, if a suspect is initially polygraphed on a set of standardized
questions, he or she is judged to be deceptive based on exaggerated
physiological responses to key questions. However, if, after the first
series of tests, the suspect confesses to the crime, later physiological
responses to the same questions (to which he or she is now telling the
truth) disappear or are greatly reduced. Not only are the phasic re-
sponses to the key questions lower, but the person's overall autonomic
activity is reduced (despite the fact that the person may now face severe
punishment because of his or her confession).
The common thread linking all of these phenomena is that lingu-
istically organizing and, perhaps, communicating aspects of personal
lNHIBITlON AND CoGNITlON 109
traumas reduce the psychological and physiological aftereffects of trau-

mas. Conversely, if an individual is inhibited or prohibited from organ-
izing trauma-relevant information, individuals may become more sus-
ceptible to disease processes. As will be seen, these cognitive and
inhibitory processes extend beyond the traditional views of the effects
of major life events (e.g., Holmes & Rahe, 1967) or the buffering effects
of social support (Cobb, 1976).
In the first section of the chapter, we will present evidence indi-
cating that the effects of major life crises-specifically sexual trauma,
divorce of parents, and death of a parent-persist over several years in
the forms of illnesses and physical symptom reporting. The results from
several survey studies will indicate that the fundamental difference in
these traumas is the degree to which individuals confide in others. Fur-
ther, failure to confide or discuss the trauma with others can exacerbate
the physiological effects of traumas. The second major section of the
chapter deals with the general problem of inhibition. Starting from a
molecular level, evidence will be presented that indicates that the in-
hibition of ongoing behavior results in heightened physiological activity.
Further, personality research suggests that chronic inhibitors evidence
generally higher autonomic activity. We will then provide preliminary
evidence that indicates that certain types of traumas require that the
person actively inhibit disclosing his or her experiences. Hence, the
combined effects of inhibition and failure to assimilate traumatic ex-
perience cognitively produce maximal long-term physiological effects.
In the final section, we will discuss the relationship of our framework
to the psychotherapy, coping, and psychoanalytic literatures.
II. PERSONAL TRAUMAS AND HEALTH
The results of the studies to follow concern the impact of major

traumatic events several years after the events occurred. It is tempting
to claim that we had clear hypotheses and a conceptual framework from
the beginning. In fact, our earlier surveys were designed to study bu-
limia and related eating disorders (see Barrios & Pennebaker, 1983).
However, as our initially serendipitous findings were uncovered, the
present framework gradually emerged. Nevertheless, in order to give
the false impression of foresight, we will present the following studies
as though they represented carefully planned programmatic research.
On the basis of the overwhelming number of confirmatory studies,
there is little doubt that major life events can adversely affect health
(e.g., Kobasa, 1982). Most work within the life change literature indicates
that the risk of serious illness increases significantly as a function of the
number and severity of the life changes. Of particular relevance to the

present investigation, virtually all of the studies using standardized life
event scales, such as the Schedule of Recent Experience (Holmes & Rahe,
1967), have focused on changes in health for six months to two years
following major life events. With the exception of a small number of
retrospective studies (e.g., LeShan, 1966; Schmale & Iker, 1971), few
investigators have examined the long-term effects of a selected group
of major life changes on health patterns.
From a preliminary study of symptom reporting (Pennebaker, 1982),
we suspected that traumatic sexual experiences, divorce of parents, or
the death of a parent could have adverse health consequences over sev-
eral years. Common sense (as well as survey results to be discussed
below) dictates that each of these events would be viewed as maximally
traumatic (and therefore roughly equivalent) by the child or adolescent.
When we first started our research, we were simply interested in the
similarities and differences among these events in their relationship to
health indicators.
The results of three surveys are discussed briefly below. The first,
which was completed by 716 female undergraduates from the University
of Virginia and the University of Mississippi, was part of a 400-item
questionnaire that assessed symptom and emotion reports, family his-
tory, eating habits, dating patterns, and the like. The second survey was
based on readership response to a questionnaire published in Psychology
Today magazine dealing with health issues in general (Rubenstein, 1982).
Although over 24,000 individuals mailed their surveys to the magazine,
only the data from a random sample of 2,020 respondents were ana-
lyzed. The third questionnaire survey was, in fact, a series of surveys
given to one class (Medical Psychology) of 80 students at the University
of Virginia over the course of the semester. Sampling problems and
related issues will be discussed in greater detail after the results are
presented.
A. Initial Survey
All 716 subjects in the first survey were contacted in various psy-
chology classes and asked to complete the questionnaire without re-
ceiving credit. Although the response rate was only 67%, a random
sample of nonrespondents was contacted by phone to learn if they dif-
fered in major respects from the respondents. No differences were found
for demographic, health, or other variables between the respondent and
nonrespondent samples. The major reasons for not completing the sur-
vey were "not enough time" and "I forgot."
INHIBITION AND CoGNITION Ill
Subjects were divided into 4 groups: Control, Death of parent, Di-

vorce of parents, and Sex Trauma (hereafter designated as Control,
Death, Divorce, and Sex Trauma). Students were assigned to the Death
group if either or both of their parents had died by the time the re-
spondent was 17. The Divorce subjects had to have natural parents who
had been divorced or legally separated by the time they were 17. Subjects
were assigned to the Sex Trauma group based on their responding af-
firmatively to the question: "Prior to the age of 17, did you have a trau-
matic sexual experience (e.g., rape, being molested)?" There are two
immediate problems with defining the Sex Trauma group. The first con-
cerns whether or not the person honestly revealed that she had had a
traumatic sexual experience and, if so, what had been the nature of this
experience. Fortunately, we were able to conduct intensive interviews
with a subsample of 17 respondents. Of the three who had answered
the sexual trauma question in the affirmative, one had been raped and
two had been molested. We of course have no idea how many of the
Control subjects had had sexual traumas. Nevertheless, we feel that the
size of our Sex Trauma group is probably conservative.
The second problem is that some of the subjects had experienced
both a divorce and a sexual trauma, a divorce and a death, or all three.
In order to maximize cell frequencies, assignment to condition was prior-
itized in the order of sexual trauma, death, divorce. It should be noted
that subjects who had experienced a sexual trauma and a divorce and/
or death did not differ statistically from the sexual trauma only subjects
on any of the major health-dependent measures.
The primary dependent measures included the degree to which
subjects experienced selected physical symptoms and emotions on "the
average day," with 1 = never and 7 = almost always. Subjects were
also asked whether or not they had been to a physician or student health
center within the past eight months for illness. Additional questions
were included asking how many very close male and female friends they
had and, if they were in trouble or very emotionally upset, how likely
it would be that they would confide in no one (where 7 = extremely
likely).
Each of the dependent measures was subjected to a one-way an-
alysis of variance. As can be seen in Table 1, those who claimed to have
had a traumatic sexual experience were far more likely to report a variety
of physical symptoms and negative emotions than any of the other
groups. Although the Death and Divorce groups tended to have higher
means on these items than the Control group, simple contrasts (using
the mean-square error term) did not yield significant differences. Al-
though the physician visit item did not attain significance, this may have
been due to a ceiling effect since a relatively high proportion of all sub-
112 ]AMES W. PENNEBAKER AND CLAUDIA W. HoovER
TABLE 1
Self-Reports of Selected Variables by Type of Trauma for the Initial Study"
Divorce of Death of Sexual
Variable Control parents parent trauma Significance
Physical symptoms
Headache 2.6 2.8 2.6 3.3 .01
Upset stomach 2.3 2.3 2.3 2.8 .05
Tense muscles 3.0 3.0 2.3 3.8 .01
Racing heart 2.4 2.5 2.3 3.2 .01
Light headed 2.2 2.6 2.0 3.0 .01
Negative emotions
Stress 4.0 4.0 3.5 4.5 .04
Depressed 3.2 3.4 3.0 3.5 ns
Guilty 2.8 2.8 2.8 3.1 ns
Angry 2.9 2.9 2.7 3.2 ns
Isolated 3.2 3.0 2.9 3.7 .09
Fatigue 4.4 4.9 3.7 4.6 .01
Other data
Number of close 8.1 7.4 9.0 8.0 ns
friends
Confided in no one 2.6 2.5 2.4 3.6 .01
Visited health center 51.4 58.0 46.3 60.3 ns
or physician
(percent)
Age 19.4 19.6 19.3 19.8 ns
N 536 81 41 58
' Symptoms and emotions were coded along 7-point scale with 1 = never experience on the average
day and 7 = always experience on the average day. Probability levels are based on overall F from a
one-way analysis of variance.
jects responded affirmatively. Finally, although the Sex Trauma group

did not differ from the others in respect to the total number of friends,
it is noteworthy that they were most likely not to confide in anyone if
they were in trouble.
B. Psychology Today Sample

In the May, 1982, issue of Psychology Today, a four-page question-
naire dealing with health was printed asking the readers to complete it
and mail it to the magazine's main office. Fortunately, the editor in
charge of the survey, Carin Rubenstein, allowed us to include several
questions and to reanalyze the data. Within two months, over 24,000
individuals returned the completed survey (about 1.2% of the maga-
zine's circulation). Of the surveys, a random sample of 2,020 was ana-
INHIBITION AND CoGNITION 113
lyzed. The data indicated that the respondents mirrored the general
readership of the magazine (mean age = 35.1, 68% female, median
income around $25,000). At the outset, it should be noted that a survey
of this nature is not truly random nor does it represent a clearly defined
population. Individuals who completed the questionnaire may well have
been far more health-conscious than nonrespondents.
On the basis of the survey responses, four groups (Control, Death,
Divorce, and Sex Trauma) were defined in the same manner as in the
initial survey described above. Overall, 22% of the women and 10% of
the men reported having experienced a sexual trauma prior to the age
of 17. Of the 367 in the Sex Trauma group, 21% had also experienced
a divorce of parents and 11% had had a parent die prior to the age of
17. Separate analyses comparing those with a sexual trauma only with
those with a sexual trauma and divorce and/or death yielded virtually
no significant differences for any of the health-dependent measures.
Unlike the initial study, subjects' reports of illnesses, physical symp-
toms, and emotions were based on binary yes-no responses to the ques-
tion, "Which of the following bothered you in the past year?" These
data are expressed as percentages of respondents answering affirma-
tively in Table 2. Finally, 10 life change items adapted from the Schedule
of Recent Experience asking whether or not the subjects had experienced
each item within the past year (e.g., death of spouse or loved one, loss
of job). These items were summed to make a total life change score.
As can be seen in Table 2, most illnesses, symptoms, and negative
emotions occurred with greater frequency among all three trauma
groups. As in the first study, these effects were most pronounced for
the Sex Trauma group. It is of interest that the groups did not differ in
the number of close friends nor in their having someone to confide in.
Finally, inspection of the total life change scores indicates that the trauma
groups had experienced more life events than the Controls. Although
statistically significant, the magnitude of differences between groups is
relatively small. It should be noted that the group differences for the
health effects remain significant when the life change score is controlled
for.
These data raise a number of important issues. First, the magnitude
of differences in the various health indicators as a function of group is
large considering that these traumatic events took place close to 20 years
earlier. Second, there are undoubtedly a number of important mediating
factors. For example, the health differences could reflect poor health
practices (e.g., the Sex Trauma group was more likely to smoke and not
eat a balanced diet), more negative emotions (which could directly cause
health problems), or simply greater attentiveness to bodily sensations.
We cannot, however, explain the results simply in terms of differences
TABLE 2
Self-Reports of Selected Variables by Type of Trauma for the Psychology Today
Sample"
Condition
Divorce of Death of Sexual

Variable Control parents parent trauma p
illness and symptoms (percentage reporting in last year)
Ulcers 9.0 10.9 7.9 14.4 .01
Indigestion or upset stomach 39.9 50.7 36.3 50.1 .01
Chronic constipation 6.9 7.9 9.6 12.8 .01
Diarrhea 8.7 12.4 7.3 14.1 .01
Frequent headaches 24.2 30.3 28.4 32.7 .01
High blood pressure 8.3 6.4 10.8 11.1 .12
Heart or chest pain 8.5 8.9 7.9 13.9 .01
Dizzy spells 10.2 12.4 7.9 14.4 .06
Kidney or bladder infection 16.7 20.9 20.4 25.6 .01
Chronic lung disease or 4.7 8.4 6.2 8.1 .03
bronchitis
Influenza, throat or lung 49.7 57.2 48.3 52.8 .17
infections
Trouble getting breath 6.2 9.9 7.3 11.7 .01
Pain or numbness in muscles 11.1 14.9 14.2 16.8 .02
or joints
Menstrual problems 19.3 27.3 19.8 26.4 .01
Cognitive and emotional reports (percentage reporting):
Difficulty concentrating 25.0 28.8 24.4 31.8 .05
Tire easily 28.6 30.3 27.2 38.2 .01
Restlessness 20.3 28.8 23.3 26.7 .01
Fatigue 28.6 30.3 27.2 38.1 .01
Nightmares 10.2 15.9 14.7 20.7 .01
Depressed or blue 49.3 62.1 55.1 60.0 .01
Guilty 17.1 22.3 21.0 30.7 .01
Lonely 32.4 38.3 39.2 41.1 .01
Angry or irritable 47.2 53.7 50.5 54.5 .01
Irrational fears 11.2 9.4 14.7 18.8 .01
Loss of interest in sex 18.9 18.4 26.7 27.5 .01
Other data:
Number of close friends 6.1 5.3 5.4 5.6 .08
Do you now confide in close 1.4 1.4 1.5 1.5 .22
friends (1 = yes 5 = no)
Number of days sick in last 4.3 6.3 5.6 5.0 .07
year
Number of physician visits 3.1 3.9 3.3 3.8 .15
last year
Number of days hospitalized 0.8 1.0 1.8 1.7 .03
last year
Age in years 36.1 31.7 38.4 35.0 .01
N (males) 530 81 56 73
N (females) 746 120 120 294
Total N 1,276 201 176 367
" Probability levels (p values) are based on overall one-way analyses of variance for each of the de-
pendent measures by condition.
in present measures of social support or recent life crises. As will be

discussed below, these results may reflect the coping strategies that the
trauma (especially sexual traumas) subjects employed during and fol-
lowing their earlier trauma.
On the original questionnaire, respondents were asked to include
their name and telephone number for possible future telephone inter-
views. Rubenstein called 15 subjects who had claimed to have had a
sexual trauma. In her article, Rubenstein (1982) notes:
One woman was raped at 16; another was a victim of incest at 8; yet another
had been fondled at the age of 5 by a man selling ponies. A 51-year-old woman
from Los Angeles told me that she had been raped, at 5, by her neighbor,
who was a friend of the family .... "I never told anyone about it. You're
the first," she said. Later on, not making the connection, she remarked, "I've
always had health problems with organs in that area ... since I was 5."
(page 34)
Rubenstein (personal communication) notes that every person that

she talked with related an experience that would objectively be defined
as sexually traumatic. Further, she found that the majority had not dis-
cussed this traumatic event with anyone at the time and, if they even-
tually did, several months or years had elapsed in the interim. The
failure of conveying and, perhaps, organizing or assimilating such a
traumatic event may have long-term health implications. Before dis-
cussing this issue in detail, however, we must first present results from
a third survey study that bears directly on the confiding issue.
C. Medical Psychology Class Survey

Each year, the first author teaches an upper-division undergraduate
course entitled Medical Psychology that is restricted to 80 students. Al-
though the course has a higher percentage of pre-med students (38%)
than most higher level courses in psychology, their mean grade point
average, health center visits, and scores on standardized individual dif-
ference measures (e.g., I-E locus of control, Jenkins Activity Survey,
Self-consciousness Scale) are virtually identical with those of students
enrolled in Introductory Psychology and Social Psychology classes. The
course was comprised of 55 women and 20 men, with a mean age of
20.5. During the semester, students complete a number of question-
naires that are pertinent to the course. The data to be presented below
are based on the responses of 75 students enrolled in the class during
the Spring of 1983. Two of the questionnaires are particularly relevant.
The first assessed the types of traumas the students had had prior to
the age of 17, how upsetting each was, and the degree to which they
116 JAMEs W. PENNEBAKER AND CLAUDIA W. HoovER
confided in others about the trauma (all along 7-point scales with 1 =
not at all and 7 = a great deal). Separate questionnaires completed later
in the semester required students to report if they had had each of 20
serious health problems (e.g., asthma, pneumonia, diabetes). Finally,
students completed a number of individual difference measures that
assess: the general proclivity to report physical symptoms (i.e., PILL
from Pennebaker, 1982), self-consciousness (including Private Self-con-
sciousness, Public Self-consciousness, and Social Anxiety from Fenig-
stein, Scheier, & Buss, 1975), Type A behavior (the student version of
the Jenkins Activity Survey from Krantz, Glass, & Snyder, 1974), cog-
nitive and social anxiety (the CSAQ from Schwartz, Davidson, & Gole-
man, 1978), and social desirability (Marlowe-Crowne Social Desirability
Scale from Crowne & Marlowe, 1964). These data are included for com-
parison purposes with other samples and will not be discussed in detail.
Unlike the previous studies, the death of parent question was ex-
panded to include death of family member or very close personal friend.
The divorce question included either divorce or separation of parents.
Overall, 10 subjects reported having had a traumatic sexual experience,
17 had experienced the death of a family member or close personal
friend, and 12 had had parents who had been divorced or separated.
The mean ages at which these traumas occurred were 12.9 for sexual
trauma, 9.3 for death, and 12.0 for divorce. The mean ratings for how
upsetting the various traumas were (with 7 = extremely upsetting) were
5.1 for sexual trauma, 5.3 for death, and 5.8 for divorce. In response to
the question, "To what degree did you confide in others about this
event" with 7 = confided a great deal, the means for the various groups
were 2.1 for sexual trauma, 3.3 for death, and 3.8 for divorce (p < .01).
Initial analyses of variance using the four groups (Control, Death,
Divorce, and Sex Trauma) as independent variables and self-reported
disease, symptom reporting, health center usage, and social factors as
dependent measures yielded results similar to the previous studies. As
before, the Sex Trauma group reported experiencing the most number
of major diseases (3.1 versus 1.8 for Divorce, 1.1 for Death and 1.1 for
Controls, p < .01), and the most visits to the student health center in
the past year (3.6 versus 1.8, 1.1, and 1.3 for Sex Trauma, Divorce, Death,
and Control groups respectively, p < .01). In addition, all trauma groups
reported higher overall frequency and severity of symptom reporting
relative to controls (means for the groups: 23.0, 22.1, 20.1, and 14.4 for
Sex Trauma, Divorce, Death, and Controls respectively, p < .01). Finally,
the four groups did not differ in the total number of close friends that
they had (8.2, 6.8, 6.8, 8.3, p = .65).
A striking aspect of the response patterns for these data was the
marked variability in the degree to which people either did or did not
confide about each of the traumas. Even though the sexual trauma
groups in each of the above studies have evidenced the greatest health
problems, we assume that this is due, in part, to their not discussing
and subsequently organizing the event. We would predict that any trau-
matic event that was not talked about would have deleterious effects.
In order to test this idea, we divided our subjects in the present exper-
iment into three groups: No Trauma, Trauma and Confided, and Trauma
and Not Confided. That is, on the basis of their responses to the three
different traumas (whether sex, death, or divorce), if any trauma was
reported they were assigned to either the Confided or Not Confided
group by their self-reported degree of confiding of the relevant trauma.
In short, no distinction was made among types of trauma. (Subjects were
assigned to the No Confide condition if, on their confiding question,
they responded with a 1, 2, or 3 along the 7-point scale).
One-way analyses of variance were performed on the disease, symp-
tom reporting, and related variables. A priori contrasts using the mean-
square error term (one-tailed tests) were also computed for each de-
pendent measure. As can be seen in Table 3, having experienced a trau-
matic event and not having confided in others about it was associated
with the highest disease incidence, most number of visits to the student
health center for illness or injury, and overall symptom reporting. In
terms of specific diseases, this trend was significant for the following:
severe acne (p = .03), tumor (p = .05), migraine headache (p = .035),
severe injury (p = .12), ulcer (p = .06), severe depression (p = .02),
and appendectomy (p = .05). No differences were found for asthma,
high blood pressure, anorexia nervosa, cancer, mononucleosis, diabetes,
severe allergy, severe obesity, or other disorder (due in part to the ex-
tremely low frequency of these disorders among the sample).
The results of this study are comparable with the previous ones that
we have reported. First, those who have had a sexual trauma report the
most health problems. Second, there are no differences among the
groups in terms of the number of close friends that they have. Third,
the primary factor that separates the groups is an event that happened,
on the average, 7 years earlier.
The distinguishing characteristic of the present survey is that not
confiding in others about a traumatic experience, irrespective of the na-
ture of the experience, appears to underlie the differences in health
status among the trauma groups. Further, our results indicate that sexual
traumas are far less likely to be confided than are other types of major
uncontrollable events such as death or divorce of parents. We must now
stand back and address what it could be about confiding or not confiding
that could affect disease processes.
118 JAMEs W. PENNEBAKER AND CLAUDIA W. HoovER
TABLE 3
Self-reports of Selected Variables by Trauma and Confiding for Medical
Psychology Class Sample
No Trauma and Trauma and
Variable trauma confide not confide Significance
Number of Diseases 1.1 1.6 2.2 .03

Number of health center 0.7 1.2 2.1 .05
visits
Number of close friends 8.3 7.9 6.1 ns
Individual difference
measures:
Symptom reporting 14.4 21.9 20.6 .01
(PILL)
Self-consciousness scale:
Private 23.6 25.5 25.1 ns
Public 26.1 29.9 27.1 .04
Social anxiety 12.5 13.5 14.8 ns
CSAQ:
Cognitive anxiety 17.0 19.4 20.1 .07
Somatic anxiety 15.2 18.6 18.7 .01
Jenkins Activity Survey: 7.8 8.8 9.1 ns
(Student version)
Marlowe-Crowne social 13.7 12.3 11.3 .07
desirability scale:
Total N 36 23 16
III. THE FAILURE To CoNFIDE: THE RoLE oF INHIBITION
There are two important aspects of the confiding issue. The first
has been addressed indirectly by the social support and social compar-
ison literatures. For example, the act of confiding assumes that the per-
son has access to intimate others. Interacting closely with others can
serve to buffer the effects of life crises (e.g., Cobb, 1976). The person
receives feedback that he or she is loved and important. In addition, the
act of confiding serves as a way of obtaining the information that the
person's thoughts and actions are normal. That is, the person can com-
pare his or her feelings about the traumatic event with similar others.
Such a social comparison process should be anxiety-reducing in that the
individual learns that he or she is not as "abnormal" or unique as he
or she may have feared (Festinger, 1954; Valins & Nisbett, 1972). Note
that these approaches assume that the traumatic event is, in and of itself,
physiologically damaging in the long run but that social factors can re-
duce the effects of this process.
An alternative way by which to view the confiding results is that
INHIBITION AND COGNITION 119
actively inhibiting or suppressing knowledge of the event to others is

in itself physiologically stressful. Although the traumatic event un-
doubtedly does produce major physiological upheavals in the short run
(Selye, 1976) and, to a lesser degree, in the long run, the act of inhibiting
by not confiding in others may require additional physiological work.
In short, the inhibitory processes of actively not confiding together with
the normal physiological upheavals of the trauma itself sum to produce
maximal long-term stress which should ultimately be manifested in
disease.
If such a process occurs, we must first demonstrate that behavioral
inhibition is associated with heightened physiological activity. Second,
it must be established that the physiological correlates of inhibition per-
sist over long periods of time. Finally, if such an explanation accounts
for the higher illness rates among those who have had traumatic ex-
periences but have not confided these events, we must demonstrate that
failure to confide reflects an inhibitory process. In this section, we will
present evidence to support each of these necessary links. We will then
discuss some of the possible physiological mechanisms that may account
for the disease processes themselves. We must first, however, briefly
discuss the general nature of inhibition.
A. The Nature of Inhibition

Few concepts in all of psychology have had a more sordid history
than inhibition. Among its difficulties are its different meanings and the
responses to which it has been applied across the various subdisciplines
of psychology. For example, on a neuronal level, the firing of one nerve
may decrease the probability of other neurons' firing. Similarly, recip-
rocal inhibition allows one group of muscles to contract while antagon-
istic muscles must not contract (e.g., when extending one's arm). In
visual and auditory perception, the firing of one group of cells may block
the firing of adjacent cells in order to enhance contrast, as in lateral
inhibition (Bekesy, 1967). In addition, certain afferent fibers can be
blocked by the cortex to aid in orienting (Hernandez-Peon, 1960) or in
reducing transmission of pain signals (Melzack, 1973). Pavlov (1927)
developed an entire taxonomy of inhibitory processes to explain phe-
nomena such as extinction of a conditioned response over time and
spontaneous recovery (internal inhibition) and the failure of a condi-
tioned stimulus to elicit a conditioned response due to the presence of
a novel stimulus (external inhibition). Spence (1936) and Hull (1950)
viewed conditioned inhibition as the main factor responsible for the
weakening of a conditioned response. One of the major criticisms of the
use of the term inhibition was leveled by Skinner (1938), who argued
that the reduction in the intensity of a response could more parsimon-
iously be explained by a reduction in excitation. Later evidence that
strongly supported the existence of active inhibitory processes has been
put forward by a number of animal researchers, including Rescorla
(1967), Konorski (1967), and Terrace (1972). Inhibition constructs have
also been used among verbal learning experts to explain differential
encoding and recall of word lists (Melton, 1963). Freud (1926/1959)
viewed inhibition as a restriction of the ego which has been either im-
pose':l as a measure of precaution or brought about as a result of an
impoverishment of energy.
Although each of the above uses of the inhibition concept has been
applied to vastly different phenomena, the concept's usefulness is in
describing a general process wherein a given event does not occur due
to the action of another event. Within the present context, we infer that
inhibition is occurring when a person is unable or unwilling to behave
in a given manner (e.g., disclosing, confiding) for fear of perceived neg-
ative sanctions (humiliation, punishment, etc.). Further, if these neg-
ative sanctions were not perceived to be present, the specific behavior
would occur. In short, inhibition in this context is an active process (cf.
Terrace, 1972).
Some supplementary remarks about this definition in regard to cog-
nitive processes are in order. In line with current thinking about per-
ceptual, cognitive, and attribution processes (Gibson, 1979; Kelley, 1972;
Neisser, 1976), we assume that individuals seek to understand their
environments and, in the case of humans, themselves. When an un-
expected and potentially self-relevant event occurs, it behooves the in-
dividual to understand and cognitively assimilate the event in order to
maintain some degree of predictability and control over the environ-
ment. Talking, writing, or even thinking about an important and com-
plex event is required in order to understand it (e.g., Jourard, 1971).
Confiding in others about a traumatic event should hasten this process.
Further, it allows them to receive information about potential coping
strategies (Lazarus, 1966). If the individual must actively inhibit this
assimiliation, both the work of inhibition and the assimilation process
should sum to produce long-term physiological stress. We will return
to these cognitive issues in the final section of the paper.
B. Inhibition and Physiological Activity

If our model is to be viable, we must first demonstrate that inhibition
is associated with physiological activity. For several years, Jeffrey Gray
has been examining the physiological substrates of inhibitory and ex-

citatory processes in animals. In an impressive integration of the liter-
ature, Gray (1975) has shown that in passive avoidance paradigms
(wherein the animal must not perform a learned behavior in order to
avoid punishment) or in temporal inhibition (wherein the organism
must not respond for a certain amount of time in order to be reinforced,
as in DRL procedures), there is an increase in activity of the septal and
hippocampal regions of the brain, which is referred to as the Behavioral
Inhibition System (BIS). When the activity of these brain systems is
blocked by drugs or destroyed, the rats exhibit a marked decline in the
ability to inhibit behaviors. Interestingly, certain drugs such as am-
phetamines or caffeine appear to increase the activity of the BIS.
In a compelling extension of Gray's work, Fowles (1980) proposed
that activity of the BIS was associated with heightened electrodem1al
activity in humans. The counterpart of the BIS, the behavioral activation
system (BAS), was posited to correlate with cardiovascular activity.
Fowles bases his argument on a number of human psychophysiological
investigations. For example, passive avoidance paradigms with humans
result in changes in electrodermal activity (EDA) but not in heart rate.
Active avoidance, on the other hand, is associated with increased heart
rate and blood pressure but not with EDA. Although we will return to
this issue in dealing with chronic inhibitors and noninhibitors later, it
is important to emphasize the link between behavioral inhibition and
specific physiological activity.
Particularly relevant to the present argument is the expanding lit-
erature related to the psychophysiological correlates of deception (see
Waid & Orne, 1981, 1982). One paradigm that has been employed ex-
tensively is the guilty knowledge test (GKT) wherein a subject is induced
to lie while various autonomic measures are monitored. One variation
of this procedure requires the subject to write down a number between
1 and 10. The experimenter then asks, "Is the number 1?""Is the number
2?" and so on. The subject is instructed to answer no to each of the
questions, including the number that was actually written down (the
lie word). Researchers are able to detect the lie word at rates far above
chance. In addition, of all autonomic channels, EDA appears to be the
most reliable indicator of deception (Waid & Orne, 1981). Drugs that
Gray (1975) posits should decrease inhibitory tendencies, such as tran-
quilizers, tend to lower EDA responses to deception (Waid & Orne,
1981). We can tentatively infer, then, that phasic inhibitory responses
are associated with autonomic activity, especially EDA.
At this point, we must be able to demonstrate that deception is
associated with inhibition of behavior and at the same time with a cor-
responding increase in physiological activity, such as EDA. In a recent
122 JAMES W. PENNEBAKER AND CLAUDIA W. HOOVER
experiment, Carol Chew and the first author (Pennebaker & Chew, in
press) directly tested this link. In the experiment, 25 women and 5 men
participated in a variant of the GKT on two occasions. All subjects se-
lected one of five words printed on separate index cards. They then
denied holding each of the five words while skin resistance was con-
tinuously monitored. Throughout the procedure, the experimenter (who
was blind to the lie word) called out the possible words from an adjacent
room. Following the initial test, which was identical for all subjects, 20
subjects were randomly assigned to the Observe condition and 10 to the
No Observe. In both conditions, subjects were given a second group of
five words to choose from. For the Observe subjects, a second experi-
menter sat directly in front of them and explained that he was attempting
to guess which word they had chosen by watching their behavior. For
the No Observe subjects, the second experimenter merely said that the
general procedure would be the same as before and then left the ~oom.
For the Observe subjects, the experimenter closely watched the sub-
ject's face throughout the second GKT. He pressed one button whenever
there was a change in eye gaze and a second one whenever there was
a change in facial expression (inter-rater reliability of this procedure
using videotapes is high, r = .87). The observer's button-press re-
sponses were recorded simultaneously with the subject's skin resistance
on a polygraph in the adjacent room. As in the initial session, the first
experimenter called out the words from the adjacent room, 14 seconds
apart.
Skin resistance level and button presses corresponding to eye move-
ment (EM) and change in facial expression (FE) were coded every two
seconds for all words across both sessions. The skin resistance measures
were converted to skin conductance and then standardized on a within-
subject basis to allow simple comparison with EM and FE. The various
measures were tabulated separately for the lie word and the mean of
the four truth words.
Between-within analyses of variance on the skin conductance scores
indicated that the overall skin conductance level (SCL) increased from
the first session to the second for the Observe and decreased for the No
Observe subjects (p < .001). As can be seen in Figure 1, SCL was sig-
nificantly higher for the lie word than for the truth words (p < .001),
especially during the 0-4 seconds after their lie or truth response. More
revealing, however, was the relationship between the overt behaviors
and SCL for the Observe subjects (behavioral measures were, of course,
not obtained in the No Observe conditions).
As depicted in Figure 2, there was a significant reduction in be-
haviors for the lie word relative to the truth words (p < .02). Closer
examination of the behavioral data reveals that subjects were least likely
5.6~------~------------------------------------------------,
5.4
5.2
(/)
0
I
::>
0
cr:. 5
~
::>
~
...J
(,)
4.8
(/)
4.6
4.4
0 2 4 6 8 10 12 14
SECONDS DURING FIRST GKT
5.6
5.4
5.2
------- ___o- __ - --- -- -o.._ --------0-------- --o... ----
5
4.8
4.6
---------.6---------.:1--
4.4
0 2 4 6 8 10 12 14
SECONDS DURING SECOND GKT
FIGURE 1. Skin conductance levels at each 2-second interval for the first and second guilty
knowledge tests. Subjects' "no" responses occurred at the 2-second point. 6-£:, No ob-
serve-Lie word; 6---6 No observe-truth words; 0-0 Observe-Lie word; 0---0
Observe-Truth words.
0.7
0.6
(I]
--4,
a: 0.5 '<
Q //~---------o------ ',,
>
<
:I:
. . . .,. . . "'0----------o
w
"'0
LL 0.4 1///
a:
w
"'
:::;:
::::>
z 0.3
0.2
0.1
0 2 4 6 8 10 12 14
SECOND BLOCKS DURING SECOND GKT
FIGURE 2. Mean number of behaviors (changes in eye movement and facial expression)
during each 2-second time block during the second guilty knowledge test (Observe subjects
only). 0-0 Lie word; 0---0 Truth words.
to change eye gaze or facial expression during the 0-4 seconds following
their verbal response on the lie word relative to the truth word (p <
.05). This behavioral inhibition coincided with the greatest increase in
SCL. Although one might be tempted to speculate that the reduction in
behavior caused the increase in SCL in a hydraulic manner, we prefer
to think that the concomitant changes in these measures reflect a cen-
trally mediated inhibitory process. That is, whenever one must inhibit
ongoing activity, there is a corresponding physiological change me-
diated, perhaps, by the septal-hippocampal brain regions. A final aspect
of these dat~ must be mentioned. It is significant that the behavioral
responses did not increase in intensity above baseline levels following
the lie. This suggests that behavioral tendencies are not like a boiling
caldron that, when suppressed, build up steam and explode once the
lid is removed. Unlike a Freudian view of energy conservation and leak-
age, behavioral inhibition is an active process involving work that is
itself measureable on a physiological level.
C. Long-term Inhibition: Personality Considerations

The research reviewed above indicates that short-term behavioral
inhibition is correlated with phasic increases in autonomic activity. What
evidence do we have to suggest that individuals who chronically inhibit

their behaviors manifest higher overall physiological levels? The re-
search evidence to support this idea derives from three areas: levels of
socialization (including sociopathy), repressive coping styles, and emo-
tional expression. Each will be discussed separately.
1. Level of Socialization
Many researchers have explicitly or implicitly assumed that indi-
viduals can be categorized along a dimension of socialization level (e.g.,
Gough, 1960; Trasler, 1978; Waid & Orne, 1981). On the lower extreme
would be the sociopath, an individual who is impulsive, selfish and
irresponsible, with a low tolerance of frustration (cf., Hare, 1978). At
the other extreme is the individual who adheres to social norms and,
in general, lacks sociopathic tendencies. For a number of reasons, far
more research has focused on those who are poorly socialized than on
those who are highly socialized. One hallmark of the sociopath is his
or her general behavior patterns suggesting a deficit in inhibition. The
sociopath typically performs poorly on passive avoidance and delayed
reinforcement tasks (see Hare, 1978, and Fowles, 1980, for reviews). In
addition, the sociopaths appear to have a low tolerance for alcohol
(Cleckley, 1976). Of particular import is the fact that sociopaths have
generally been found to have lower tonic SCLs and, during noxious
stimulation, reduced phasic EDA (Hare, 1978).
Complementing this work, Waid and Orne (e.g., 1981, 1982) have
examined subjects who are classified either as highly or poorly socialized
as measured by the Socialization scale of the CPI developed by Gough
(1960). Across a number of studies, these researchers find that poorly
socialized college students consistently yield smaller skin conductance
responses to deception than highly socialized subjects. If we make the
inferential leap that those individuals who are highly socialized tend to
inhibit behaviors in general, we can see how chronic inhibition is cor-
related with greater physiological activity (at least that related to the BIS)
in general.
2. Repressive Coping Styles
Within the humanistic and psychoanalytic literatures, an individual

who is defined as a repressor is one who excludes from consciousness
both cognitive and emotional factors that are potentially threatening to
the ego (Ross & Abrams, 1965). Such an individual may deny that he
or she is anxious or upset even though behavioral and physiological
measures would indicate otherwise. Early work on repressors found that
they had higher recognition thresholds for anxiety-producing stimuli
126 JAMES W. PENNEBAKER AND CLAUDIA W. HoovER
(e.g., Eriksen, 1966). One problem that has impeded empirical research
on this personality measure has been the lack of ways of assessing
whether or not a person is, in fact, repressing.
An intriguing approach to this problem was proposed by Crowne
and Marlowe (1964), who developed the Social Desirability Scale. In-
dividuals who score high on this scale publicly adhere to social norms
to an extreme degree. Further, these authors indicate that such indi-
viduals are affectively inhibited and defensive. A recent important study
by Weinberger, Schwartz, and Davidson (1979) sought to define re-
pressors as those who score high on the Social Desirability Scale and
low on the Taylor (1953) Manifest Anxiety Scale (MAS). These individ-
uals were separated from those who were defined as high anxious (from
the MAS-irrespective of social desirability) and truly low anxious (low
MAS and low social desirability). During the actual experiment, subjects
participated in a phrase association task wherein they had to respond
to aggressive, sexual, or neutral sentences with a phrase that completed
the thought expressed in each stimulus sentence. In addition to reaction
time, heart rate, skin resistance, and forehead muscle tension were
recorded during the task. Overall, repressors exhibited significantly
higher forehead muscle tension and skin resistance responses than
either the high or low anxious group throughout the task. Both the
repressors and high anxious subjects evidenced elevated heart rate rela-
tive to the low anxious subjects.
These results are important for two reasons. First, they suggest that
a repressive coping style as measured by the Social Desirability Scale
(which could be viewed as analogous to a high level of socialization) is
associated with increased physiological responding. Second, measures
of repression have been associated with cancer (Kissen, 1966), high
blood pressure (Davies, 1970), and physical disease-in general (Black-
burn, 1965). We will return to the relationship between repression and
inhibition in a later section.
3. Emotional Expression
The relationship between the expression of emotion and the au-
tonomic correlates of the emotion has been an area of intensive inves-
tigation. Beginning with the pioneering work of Jones (1960), it was
found that the more emotionally expressive a child was, the smaller his
or her electrodermal responses. According to Jones, the act of sociali-
zation forced the child to inhibit emotional expression, outwardly caus-
ing the efferent discharge to occur along "hidden" physiological chan-
nels. Jones further notes that the socialization process encourages the
inhibition of overt responding.
One aspect of the emotional expression research concerns the re-

lationship between "sending accuracy" along nonverbal channels and
autonomic activity (EDA and heart rate). Sending accuracy is measured
by the ability of observers to detect a person's emotional state by viewing
facial expression. In an intriguing series of studies, Buck (1979) reports
large and consistent negative correlations between sending accuracy of
the sender and the sender's EDA and heart rate. Similarly, Notarius and
Levenson (1979) found that subjects who were classified as natural in-
hibitors (on the basis of facial expressiveness) were more physiologically
reactive in terms of heart rate, skin conductance, and respiration rate
when exposed to threat of a shock than were subjects classified as natural
expressers. These data are in line with Jones's findings that the more
expressive, the lower the physiological response.
It should be noted that a number of investigators have found sig-
nificant positive rather than negative relationships between facial ex-
pressiveness and various autonomic indices on a within-subject (as op-
posed to a between-subject) basis (e.g., Laird, 1974; Lanzetta,
Cartwright-Smith & Kleck, 1976). Whether these inconsistencies reflect
methodological differences is beyond the scope of this paper.
4. Summary
Each of the lines of research discussed above is relevant to the gen-

eral inhibition concept. Individuals who are highly socialized, high in
social desirability, and lacking in emotional expressivity evidence higher
tonic and phasic EDA and, for some studies, heart rate and muscle
tension. Unfortunately, none of the studies cited above have examined
illness rates or disease processes. Although there is some evidence that
repressors are more prone to hypertension, cancer, and disease states
in general, we would expect that traumatic experiences would interact
with each of these individual difference measures to increase the overall
probability of illness.
It must be emphasized that the link between each of these individual
difference measures and inhibition processes is not firmly established.
Whether or not the individual who is characterized as exhibiting a re-
pressive coping style, lacking in emotional expressiveness, and/or be-
having in a highly socialized way would be less likely to confide fol-
lowing a traumatic experience is an empirical question worthy of future
research. Common sense would argue that each of these personality
types would be more hesitant to confide simply because of fears of social
sanctions.
D. Failure to Confide as Inhibition
To this point, we have demonstrated that short-term behavioral

inhibition is associated with increased physiological activity. We have
presented data from a number of studies indirectly suggesting that long-
term inhibition, as measured by specific individual difference variables,
is associated with overall higher autonomic activity. Together, these two
lines of research indicate that we can view inhibition as physiological
work or, in Selye's (1976) terms, as a type of stressor. What evidence
do we have to indicate that not confiding about a traumatic event rep-
resents an inhibitory process? Unfortunately, our research into this ques-
tion is merely in the planning stage. A small number of in-depth inter-
views that we have conducted together with sporadic case studies in
the literature strongly indicate that not confiding does, in fact, require
considerable inhibitory effort.
Three recent in-depth interviews of adult women who had had sex-
ually traumatic experiences and did not confide these events to anyone
offer interesting insights. Person A was molested by her stepfather be-
tween the ages of 11 and 14. She noted that she simply could not tell
her mother because "it would break her heart." She wanted to tell her
friends, but she knew they would never understand. Another reason
for not telling anyone was that she felt that it was her fault that she was
being molested. Person B witnessed her mother in a sexualliason with
a friend of the family at age 13. She noted that she desperately wanted
to tell her father and often thought and even dreamed of doing so. In
fact, she said that she had "to actively restrain from telling Dad on those
(frequent) days when there was a family fight." Person C was sexually
assaulted by a cousin at age 9. After the event, she became shy and
withdrawn. Although she wanted to tell somebody about the event, she
assumed that she would be punished for provoking it. In each of these
cases, the victim constantly thought about the event for years, wanted
to tell others, but did not feel that she could.
The literature on rape and sexual assault confirms these anecdotal
reports. Binder (1981) conducted a survey of adult women dealing with
sexual assault. Although her response rate was only 33%, of those re-
porting having been sexually assaulted, only about 15% reported the
assault to the police. Unfortunately, the author does not report how
many of these women did not tell anyone. Nevertheless, the case quo-
tations within the article suggest that many did not confide in anyone
because of embarrassment, guilt, fear of the offender, or a feeling that
it was the victim's own fault. Burgess and Holmstrom (1974) studied
the reactions of 92 adult women rape victims and identified a general
rape trauma syndrome which includes increased somatization, phobic

reactions, self-blame, and nightmares (see also Kilpatrick, Resick, & Ve-
ronen, 1981, for discussion of long-term somatic and emotional distur-
bances of rape victims). In their article, the authors point out that a
significant proportion of rape victims do not report the event to anyone.
Hence, clinicians should be especially attentive to rape trauma
symptoms.
We readily admit that case reports are the last refuge of scientific
scoundrels. We feel, however, that the overall picture suggests that fail-
ure to confide a traumatic event represents a significant risk to one's
health.
IV. SuMMARY AND FINAL CoNSIDERATIONS
People seek to understand their world. To do this, they must often

organize and assimilate complex and personally traumatic events. One
of the most efficient and natural ways of accomplishing this is by trans-
lating thoughts and feelings into language. The traumatic event and the
assimilation process themselves can be viewed as direct stressors on the
body. On occasion, individuals may not be able to accomplish this or-
ganization process for fear of punishment, humiliation, or embarrass-
ment. Actively to inhibit behavior requires additional work which is
manifested in physiological activity. The longer one must inhibit this
cognitive organization process and the greater the magnitude of the
trauma, the more the inhibitory processes are engaged. Over time, the
combined physiological stresses of the organization process and the in-
hibitory process produce increased wear and tear on the body, thus
increasing the probability of disease.
Although a great deal of evidence has been put forward in support
of this general conceptual framework, there are a number of empirical
issues that remain to be resolved. At this point, we are not able to specify
the precise physiological mechanisms that underlie the inhibitory pro-
cess. Gray's (1975) hypotheses concerning relevant brain structures as-
sociated with inhibition offer a promising beginning. Undoubtedly,
other central nervous system structures are involved. We must also
await research linking these inhibitory processes to immune, hormonal,
and autonomic nervous system functions. At this point, it is premature
for us to "physiologize" too much. We can, however, speculate about
the relationship of our framework to a number of related research en-
deavors relevant to psychotherapy, self-disclosure, and coping pro-
cesses in general.
A. Inhibition and Psychotherapy

There is little doubt that psychotherapy is an effective way by which
to reduce anxiety and even the incidence of psychosomatic disease. Jer-
ome Frank (1974) and a number of recent researchers (e.g., Nicholson
& Berman, 1983; Smith, Glass, & Miller, 1980) indicate that virtually all
forms of psychotherapy (whether cognitive, behavioral, or psychodyn-
amic) yield significant improvements in patients' self-reports and be-
haviors by the end of treatment which continue to persist several months
or years after treatment. We suggest that one of the benefits of psycho-
therapy is that it promotes the cognitive reorganization and assimilation
of important and, on occasion, traumatic events. In accomplishing this
goal, the person no longer needs to inhibit his or her urge to assimilate
the event by talking with others (e.g., Pennebaker & O'Heeron, 1984).
Much of psychotherapy, then, reduces stress directly by allowing the
cognitive reorganization process to occur and, indirectly, by nullifying
the need to inhibit (see also excellent discussions of similar processes
by Horowitz, 1976, and Silver, Boon, & Stones, 1983). One can readily
see how this process could also be accomplished through confession,
talking to a stranger on a plane, or confiding in a close friend.
Note that we are in no way denying other benefits of psychother-
apy, such as the acquisition of new coping skills, stress management
techniques, and the like. Many problems that psychotherapists must
deal with have been previously disclosed and assimilated but continue
to persist, for example, snake phobias. In short, the processes on which
we have focussed represent only a portion of the variables that are at
play within the psychotherapeutic mileu.
B. Cognitive Organization and Self-disclosure

In his book, The Transparent Self, Jourard (1971) argues that the act
of self-disclosure allows for one's feelings and thoughts to become more
concrete. Self-disclosure, then, helps the person to know himself better.
Interestingly, Jourard posited that self-disclosure had motivational prop-
erties, and if this motive was blocked adverse physiological problems
would result (see also Archer, 1980, for an excellent review).
Neither Jourard nor we think that public self-disclosure is the only
method by which to organize and assimilate self-relevant information.
For example, expressing one's experiences on paper may be almost as
effective in allowing for the assimilation and subsequent reduction in
inhibition processes to occur. We base this idea on both personal ex-
periences and a recent pilot study.
In the experiment (conducted by Melissa Conrad and the first au-

thor), subjects wrote two essays while heart rate was continuously mon-
itored. Half of the 16 subjects were required to write their first essay
about an intensely personal experience that they had never told anyone.
We assured the subjects that they could keep the essays and that we
would not read them. The remaining subjects wrote about what had
happened to them the day before the experiment. For the second essay,
all subjects were to describe how their bedroom was decorated. Those
writing a personal essay evidenced an increase in heart rate during the
initial essay but a significant heart rate decrease (below initial baseline)
during the second essay. The control subjects' heart rates did not deviate
from basal heart rate levels throughout the experiment. Finally, self-
reports of physical symptoms dropped significantly from the beginning
of the study to the end for the personal essay subjects but were un-
changed for the control group.
Ther~ are a number of alternative explanations for these findings
other than increased assimilation and decreased inhibition (e.g., op-
ponent process, contrast effects). Nevertheless, from a theoretical per-
spective, such a treatment approach involving the use of personal diaries
or journals may have merit.
C. Links to Other Theories

Many aspects of our approach have been suggested by others in
one form or another. In this final section, we will focus on the similarities
and differences between the present perspective and those within the
psychodynamic and coping literatures.
a. Psychodynamic Views of Inhibition. Within the psychodynamic lit-
erature, a number of terms are used that bear a certain degree of sim-
ilarity to our use of inhibition. For example, the defense mechanism of
suppression is invoked to define the active process of putting certain
disturbing thoughts out of one's mind. Freud's use of the inhibition
construct is similar to ours in that it requires work or energy (Freud,
1926/1959). At this point, the psychodynamic perspective of inhibition
and our own diverge. Inhibition and suppression, according to a Freud-
ian analysis, represent work in that they oppose some intrapsychic pro-
cess. They act to allow the ego to repress forbidden desires. Note that
in Freud's discussion of symptoms and Alexander's (1950) analyses of
illness problems develop when these defense mechanisms fail. Hence,
disease processes are not the result of effectively used psychological
processes such as inhibition or suppression. Our model posits that the
inhibitory process itself is stressful.
132 JAMES W. PENNEBAKER AND CLAUDIA W. HoovER
As we use the term, an active process must be engaged in order

not to do or say something. This is the central feature in the connection
between life traumas and illness. For Freud, the life trauma itself may
be reminiscent of an earlier conflict which has been repressed, thus
bringing the long-since buried ideational content to the psychic surface,
which itself is damaging. Once it surfaces, either the ego must again
perform the energy-consumptive task of repressing the ideational con-
tent or, if this process is incomplete, neurotic symptoms should develop.
In short, Freud focuses on inhibition (or suppression) in the service of
avoiding potentially greater intrapsychic conflict, whereas we have
pointed to the process of inhibition itself as a potential stressor.
b. Coping and Appraisal. An important issue in the link between life
traumas and disease is the role that cognitive style takes in mediating
the effects of stress. Lazarus and his colleagues (e.g., Lazarus, 1966;
Folkman & Lazarus, 1982) have developed a broad theory of stress and
coping which posits a person's cognitive appraisal of threats as a critical
determinant of the coping process. The theory posits that coping is de-
termined by three appraisal processes. The primary appraisal concerns
the evaluation of the consequences of a situation for the individual's
immediate well-being. The secondary appraisal process entails an eval-
uation of what can be done about the situation, that is, a mental "taking
stock" of available resources that may ameliorate the effects of the stres-
sor(s). After these appraisal processes have initially occurred, a third
process-reappraisal-is initiated that allows for the assessment of how
accurately the first two processes reflect the situation. At this point, a
person may reassess and change his or her earlier assumptions and
decisions.
Following the appraisal process, a coping strategy is adopted. Folk-
man and Lazarus (1982) posit that coping strategies will be directed
either to the problem itself or to the emotional reaction to the problem.
The four general modes of coping include direct action, information
seeking, inhibition of behavior, and cognitive changes that could include
restructuring, rationalization, and so forth. Although the bulk of re-
search on these coping processes has been concerned with the relatively
short-term behavioral and physiological responses to both major and
minor life experiences (e.g., death of a child, reactions to surgery), our
work suggests that we should begin focusing on coping with personal
traumas that may take place for years after the event. Clearly, after
several months or years following a trauma, the only coping strategies
that would most likely be employed would include inhibition of behavior
and, to a lesser extent, continued cognitive changes.
D. Summary
In this chapter, we have argued that individuals seek to understand
and organize personally traumatic events. Certain events, such as rape,
death of a parent, or divorce of parents, require considerable effort to
understand and assimilate. Some major events cannot easily be dis-
cussed with others because of perceived negative social sanctions, em-
barrassment, or fear. Not being able to confide about such events re-
quires the person actively to inhibit his or her behavior. As we have
indicated, behavioral inhibition is associated with physiological activity.
Much of our earlier work suggests that victims of sexual assault are often
loathe to discuss the event with anyone. These individuals-who, we
are hypothesizing, must actively inhibit their urge to tell others of their
experience-are far more likely to report a number of physical symp-
toms, illnesses, and use of medical facilities. Although life stressors such
as these are themselves physiologically damaging, the act of inhibition
compounds the wear and tear on the body.
The research that we have presented is just beginning. Many of the
unifying links in our model are based on speculation or case studies.
Although we are certain that a number of revisions will be necessary in
the future, we are optimistic about the theoretical and clinical implica-
tions of our initial findings.
REFERENCES
ALEXANDER, F. (1950). Psychosomatic medicine: Its principles and applications. New York:
Norton.
ARCHER, R. L. (1980). Self-disclosure. In D. M. WEGNER & R. R. VALLACHER (Eds.), The self
in social psychology. New York: Oxford, pp. 183-205.
BARRIOS, B. A., & PENNEBAKER, J. W. (1983). A note on the early detection of bulimia
nervosa. The Behavior Therapist, 6, 18-19.
BEKESY, G. V. (1967). Sensory inhibition. Princeton, NJ: Princeton University Press.
BINDER, R. (1981). Why women don't report sexual assault. Journal of Clinical Psychiatry,
42, 437-438.
BLACKBURN, R. (1965). Emotionality, repression-sensitization, and maladjustment. British
Journal of Psychiatry, 111, 399-400.
BucK, R. W. (1979). Individual differences in nonverbal sending accuracy and electrodermal
responding: The externalizing-internalizing dimension. In R. RosENTHAL (Ed.), Skill
in nonverbal communication. Cambridge: Oelgeschlager, Gunn & Hain, pp. 44-67.
BuRGEss, A. W., & HoLMSTROM, L. L. (1974). Rape trauma syndrome. American Journal of
CLECKLEY, H. (1976). The mask of sanity. St. Louis: Mosby.
Coss, S. (1976). Social support as a moderator of life stress. Psychosomatic Medicine, 38,
300-314.
CROWNE, D.P., & MARLOWE, D. (1964). The approval motive: Studies in evaluative dependence.
New York: Wiley.
DAVIES, M. (1970). Blood pressure and personality. Journal of Psychosomatic Research, 14,
89-104.
ERIKSEN, C. W. (1966). Cognitive responses to internally cued anxiety. In C. D. SPIELBERGER
(Ed.), Anxiety and behavior. New York: Academic Press, pp. 317-360.
FENIGSTEIN, A., ScHEIER, M. & Buss, A. (1975). Public and private self-consciousness: As-
sessment and theory. Journal of Consulting and Clinical Psychology, 43, 522-527.
FESTINGER, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-
140.
FoLKMAN, S., & LAZARUS, R. S. (1982). Stress and coping theory applied to the investigation
of mass industrial psychogenic illness. In M. ]. COLLIGAN, ]. W. PENNEBAKER, & L.
MuRPHY (Eds.), Mass psychogenic illness: A social psychological perspective. Hillsdale, NJ:
Erlbaum, pp. 237-255.
FowLES, D. C. (1980). The three arousal model: Implications of Gray's two-factor theory
for heart rates, electrodermal activity, and psychopathy. Psychophysiology, 17, 87-104.
FRANK, J. (1974). Persuasion and healing. New York: Schocken Books.
FREUD, S. (1926/1959). Inhibitions, symptoms, and anxiety. In J. Strachey (Ed.), Complete
psychological works of Sigmund Freud (Vol20). London: Hogarth, pp. 77-174.
GIBSON,]. J. (1979). The ecological approach to visual perception. Boston: Houghton-Mifflin.
GouGH, H. G. (1960). Theory and measurement of socialization. Journal of Consulting Psy-
chology, 24, 23-30.
GRAY,]. A. (1975). Elements of a two-process theory of learning. New York: Academic Press.
HARE, R. D. (1978). Electrodermal and cardiovascular correlates of psychopathy. In R. D.
HARE & D. ScHALLING (Eds.), Psychopathic behavior: Approaches to research. New York:
Wiley.
HERNANDEZ-PEON, R. (1960). Neurophysiological correlates of habituation and other man-
ifestations of plastic inhibition (internal inhibition). In H. Jasper & G. Smirnov (Eds.),
The Moscow colloquium on electroencephalography of nervous activity. EEG Clinical Neuro-
physiology, Supplement 13, 101-112.
HoLMES, T. & RAHE, R. (1967). The social readjustment rating scale. Journal of Psychosomatic
Research, 4, 213-218.
HoROWITZ, M. J. (1976). Stress response syndromes. New York: Jacob Aronson.
HuLL, C. L. (1950). Simple qualitative discrimination learning. Psychological Review, 57, 303-
313.
}oNES, H. E. (1960). The longitudinal method in the study of personality. In I. IscoE & H.
W. STEVENSON (Eds.), Personality development in children. 'thicago: University of Chicago
Press.
}OURARD, S. M. (1971). Self-disclosure: An experimental analysis of the transparent self. New
York: Wiley.
KELLEY, H. H. (1972). Attribution in social interaction. New York: General Learning Press.
KILPATRICK, D. G., REsiCK, P. A., & VERONEN, L. J. (1981). Effects of rape experience: A
longitudinal study. Journal of Social Issues, 37, 105-122.
l<IssEN, D. M. (1966). The significance of personality in lung cancer in men. Annals of the
New York Academy of Science, 125, 820-826.
KoBASA, S. (1982). The hardy personality: Toward a social psychology of stress and health.
In G. S. SANDERS & ]. Suis (Eds.), Social psychology of health and illness. Hillsdale, NJ:
Erlbaum, pp. 3-32.
KoNORSKI, ]. (1967). Integrative activity in the brain. Chicago: University of Chicago Press.
KRANTz, D., GLASS, D. & SNYDER, M. (1974). Helplessness, stress level and the coronary-
prone behavior pattern. Journal of Experimental Social Psychology, 10, 284-300.
LAIRD, J. D. (1974). Self-attribution of emotion: The effects of expressive behavior on the

quality of emotional experience. Journal of Personality and Social Psychology, 20, 475-
486.
LANZETIA, ]. T., CAR1WRIGHT-SMITH, ]., & KLECK, R. E. (1976). Effects of nonverbal dissi-
mulation on emotional experience and autonomic arousal. Journal of Personality and
Social Psychology, 33, 354-370.
LAZARUS, R. (1966). Psychological stress and the coping process. New York: McGraw-Hill.
LESHAN, L. L. (1966). An emotional life-history pattern associated with neoplastic disease.
Annals of the New York Academy of Sciences, 125, 780-793.
MELTON, A. W. (1963). Implications of short-term memory for a general theory of memory.
Journal of Verbal Learning and Verbal Behavior, 2, 1-21.
MELZACK, R. (1973). The puzzle of pain. New York: Basic.
NEISSER, U. (1976). Cognition and reality. San Francisco: Freeman, 1976.
NICHOLSON, R. A., & BERMAN,]. S. (1983). Is follow-up necessary in evaluating psycho-
therapy? Psychological Bulletin, 93, 261-278.
NoTARIUS, C. I., & LEVENSON, R. W. (1979). Expressive tendencies and physiological re-
sponse to stress. Journal of Personality and Social Psychology, 37, 1204-1210.
PAVLOV, I. (1927). Conditioned reflexes. New York: Oxford.
PENNEBAKER, J. W. (1982). The psychology of physical symptoms. New York: Springer-Verlag.
PENNEBAKER, J. W., & CHEW, C. H. (in press). Deception, electrodermal activity, and in-
hibition of behavior. Journal of Personality and Social Psychology.
PENNEBAKER, J. W., & O'HEERON, R. C. (1984). confiding in others and illness rates among
spouses of suicide and accidental death victims. Journal of Abnormal Psychology, 93,
473-476.
REscORLA, R. A. (1969). Pavlovian conditioned inhibition. Psychological Bulletin, 72, 77-94.
Ross, N., & ABRAMS, S. (1965). Fundamentals of psychoanalytic theory. In B. Wolman
(Ed.), Handbook of clinical psychology. New York: McGraw-Hill, pp. 303-339.
RuBENSTEIN, C. (1982). Wellness is all: ·A report on Psychology Today's survey of beliefs about
health. Psychology Today, 16, 28-37.
SACKHEIM, H. A., & GuR, R. C. (1978). Self-deception, self-confrontation, and conscious-
ness. In G. E. ScHWARTZ & D. SHAPIRO (Eds.), Consciousness and self-regulation (Vol. 2).
New York: Plenum Press.
ScHMALE, A. H. & IKER, H. (1971). Hopelessness as a predictor of cervical cancer. Social
Science and Medicine, 5, 95-100.
ScHWARTZ, G., DAVIDSON, R. & GoLEMAN, D. (1978). Patterning of cognitive and somatic
processes in the self-generation of anxiety: Effects of meditation versus exercise. Psy-
chosomatic Medicine, 40, 321-328.
SELYE, H. (1976). The stress of life. New York: McGraw-Hill.
SILVER, R. L., BooN, C., & STONES, M. H. (1983). Searching for meaning in misfortune:
Making sense of incest. Journal of Social Issues, 39, 81-102.
SKINNER, B. F. (1938). The behavior of organisms. New York: Appleton-Century-Crofts.
SMITH, M. L., GLASS, G. V., & MILLER, T. I. (1980). The benefits of psychotherapy. Baltimore:
Johns Hopkins University Press.
SPENCE, K. W. (1936). The nature of discrimination learning in animals. Psychological Review,
43, 427-449.
TAYLOR, J. (1953). A personality scale of manifest anxiety. Journal of Abnormal and Social
TERRACE, H. S. (1972). Conditioned inhibition. In R. A. BoAKES & M. S. HALLIDAY (Eds.),
Inhibition and learning. London: Academic Press, pp. 99-119.
TRASLER, G. (1978). Relations between psychopathy and persistent criminality: Methdol-
ogical and theoretical issues. In R. D. HARE& D. ScHALLING (Eds.), Psychopathic behavior:
Approaches to research. New York: Wiley, pp. 273-298.
VALINS, S., & NISBETI, R. E. (1972). Attribution processes in the development and treatment
of emotional disorder. In E. E. }ONES et al. (Eds.), Attribution: Perceiving the causes of
behavior. Morristown, NJ: General Learning Press, pp. 137-150.
WAID, W. M., '&ORNE, M. T. (1981). Cognitive, social and personality processes in the
physiological detection of deception. In L. BERKOWITZ (Ed.), Advances in experiment social
psychology (Vol. 14). New York: Academic Press, pp. 61-106.
WAID, W. M., & ORNE, M. T. (1982). Reduced electrodermal response to conflict, failure
to inhibit dominant behaviors, and delinquent proneness. Journal of Personality and
Social Psychology, 43, 769-774.
WEINBERGER, D. A., ScHWARTZ, G. E., & DAVIDSON, R. ]. (1979). Low-anxious, high-anxious,
and repressive coping styles: Psychometric patterns and behavioral and physiological
responses to stress. Journal of Abnormal Psychology, 88, 369-380.
WoLF, S., & GooDELL,]. (1968). Harold G. Wolff's stress and disease. Sp~gfield, IL: Thomas.
6 Mechanisms of
Biofeedback Control
On the Importance of Verbal (Conscious) Processing
J. MICHAEL LACROIX
I. INTRODUCTION
A young woman sitting alone in a small room watches a television screen

which provides her with an analog on-line display of her heart rate, and,
presumably using the information provided on the screen, attempts to raise
and lower her heart rate appropriately when cued to do so by the letter R
or the letter L shown in the upper left-hand corner of the screen.
This is the paradigmatic procedure for biofeedback training. The

typical outcome of practice with this paradigm is that after some training
the subject comes to produce larger changes of the appropriate nature
in the target physiological system in response to instructions than he
or she did at the outset of training. This acquisition in response control
is usually attributed to learning (but see Riley & Furedy, 1981, for a
contrasting view).
There are at least two general questions arising from the use of this
paradigm and the response control to which it leads that should be, and
indeed have been, of interest to psychologists. The first pertains to po-
tential applications of the biofeedback paradigm to the treatment of psy-
chosomatic disorders. The second concerns the learning mechanisms
that underlie the acquisiton of response control in the biofeedback par-
adigm, and, reciprocally, whether what takes place in the context of the
biofeedback paradigm can tell us anything about the learning process
generally. It is to this second question that this chapter is addressed. It
is important to point out, in the context of conflicting claims about the
value of biofeedback, that these two issues are orthogonal to one an-
J. MICHAEL LACROIX e Department of Psychology, Glendon College, York University, To-

ronto M4N 3M6.
137
138 J0 MICHAEL LACROIX
other: biofeedback may well be useful in the treatment of a variety of

disorders and yet prove to be theoretically uninteresting.
I will begin by providing a brief historical review of work on the
question of whether the biofeedback paradigm fosters a unique or at least
distinctive kind of learning, which I will refer to as the biofeedback
phenomenon. The outcome of that review will be essentially negative. I
will then call upon a number of cognate literatures suggesting that cer-
tain features of the biofeedback paradigm as it is used presently make
it singularly inappropriate for promoting exactly the kind of biofeedback
learning phenomenon that many had thought to be characteristic of the
paradigm. Finally, I will develop a two-process theory of learning in
biofeedback that takes into account what is going on in the paradigmatic
case and that also makes predictions as to manipulations that should
foster the distinctive biofeedback phenomenon.
II. A PROMISE UNFULFILLED
One of the attractions of the biofeedback paradigm has been the

promise that it might tell us something important about the learning
process. Indeed, historically, the interest in what was then known as
operant autonomic conditioning stemmed primarily from the fact that
researchers thought the paradigm potentially critical for two-process
learning theories (Katkin & Murray, 1968). This particular line of interest
soon faded, not because of any problems with the paradigm, but pri-
marily because of a loss of interest in two-process learning theories and
in traditional learning theories generally, which carne under heavy at-
tack from a number of directions in the late 1960s and early 1970s. With
the waning of interest in two-process theories, the bulk of research in
biofeedback began to focus on applications. However, the promise of
the biofeedback paradigm for Learning Psychology was not forgotten,
and I think it is accurate to say that most biofeedback researchers and
practitioners maintained a belief throughout this period that there was
something different about the kind of learning fostered by the paradigm,
but they were unable to specify what it was because the theoretical
structures of Learning Psychology were shifting too rapidly.
In a series of papers beginning at a conference in Chapel Hill in
1972, Jasper Brener sought to move our conceptualizations of the learn-
ing process in biofeedback away from those of Learning Psychology and
to bring them closer to those of neurophysiology (Brener, 1974a, b;
1977a, b; Brener, Ross, Baker, & Clemens, 1979). Brener also transformed
the promise of biofeedback from shimmering glimmer to testable prop-
osition. Before describing his initial conceptualization of learning in the
MECHANISMS oF BIOFEEDBACK CoNTROL 139
biofeedback paradigm (it has since changed significantly), it is important

to draw attention to a distinction which is implicit in Brener's earlier
writings and has been made explicit in his latest work (Brener, 1984),
and to which I will return at greater length later in the chapter. This is
the notion that learning can proceed in one of two basic ways. First, it
can involve simply a reshuffling of existing behavioral/cognitive ele-
ments at a central level, as when a subject learns to associate two words
already known to him or her in a paired-associates learning task, or
when a new driver learns to associate a previously known stimulus (a
red traffic light) with a variant of a behavior previously used in a number
of other contexts (lifting the right foot and moving it over slightly).
Second, learning can involve the acquisition of skills not previously
within the subject's behavioral repertoire and thus may require an ex-
tensive recalibration of sensorimotor systems in order to construct an
appropriate motor program. It is presumably this second type of learning
that operates when a person learns to speak a foreign language or a
baby learns to walk. In Brener's initial theoretical formulation, learning
in the biofeedback paradigm was considered an instance of this second
type of learning.
In this context, Brener envisaged the process of acquiring control
of the target response in the biofeedback paradigm as primarily one of
learned discrimination: the subject had to learn to discriminate and to
identify interoceptive afferentation related to the target response, in
order to exhibit control of the response. The feedback stimulus used in
biofeedback was viewed as serving simply to identify the relevant af-
ferentation. This was accomplished by a process of association or cali-
bration, through repeated pairings of the exteroceptive feedback stim-
ulus, which is by its very nature discriminable by the subject and some
feature of which is directly related to appropriate changes in the target
response, and the interoceptive afferentation which is also generated as
a consequence of changes in that response. Thus, as biofeedback training
progressed, the subject was expected to become gradually more profi-
cient at discriminating changes in the target response. The set of inter-
oceptive stimuli which the subject comes to identify with the production
of the target response was referred to in Brener's terminology as the
response image. Once the subject had formulated an appropriate response
image, activation of this response (e.g., by appropriate instructions to
produce the response) was envisaged to lead automatically to control
of the response. The theory also predicated that the response image
underwent refinement as a function of practice in the biofeedback par-
adigm. As variability occurred in the response complex produced during
repeated attempts to control the response, the exteroceptive feedback
was expected to become a reliable predictor of a progressively smaller
140 J. MICHAEL LACROIX
set of interoceptive stimuli, eventually to become a reliable predictor

only of those interoceptive stimuli arising from changes in the target
response itself-provided sufficient training was given and subject to
operating biological constraints. Thus, over the course of biofeedback
training, both control and discrimination of the target response were
expected to become gradually more specific to the target response.
In sum, Brener viewed learning in the biofeedback paradigm as
resting upon the development of new, discriminative skills, which were
necessary and sufficient for the development of an appropriate motor
program for controlling the target response. It is important to emphasize
that this view is and continues to be shared by many, if not most bio-
feedback investigators. A few examples should suffice. Thus, in an-
swering the question, What is biofeedback? Turk, Meichenbaum, and
Berman (1979) wrote: "By trial and error the subject learns to recognize
the subjective state and subtle internal changes associated with an al-
teration in physiological activity" (p. 1322; italics added). Similarly, Sur-
wit and Keefe (1983) stated that biofeedback is "based on the notion
that increasing awareness of the physiological response will lead to im-
proved voluntary control of those responses" (p. 2; italics added). Or
again, in discussing how biofeedback works in reducing frontalis EMG
and tension headaches, Cox and Hobbs (1982) suggested that it served
"to increase the patients' awareness of physiological dysregulation" (p.
387; italics in original).
Does practice in the biofeedback paradigm lead to a biofeedback phe-
nomenon, a form of learning characterized by a recalibration of senso-
rimotor systems essential to the development of an appropriate motor
program? Such a phenomenon, if we take Brener's formulations as its
testable expression, should be evidenced by the following observations.
First, subjects who have undergone training to control a given target
response in the biofeedback paradigm should perform well on tests of
their ability to discriminate changes in that response. Second, if the
presence of an appropriate response image is indeed a sufficient con-
dition for control of the response, subjects who can successfully dis-
criminate changes in a given autonomic or other "internal" response
should perform well on tests of their ability to control that response.
Third, practice in the biofeedback paradigm, if it results in a gradual
refinement of the response image, should lead to the gradual elimination
of components of the response complex that are not strongly correlated
with the target response; thus, control (and discrimination) should be-
come more specific to the target response as a function of practice in
the biofeedback paradigm.
Four years ago I carried out an exhaustive review of the literature
with respect to those predictions from Brener's theory (Lacroix, 1981).
MECHANISMS OF BIOFEEDBACK CONTROL 141
I will not duplicate that review here. Suffice it to say that the conclusion
from that review was essentially negative. Moreover, because of the
methodological diversity which characterized the literature, Anne
Gowen and I (Lacroix & Gowen, 1981) carried out an extensive exper-
iment which tested simultaneously the three major theoretical predic-
tions outlined above. These were tested in a comparative design (Rob-
erts, 1974) that contrasted the evolution of control and discrimination
with respect to two target responses, heart rate and skin conductance,
under homologous conditions. Again, the results were essentially neg-
ative. Moreover, I have seen nothing in the literature in the past four
years to convince me that the biofeedback paradigm as it is typically
used leads to the development of a biofeedback phenomenon. Training
in the biofeedback paradigm does not necessarily enhance subjects' abil-
ity to discriminate changes in the target physiological response. Per-
formance on tests of response discrimination are not usually correlated
with performance on tests of response control. And training in the bio-
feedback paradigm does not usually lead to greater response specificity
as a function of practice. There are a few notable exceptions to these
generalizations, and I will deal with them shortly.
Before dealing with these exceptions, however, and with the larger
question of the nature of the learning process in the biofeedback par-
adigm, I would like to consider two challenges to these general conclu-
sions. First, there has been mounting criticism of the various methods
used to measure subjects' ability to discriminate autonomic and other
internal responses. These have been voiced particularly strongly by Rob-
erts on theoretical grounds (Roberts, 1977; Roberts, Williams, Marlin,
Farrell, & Imiolo, 1984) and by Katkin, (Katkin, Morell, Goldband, Bern-
stein, & Wise, 1982) and Ross and Brener (1981) on empirical grounds.
The thrust of those criticisms is on the one hand that inability to perform
on a discrimination task often simply reflects inadequacies in the task.
On the other hand, and conversely, successful performance on a dis-
crimination task often reflects a lack of control over potentially con-
founded variables, suggesting that subjects can discriminate internal
responses that they cannot in fact discriminate and leading to a spurious
relationship between discrimination and control. The upshot of those
criticisms is that the first two predictions from Brener's theory (that
subjects who can control a response should also be able to discriminate
that response and that subjects who can discriminate a response should
also be able to control it) simply cannot be evaluated cleanly, leading
some investigators (Roberts & Marlin, 1979; Roberts et al.; 1984) to seek
alternative ways of assessing awareness of the response in biofeedback.
I do not think that the criticisms that have been offered are necessarily
fatal to the use of any response discrimination measure, and I think that
our own work which shows no relationship between response control

and response discrimination deals adequately with both lines of criticism
(Lacroix & Gowen, 1981). Nevertheless, even if it is granted that the
first two predictions from Brener's theory cannot be assessed unambig-
uously, the third prediction from the theory (that biofeedback training
should lead to greater response specificity as a function of practice) re-
mains unconfirmed.
The second challenge to my conclusion that use of a biofeedback
paradigm does not foster a biofeedback learning phenomenon is the
argument that training in the paradigm is seldom carried out extensively
enough to allow a biofeedback phenomenon to become manifest. Miller
(1981) made this point rather colorfully when he suggested that much
of the literature on learning in the biofeedback paradigm could be lik-
ened to a literature concerned with skating behavior in which all the
studies would focus on people putting on their skates and tying their
laces, and in which the general conclusion would be that skating is very
much like walking. The desirability of using longer training regimens
in biofeedback studies is a point well taken and has also been made by
others (e.g., Steiner & Dince, 1981). However, it surely cannot be as-
sumed that the form of the learning process will invariably be altered
by adding X number of additional sessions of training to any experi-
mental regimen, particularly as the literature varies widely with respect
to the amount of training variable. Moreover, the requisite number of
sessions for finally looking at "skating behavior" remains undefined.
These points notwithstanding, I would like to present some em-
pirical evidence relevant to the insufficient training argument. This evi-
dence comes from a recent experiment which focused on biofeedback
training of frontalis EMG in muscle-contraction headache patients (Lac-
roix, Clarke, Bock, & Doxey, 1984). Training was extensive and I pre-
sume that it was sufficient to allow a putative biofeedback phenomenon
to become apparent, as the patients obtained significant clinical benefits
from the training, and these were maintained for at least six months
after training. Figure 1 presents the response profiles obtained in these
patients. The figure depicts the mean resting levels in four physiological
functions during two sessions prior to any training (left-hand side) and
two post-training baseline sessions (right-hand side). Training was pro-
vided over sixteen sessions, each of which included some programmed
generalization (i.e., some practice without feedback or relaxation tapes,
depending on the groups). It should be emphasized that the patients
were instructed to practice the skills developed over the course of train-
ing during the final baseline sessions. Subjects in the biofeedback group
received training to lower frontalis EMG (EMG 2 in the figure), subjects
in the relaxation group received a fixed series of relaxation tapes, sub-
EMGZ EMG I PR T0
90
18 95
BIOFEEDBACK GROUP
6
15 e5 94
12 4 93
eo
9 3 92
6 2 91
75
3 90
0 0 70 BAS . 3+4
BAS . I +2
90
16 6 95 BIOFEEDBACK AND RELAXATION
15 65 94
12 4 93
9
eo 92
6 91
75
3 90
0 0 70
BAS. I+Z
90
18 95 RELAXATION GROUP
15 65
12 4 93
eo
9 92
3 90
0 0 70
BAS. 1+2 8AS . 3+4
90 CONTROL GROUP
18 95
85 94
12 4 93
80
9 92
6 2 91
75
90
0 0 70
BAS . I+ 2 BAS. 3+4
FIGURE 1. Mean resting levels in frontalis EMG (EMG-2), longissimus or posterior neck
EMG (EMG-1), pulse rate (PR) and peripheral digit temperature (T" in degrees Fahrenheit)
during two baseline sessions prior to, and two baseline sessions following an experimental
treatment. The treatments were, respectively, biofeedback for decreases in EMG-2, bio-
feedback for EMG-2 and relaxation training, relaxation training only, and no treatment.
The subjects were all muscle-contraction headache patients. (From Lacroix eta/., 1984).
jects in the biofeedback and relaxation group received both frontalis

EMG training and the relaxation tapes, and subjects in the control group
received no training of any kind between the first two and the last two
baseline sessions. The figure depicts, successively, the target EMG ac-
tivity, EMG recorded from a different site (back or neck), pulse rate, and
hand temperature.
What I find most interesting about these data is that the response
profiles are so similar. Thus, a comparison of the profiles in the bio-
feedback and the relaxation groups reveals in both cases decreases in
the two EMG measures and negligible increases in pulse rate and pe-
ripheral temperature. Indeed, analyses of variance carried out on these
data yielded only two significant effects (a Time main effect for EMG 2,
F(l/51 = 16.56, p < .001, and a Groups main effect for pulse rate, F(3/
51) = 2.9, p < .05): none of the Groups x Time interactions approached
statistical significance. It might be argued that differences in level of
learning in different subjects might add so much noise to the data as to
mask important differences between groups with respect to response
specificity. We evaluated this possibility by carrying out an identical
analysis, but using only the data from subjects who showed significant
learning from the first two to the last two baseline sessions, as defined
by the absence of overlap between the standard errors of the mean. The
pattern of data was essentially the same: no group differences emerged
with respect to response profiles. The message from these data, clearly,
is that longer training regimens do not necessarily reveal a biofeedback
phenomenon.
If it is not a biofeedback phenomenon that accounts for learning in
the biofeedback paradigm, what does? I suggested earlier that although
learning could entail an extensive recalibration of sensorimotor systems,
it could also entail a simpler reshuffling of behavioral-cognitive elements
at the central level. I have argued before (Lacroix, 1981) that this latter
process is in fact the sort of process that is going on (usually) when
subjects undergo training in the biofeedback paradigm. Briefly, what I
think happens in the usual case is that the instructions given to the
subject identify a set of existing behaviors within his or her behavioral
repertoire. Insofar as this set is at all effective in producing appropriate
changes in the feedback display, this set is then simply magnified over
the course of training. This position will be developed at greater length
later in the chapter, but to keep the skeptical reader at least open to the
possibility I would like to introduce here one small bit of empirical
evidence.
Figure 2 (from Lacroix, Clarke, Bock, Doxey, Wood, and Lavis, 1983)
presents within-session changes in peripheral hand temperature in sub-
jects suffering from migraine headaches trained to control peripheral
MECHANISMS OF BIOFEEDBACK CoNTROL 145
-------- ------
5
,.
I
"'c
"'
0 4
.<:
(.)
·-
c
0 "'c
.s::;
"iii ~
Vl
"'"' .s::;
~ 3
:c
c
"'"'
~
it
.,
c>
0
c 2
., 0
:::;:
- .... -Thermal Biofeedback, Training
-Thermal Biafeedback,Test
o E MG Biofeedback, Test
" Relo•ation Training, Test
182 384 58 6 Test Sessions

1-3 4-7 8-10 Training Days
FIGURE 2. Mean within-session change in skin temperature in three groups of migraine

headache patients who received biofeedback for increases in peripheral digit temperature,
biofeedback for decreases in frontalis EMG, or relaxation training. Performance is displayed
for the three groups on pairs of test (no feedback) sessions given before training, half-way
through, and at the end of training, and in the thermal biofeedback group on training
(feedback) sessions from the first, middle, and last third of training. (From Lacroix et al.,
1983).
temperature or frontalis EMG or trained in progressive relaxation. The

figure presents the hand temperature data from test (no feedback) ses-
sions before training, half-way through an 18-session training regimen,
and after the 18 training sessions, in the three groups, and also presents
the comparable data from training (feedback) sessions in the early, mid-
dle, and late stages of training, in the skin temperature feedback group.
What is particularly interesting here is that the temperature biofeedback
group was able to produce substantial increases in skin temperature on
training days from the outset of training: biofeedback training appar-
ently resulted simply in generalization of this performance to no-feed-
back test conditions. Clearly, the temperature control task (which was
accompanied by pronounced changes in migraine attacks in these pa-
tients, who had suffered from migraines for 20 years on average) was
well within the subjects' behavioral repertoire.
146 J0 MICHAEL LACROIX
Is the implication to be taken from these data and arguments that

the biofeedback phenomenon is simply not to be found in the biofeed-
back paradigm, and that the paradigm is therefore of little interest to
psychologists whose primary concern is with the learning process? Not
quite. Although the biofeedback phenomenon appears to be rather like
the Cheshire cat, there are some tantalizing suggestions of its presence
in the literature. Two lines of evidence appear particularly promising.
The first stems from a number of studies carried out before the bio-
feedback paradigm acquired the wide currency that it now holds, when
scientists interested in the general area still referred to it as "operant
autonomic conditioning." Although there are a number of procedural
differences between biofeedback studies and operant autonomic con-
ditioning studies, one that strikes me as particularly important pertains
to the information given to the subjects. In the present biofeedback par-
adigm subjects are invariably given complete information about the re-
sponse to be controlled and the direction (and sometimes the magnitude)
of desired changes and are often (particularly in clinical studies) pro-
vided with specific suggestions as to response strategies that might be
effective in controlling the target response. In contrast, in the earlier
operant conditioning paradigm subjects were invariably kept ignorant
of the response to be controlled, the contingency between the response
and the reinforcer that was used, and often even of the very existence
of any contingency on their behavior; in many studies, moreover, sub-
jects who guessed correctly either the target response or the nature of
the contingency were dropped from the analyses.
Yet it is precisely from those kinds of studies that the most en-
couraging evidence emerges for gradual acquisition functions (e.g. for
heart rate: Ascough & Sipprelle, 1968; Levene, Engel, & Pearson, 1968;
Shapiro, Tursky, & Schwartz, 1970; for electrodermal activity: Greene
& Wirth, 1974; Kimmel & Baxter, 1964; Van Twyver & Kimmel, 1966;
for blood pressure: Shapiro, Schwartz, & Tursky, 1972; Shapiro, Tursky
& Schwartz, 1970; Shapiro, Tursky, Gershon, & Stern, 1969). Moreover,
the most encouraging evidence for specificity of control also comes from
some of these studies. For example, Ascough and Sipprelle (1968) re-
ported an independence of target heart-rate changes from respiratory
changes, and Cohen (1973) reported an independence between target
heart-rate changes and EMG changes. Yet in the biofeedback paradigm
our fine-grained analyses of the development of cardiac control indicate
that instructions to control heart rate activate respiratory changes and
that the acquisition of heart-rate control then takes place in the context
of somatomotor activation (Kelly & Lacroix, 1983; Lacroix & Gowen,
1981), and this is in keeping with more molar analyses carried in other
investigations (Clemens & Shattock, 1979; Lacroix & Roberts, 1978; Lev-
MECHANISMS OF BIOF!lEDBACK CoNTROL 147
enson & Ditto, 1981; Obrist et al., 1975; Schober & Lacroix, 1985). Sim-
ilarly, Van Twyver and Kimmel (1966) reported no relationship between
target GSR changes and changes in forearm EMG or respiration fre-
quency, whereas in the biofeedback paradigm we have invariably found
target electrodermal changes to be embedded in a somatomotor-respi-
ratory complex (Lacroix & Gowen, 1981; Lacroix & Roberts, 1978).
Not all operant autonomic conditioning experiments have yielded
gradual acquisition functions or specificity of control, of course, and
even those studies that have provided such evidence of a biofeedback
learning phenomenon are far from entirely convincing. My point here
is simply to draw attention to these studies, because they at least provide
suggestive evidence of a learning process akin to the biofeedback phe-
nomenon, whereas studies carried out in the biofeedback paradigm sim-
ply do not. One final point about the specificity of control which is
(sometimes) evident in these earlier investigations: it is interesting to
note that the specificity which is apparent in these studies does not
develop in the sense that components of the response complex initially
activated by the subject are eliminated as a function of training. Rather,
control appears specific to the target response from the outset of training
(e.g. Schwartz, 1972; Shapiro, Tursky, & Schwartz, 1970).
The operant autonomic conditioning studies constitute the first line
of evidence which suggests that the biofeedback phenomenon might
provide a description of the learning process in some cases. The second
line of evidence stems from studies carried out within the current bio-
feedback paradigm, but with patients suffering from various physio-
logical dysfunctions. Evidence of gradual acquisition functions and of
control specific to the target response has been provided by Engel and
Bleecker (1974), for example, on the control of cardiac arrhythmias and
by Miller and Brucker (1979) on the control of diastolic blood pressure
in a hypotensive patient. I would like to indicate here that the variable
that distinguishes these studies from other studies with the biofeedback
paradigm that have not yielded any evidence of a biofeedback phenom-
enon is not simply the clinical status of the subjects, since our work with
clinical populations, referred to earlier, has not not provided any evi-
dence of learning in terms of a biofeedback phenomenon in these pa-
tients. Presumably it is the specific nature of the disorder and not simply
the presence of a psychophysiological disorder which is critical. I will
return to this point and these studies in the last section of the chapter.
III. THE VERBAL (CoNSCIOUSNESS?) INTERFACE
In the last couple of years I have become intrigued with the impli-
cations of the differences in outcome pertaining to the biofeedback phe-
nomenon between studies carried out with the biofeedback paradigm

and (some) studies carried out in the paradigmatic context of operant
autonomic conditioning. Could the mere presence of a verbal label des-
ignating a target response have such dramatic effects as to affect the
very form of the learning process? In this section I would like to focus
on this general question, by first drawing upon a number of literatures
cognate to the biofeedback literature that point to the importance of
verbal labels and of verbal processing generally. I will then offer some
speculations on the role of the verbal labels in the biofeedback paradigm,
and these speculations will then be formalized in the general theory of
learning in the biofeedback paradigm which is outlined in the following
section.
First, it should be acknowledged that the putative importance of
the verbal information given to the subjects in the context of biofeed-
back-operant conditioning studies has not been completely ignored in
the literature. There has been some attention to the effects of providing
the target response with different labels. Thus, Blanchard, Scott, Young,
and Edmundson (1974) observed larger heart-rate changes when heart
rate was designated as target and subjects were given feedback for heart
rate than when the target response was designated as skin conductance
or as an unspecified internal response and feedback was nonetheless
provided for heart-rate changes. Quy and Kubiak (1974) examined the
issue of "awareness" of the response. Two groups of subjects were
punished by an electrical pulse for spontaneous electrodermal re-
sponses, but one group was given information about the target response
and about the response-reinforcer contingency whereas subjects in the
other group were told simply that the study focused on their physio-
logical reactions to the electrical stimuli. Although a careful look at their
study reveals statistical problems that make the results uninterpretable,
their conclusions are worth noting:
A subject's awareness of the process of control could be a condition which
impedes, rather the encourages learning. This implies that the straightfor-
ward instrumental learning procedures with "naive" subjects may in fact
produce more rapid and enduring control over autonomic functions than
"aware" conditions of learning. (p. 505)
There has also been some interest in the effects of providing subjects
with suggestions as to response strategies for controlling the target re-
sponse. In our own work on this issue (Lacroix & Roberts, 1978), we
have found the provision of such strategies to be deleterious to control
of both electrodermal and cardiac target responses, and others have
reported similar evidence (e.g. White, Holmes, & Bennett, 1977). Al-
though the above studies are indicative of a peripheral interest in the
biofeedback literature with issues related to the importance of verbal
labels and verbal processing, it is clear that their concern was with the
effects of different verbal information on level of control of the target
response, that is, on performance. My concern here goes deeper, how-
ever: can such information affect not only performance level but also
the form of the learning process?
It is interesting to note that the effect of providing a subject with
the information as to the response to be controlled and as to the re-
sponse-feedback contingency is to make the subject aware of that re-
sponse, or to make the response conscious. The last several years have
seen increasing attention to the question of whether being aware of
information is necessary in order to act on that information, and to the
meaning and the nature of consciousness. I do not plan to review here
what is now a voluminous literature with ramifications in all areas of
psychology. I would like simply to highlight some aspects of that lit-
erature which I think can shed some light on the issue of learning in
the biofeedback paradigm.
First, it is clear that the brain comprises a number of different in-
formation-processing systems. Moreover, it appears that information
can be processed independently by the different systems (or, at least, some
of the information, some of the time, with respect to some of the in-
formation-processing systems). Evidence for this generalization can be
adduced from cognitive (Nisbett & Wilson, 1977), clinical (Beahrs, 1982;
Hilgard, 1976), and neurophysiological data (LeDoux, Wilson, & Gaz-
zaniga, 1979) and is sensible on phenomenological grounds (I can drive
my car without running into trees while my conscious attention is fo-
cused on the news being read on the radio). Thus for example, in a
series of clever experiments, Nisbett and Wilson manipulated appar-
ently innocuous variables that had significant effects on subjects' be-
havior. Yet the subjects attributed their behavior to variables that were
controlled by the experimenters and therefore could not have been the
basis for the behavior. Nisbett and Wilson concluded that subjects do
not have access to their cognitive processes. Pennebaker (1982) recently
described the same sort of phenomenon with respect to subjects' per-
ception of physical symptoms. Following certain manipulations subjects
would report certain symptoms consciously, and would invariably at-
tribute these to internal factors rather than to the experimenter-con-
trolled variables.
Yet another and dramatic example of appropriate behavior being
produced in the absence of conscious realization is given by the results
of a study by Risse and Gazzaniga (cited in LeDoux et al., 1979). In this
study, patients scheduled to undergo brain surgery had their left hem-
isphere anaesthetized by sodium amytal and had objects placed in their
left hand (which is governed by the right hemisphere). When the left
hemisphere had recovered from the anaesthetic, the patients were un-
able to name the objects but could pick them out correctly, by pointing,
from among a larger group of objects, all along insisting that they were
only guessing.
This last experiment is important in that it not only reinforces the
argument that learning can proceed in information-processing systems
other than the verbal system without the knowledge of the latter, it also
develops this argument one small step further by pointing out that the
different information-processing systems do not necessarily have access
to each other's memories. If we agree, then, that behavior is governed
by a number of information-processing systems that overlap only par-
tially, how does one come to present a coherent self to the world, and
for that matter to oneself: what is consciousness within this apparently
anarchic universe within the brain/mind? LeDoux et al. (1979) have re-
cently provided an elegantly simple answer to these questions.
LeDoux et al. argue that, given the importance of language, over
the course of ontogeny the activities in the various information-pro-
cessing systems come to be monitored more and more by the infor-
mation-processing system on which we come to rely more and more:
the verbal system. Gradually a concept of self develops which is equiv-
alent to what the verbal system knows about itself and the others. In
their view, then, consciousness is equivalent to the activation of the verbal
information-processing system. This conceptualization gives the verbal sys-
tem a triple importance. It is important in its own right, in the processing
of verbal information. It is important as an interface between the other
information-processing systems in the brain/mind through its monitor-
ing activities. And it is important as a major interface between the person
and the outside world. Thus, most demands on the individual coming
from the outside world, because of their verbal nature, are first "han-
dled" by the, verbal system. It is through the verbal system interface
that the other information-processing systems learn of these demands.
The demands of the outside world are then processed by the verbal
system and presumably by others, but since it is again the verbal system
that interprets the outputs of the other systems to the outside world,
the verbal system will also put its "stamp" on behavior generated by
the other systems in the brain/mind.
That the verbal system interprets in its own "language" the outputs
of the other information-processing systems was demonstrated dra-
matically in a series of split-brain studies by Gazzaniga and LeDoux
(1978). Thus, when the patient's left and right hemispheres were pre-
sented simultaneously with different demands, leading to behavior
which could be predicted only in part by the left (verbal) hemisphere,
the verbal hemisphere nonetheless provided an instantaneous inter-
pretation of the patient's behavior in terms of its own referents. For

example when the patient's right hemisphere was instructed to scratch,
and did so, the left hemisphere simply indicated that he was itchy.
The domination of the various information-processing systems by
the verbal system is never complete, as any good clinician can attest.
However, it is usually sufficient to provide enough of a semblance of
unity that we can appear as a coherent person both to the outside world
and to the verbal-conscious self. However, the partial nature of verbal
predominance is revealed easily under a variety of conditions that favor
the temporary ascendancy of other systems, such as hypnosis or dis-
sociative disorders, and in patients with a severed corpus callosum or
an anesthetized left hemisphere. The partial nature of verbal predom-
inance is also evident in more normal conditions, as when we can pro-
cess complex information (e.g., driving) while consciously-verbally en-
gaged in some other task, or when nonverbal information-processing
systems find nonverbal (and sometimes verbal) means of expression, as
in the patient with a conversion reaction who chuckles while describing
her symptomatology (or in the case of Freudian slips).
What are the implications of this conceptualization for the learning
process? At a general level, the major implication is that we can expect
that demands placed on the individual by the outside world (including
demands to learn some new behavior) will be conveyed to the network
of information-processing systems in the brain/mind by the verbal sys-
tem and will be processed by that system unless conditions make such
processing impossible. Moreover, any processing of information by
other systems will be interpreted by the verbal system and will generate
some sort of verbal coding. In terms of the general distinction made
earlier, learning in the typical case will involve a reshuffling of behav-
ioral-cognitive elements that are verbally coded. However, there are two
sorts of conditions under which we may expect a demand from the
outside world not to involve the verbal system. The first is in the case
of highly predictable, well-rehearsed tasks that have been repeated so
frequently that their processing has become automatized. Cognitive psy-
chologists have been understandably concerned with the differences
between highly automatized tasks and tasks requiring conscious-verbal
processing and have drawn distinctions between these on a number of
bases (e.g., Hasher & Zacks, 1979; Schneider & Shiffrin, 1977). The sec-
ond category of conditions under which the verbal system might be by-
passed comprises those conditions wherein the verbal system is unable
to deal with the demands of the task. This might occur in cases in which
the verbal system lacks the "language" or information or the proper
interfacing with other systems to deal with the demands. Alternatively,
this might occur in cases in which the task demands are not phrased in
such a way as to allow the verbal system to access the relevant infor-
mation that it has. These conditions are most likely to be met when the
task to be learned is some sort of "skill." In terms of the general dis-
tinction made earlier, learning under those conditions will involve an
extensive recalibration of sensorimotor systems in order to construct an
appropriate motor program.
We can now see the relevance of this conceptualization to learning
in the biofeedback paradigm. In this context, we might profitably en-
visage the verbal system in terms of an analogy, as a central library
through which information pertaining to behavior is normally accessed.
Demands are placed on a subject, for example, to raise his heart rate.
Unless the behavioral strategy for producing increases in heart rate has
become automatized, the subject has to "look up" what the task entails.
If relevant information is found in the verbal library, including appro-
priate interfacing with the relevant behavior-production systems, and
if implementation of the behavior program thus identified is successful
(the feedback arrow goes up), there is no incentive to continue the search
outside the verbal library system. Indeed, there is an extensive literature
that shows that organisms do not work for redundant information (e.g.,
Egger & Miller, 1963; Kamin, 1968). The fact that the changes produced
in this way may not be maximal is seldom, if ever, indicated to the
subject. Nor is the subject typically instructed to produce changes spe-
cific to the target response. Thus, it is only in those instances wherein
the verbal library cannot access information relevant to the appropriate
behavioral programs (the operant autonomic conditioning studies?) or
wherein the verbal system lacks the language or the interfacing to inform
the subject as to what to do (the clinical cases such as those of Miller &
Brucker, 1979) that learning could be expected to proceed along the lines
of a biofeedback phenomenon.
IV. A Two-PRocEss THEORY

In this section I would like to develop a theory of how human sub-
jects learn to control autonomic or other "internal" responses when
provided with information about the target response in the form of a
biofeedback signal in the biofeedback paradigm, or a reinforcer for ap-
propriate responding in an operant autonomic conditioning paradigm.
In keeping with the thrust of the previous section, the theory places
considerable emphasis on the role of the verbal system both as the in-
terface with the outside world that sets the task requirements and as
the interface with the other information-processing/behavior-generating
systems of the brain/mind. The theory presented here is basically an
MECHANISMS OF BIOFEEDBACK CONlROL 153
elaboration of the theory that I introduced in outline form in 1981 (Lac-

roix, 1981). However, it also echoes suggestions made by a number of
other biofeedback investigators. Thus, the emphasis that I will place on
what are called here feedforward processes parallels a similar emphasis
in .the recent work of Brener (1982), Qualls and Sheehan (1981) and
Stilson, Matus, and Ball (1980). Moreover, the relatively large impor-
tance given to the task instructions owes much to Roberts's conceptual
analysis of the environment in which biofeedback training takes place
(Roberts & Marlin, 1979; Roberts et al., 1984), although my emphasis
here is on how the task demands are represented in the subject's verbal
system and on the verbally mediated associations thus generated.
The theory is presented in the form of a flow-diagram in Figure 3.
The diagram is not meant to be exhaustive of the operations involved,
but for the sake of clarity the more obvious operations have been as-
sumed (e.g., monitoring the effect of "trying harder" on the feedback
display-upper right-hand quadrant). The theory envisages that control
of autonomic responses may be acquired through two different kinds
of processes. These are labeled feedforward and feedback processes (above
and below the dotted line) in keeping with Brener's recent analysis of
the neurophysiology of learning of skeletal-motor responses (Brener,
1984). Feedforward and feedback processes correspond to what I had
earlier (Lacroix, 1981) referred to as efferent and afferent processes, re-
spectively; the earlier terminology generated some confusion and I am
happy to homogenize my nomenclature with that of Brener. Feedfor-
ward and feedback processes also correspond to the two general strate-
gies in which learning can proceed and to which I made reference earlier:
a central reshuffling of existing behavioral-cognitive elements and a re-
calibration of sensorimotor systems. One advantage of the present con-
ceptualization is that it provides for possible interactions between these
two general strategies, that is, for subjects to change from one strategy
to the other when their initial selection is no longer effective or optimal.
Briefly, control of autonomic responses is envisaged to be acquired
as follows. The instructions given to the subject lead him to engage in
a search of his verbal library to determine whether it contains behavioral
programs which are potentially relevant for control of the target re-
sponse and are appropriately interfaced with pertinent nonverbal in-
formation-processing/behavior-generating systems. The behavioral pro-
gram judged likely to be the most effective is then implemented and its
effectiveness is evaluated in terms of whether the feedback display
changes as expected. If it does, the subject may decide to try to improve
on his performance either by carrying out the various components of
the selected behavioral program with greater intensity, or by directing
his attention to constituent elements of the programs which are acces-
......
~
TASK
INSTRUCTIONS
I
N
T
E
R
F
A { FEEDFORWARD
c PROCESSES]
E
(FEEDBACK
PROCESSES]
':""'
~
~
r
t""'
FIGURE 3. Schematic representation of the processes and operations involved in learning to control an "internal" response in biofeedback-like ~
situations. ~
sible to (verbal) consciousness and, in as much as possible, identify a

more effective subset. This may continue until the subject judges further
improvements in performance to be unlikely or unworthy of the effort,
and performance should remain at asymptote thereafter. If, however,
the behavioral program initially activated is not effective in producing
the desired change in the feedback display, the subject implements the
next program in his hierarchy of potentially effective programs, more
or less in keeping with Hull's notions of habit family hierarchies (Hull,
1952). The subject may thus implement a number of behavioral pro-
grams, successively, either until one is found to be effective or until all
the programs identified from the verbal library have been exhausted
and shown to be ineffective. At this point the subject's general strategy
would switch from one that emphasizes feedforward processes to one
that emphasizes feedback processes.
In sum, in most instances of autonomic learning the subject is be-
lieved to adopt a general learning strategy that emphasizes feedforward
processes: he or she selects a previously constructed motor program and
simply "plugs it in," with relatively little attention to feedback infor-
mation from interoceptive afferentation related to the target effector.
Instead, the "feedback" provided by the feedback display serves pri-
marily to determine the appropriateness of the program which has been
selected. Improvements in performance stem either from doing more of
the same or from stressing components of the program that may be
particularly effective. Even in the latter case, however, acquisition would
not be expected to result in a gradual refinement in performance, as
would be evidenced by the gradual elimination of responses not cor-
related with the target response. Instead, acquisition is conceptualized
in terms of an additive modet with the smaller components of the be-
havioral program which have been identified by feedback training as
particularly effective being superimposed on the other components, in-
sofar as it is still the same program (albeit with minor modifications)
which is activated.
Although feedforward processes would be expected to be employed
in most instances of autonomic learning, they would not always be se-
lected. There are a number of conditions that should lead subjects to
emphasize feedback processes instead in learning to control the target
response. First, all the programs identified from the verbal library may
be ineffective. Second, the verbal library may fail to identify potentially
relevant programs. Third, the library may identify relevant programs
but these may be poorly interfaced with the nonverbal information-pro-
cessing/behavior-generating systems such that the programs are judged
to be impossible to carry out. In all these cases the subject would adopt
a general learning strategy focusing on feedback processes; learning

would then proceed along the lines of a biofeedback phenomenon.
For example, in the typical operant autonomic conditioning para-
digm the subject would be unable to plug in an already existing centrally
organized motor program to control the target response since the in-
structions may not provide enough information to implement a suc-
cessful search; the subject would therefore have to go about constructing
an appropriate program. Initially, this would involve monitoring of in-
teroceptive afferentation from potentially relevant organ systems, with
a view to determining which of the afferent processes monitored are
correlated with variations in the feedback display. If a correlation can
be established, the subject may then guess as to the nature of the target
response and implement a new verbal library search on the basis of that
new information, and this may then lead her to change to a strategy
emphasizing feedforward processes (recall that I suggested earlier that
subjects do not work for redundant information). However, should the
target response not be recognizable or not readily map onto existing
behavioral programs, acquisition of the response would then proceed
in terms of the kinds of operations that Brener originally suggested
(1974a) and that I summarized at the beginning of section II (construction
of a response image; gradual development of specificity of control to the
target response). In keeping with the emphasis that I have placed on
the verbal system as the interface with the other information-processing
systems, I surmise that successful construction of a new motor program
on this basis would then result in some updating of the verbal library.
The theoretical framework presented here is empirically testable.
Indeed, there are existing data that can be brought to bear on the theory
and these have so far fitted rather nicely.
A first critical point at which the form of the acquisition process can
be influenced is at the level of the task instructions. Presumably the
more complete the verbal instructions the greater the likelihood that
they will map onto existing behavioral programs, and the smaller the
likelihood that acquisition will proceed in keeping with a biofeedback
phenomenon (as operationalized by an ability to discriminate afferen-
tation related to the occurrence of the target response, or by a developing
specificity of control). In keeping with this analysis, I referred earlier to
a report by Blanchard et al. that subjects performed differently in a heart-
rate control task as a function of whether the target response was defined
as heart rate, skin conductance, or an unspecified internal response.
This is evidence of the importance of the verbal labels in determining
performance, presumably as a function of the behavioral programs
which the labels elicit from the verbal library. Also in keeping with this
analysis are our findings (Lacroix & Roberts, 1978) that providing sub-
jects with strategy suggestions led to a deterioration of performance on

both heart-rate and skin conductance control tasks: again, this points
to the importance of the verbal information provided to the subjects
in determining which behavioral programs are selected from the ver-
bal library. More generally, I think that the differences discussed
earlier between the data obtained in the biofeedback paradigm and
those obtained (sometimes) in the operant autonomic conditioning para-
digm also are attributable to differences in the outcome of the verbal
library search stemming from differences in the instructions given the
subjects.
A second point at which the form of the acquisition process can be
influenced is at the level of the verbal library search itself. More spe-
cifically, the outcome of that search is likely to be influenced by a number
of variables in addition to the instructions. Among these variables are
individual difference factors: some subjects may be likely to engage in
more extensive library searches or may possess "better" libraries-per-
haps in terms of their verbal cross-references. In keeping with this view,
we have found (Offutt & Lacroix, 1983) as have others (Qualls & Shee-
han, 1981) that subjects who score high on the Absorption scale exhibit
very good control of their internal responses without need for feedback
training and indeed are not helped by biofeedback training, whereas
low absorbers exhibit little control of the target internal response without
feedback and improve with biofeedback training. Absorption is a trait
which correlates with hypnotic susceptibility and reflects subjects' ability
to tune out the world and draw on their own inner resources (Tellegen
& Atkinson, 1974). The clinical data to which I referred earlier (Engel &
Bleecker, 1974; Miller & Brucker, 1979) point to another variable likely
to affect the outcome of the verbal library search operation. The patients
in these studies may represent cases in which, although potentially ef-
fective behavioral programs may be referenced in the verbal library, the
nature of the patient's injury may have affected the interface between
the verbal library and other nonverbal information-processing/behavior-
generating programs such that the patient cannot implement the pro-
gram selected.
A third point at which the present theoretical framework lends itself
to empirical test is at the level of the feedback provided to the subject
as to the effectiveness of the initial (and subsequent) behavioral pro-
grams selected. Appropriate manipulations of the feedback display
should lead subjects either to maintain their initial program or to aban-
don it in favor of an alternative. Changes (or retention) of the behavioral
programs selected should be evident both in terms of patterns of phys-
iological activity during attempts to control the target response and in
terms of the subjects' verbal reports of what they are doing.
Manipulations of the feedback display also allow a test of the theory

to be carried out at another point. If a subject is following a learning
strategy that emphasizes feedforward processes, once the subject has
decided to stay with a particular behavioral programme, her next task
is to determine whether further improvements in performance can be
accomplished. At this point, manipulations of the feedback display that
suggest that the subject is doing very well should lead to asymptotic
performance early, whereas manipulations of the display that suggest
that there is considerable room for improvement should lead the subject
to "try harder." In keeping with this analysis, Williamson, Monguillot,
Hutchinson, Jarrell, and Blouin (1981) reported that subjects produced
larger heart-rate changes in a heart-rate control task with low sensitivity
feedback displays than with high sensitivity displays. It is worth noting
here that these results are opposite to predictions from traditional motor
skills theories, although they fit nicely with the present theory.
One final testable proposition from the theory may be mentioned.
Assume that a subject follows a learning strategy that emphasizes feed-
back processes and proceeds to recalibrate his sensorimotor systems so
as to construct a new motor program, that is, the subject learns in terms
of a biofeedback phenomenon. At some point the verbal (consciousness)
system, which monitors the other information-processing systems,
should seek to code the new behavior program in terms of its own
referents. This is exactly what Roberts et al. (1984) have recently re-
ported. Their subjects were trained either to control heart rate or to
produce lateralized changes in skin conductance, but possible clues as
to the nature of the task were specifically concealed. In terms of the
present conceptualization, this should have fostered a learning strategy
that emphasizes feedback processes. Under these conditions, the sub-
jects who learned to control the target response (but not those who failed
to learn) were able to provide a verbal description of what they were
doing which, although not always well articulated, nonetheless mapped
onto some of the properties of the target response system.
In closing, I would like simply to return to the general issue raised
in the introduction. Does the biofeedback paradigm foster a unique kind
of learning, a biofeedback phenomenon such that it should be of particular
interest to psychologists interested in the learning process? My attempt
to answer this question has led me far afield and has led me to engage
in considerable theoretical speculation. At this point my answer is that
biofeedback training can lead to a biofeedback phenomenon, but that
this kind of learning is evident only under certain (now unusual) con-
ditions. The theory presented in this section attempts to specify what
those conditions are and also to provide a basis for understanding what
and how subjects learn in biofeedback situations more generally.
REFERENCES
AscouGH, ]. C., & SIPPRELLE, C. N. (1968). Operant verbal conditioning of autonomic re-
sponses. Behavior Research and Therapy, 6, 363-370.
BEAHRS, ]. 0. (1982). Unity and multiplicity: Multilevel consciousness of self in hypnosis, psy-
chiatric disorder and mental health. New York: Brunner/Mazel.
BLANCHARD, E., Scorr, R. E., YouNG, L. D., & EDMUNDSON, E. D. (1974). Effect of knowl-
edge of response on the self-control of heart rate. Psychophysiology, 11, 251-264.
BRENER, ]. A. (1974a). A general model of voluntary control applied to the phenomena of
learned cardiovascular change. In P. A. 0BRJST A. H. BLACK, J. BRENER, & L. V. DICARA
(Eds.), Cardiovascular psychophysiology: Current issues in response mechanisms, biofeedback,
and methodology. Chicago: Aldine.
BRENER, J. (1974b). Factors influencing the specificity of learned cardiovascular change. In
L. V. DICARA (Ed.), Limbic and autonomic nervous systems research. New York: Plenum
Press.
BRENER, J. (1977a). Sensory and perceptual determinants of voluntary visceral control. In
G. E. ScHwARTZ & J. BEAITY (Eds. ), Biofeedback: Theory and research. New York: Academic
Press.
BRENER, J. (1977b). Visceral perception. In]. BEAITY & H. LEGEWIE (Eds.), Biofeedback and
behavior. New York: Plenum Press.
BRENER, J. (1982). Psychobiological mechanisms in biofeedback. In L. WHITE & B. TURSKY
(Eds.), Clinical biofeedback: Efficacy and mechanisms. New York: Guilford.
BRENER, J. (1984). Operant reinforcement, feedback and the efficiency of learned motor
control. In M. CoLES, E. DoNCHIN, & S. PoRGES (Eds.), Psychophysiology: Systems, pro-
cesses and applications (Vol. 2). New York: Guilford.
BRENER, J., Ross, A., BAKER, ]., & CLEMENS, W. J. (1979). On the relationship between
cardiac discrimination and control. InN. BIRBAUMER& H. D. KiMMEL (Eds.), Biofeedback
and self-regulation. Hillsdale, NJ: Erlbaum.
CLEMENS, W. J., & SHAITOCK, R. J. (1979). Voluntary heart rate control during static mus-
cular effort. Psychophysiology, 16, 327-332.
COHEN, M. J. (1973). The relation between heart rate and electromyographic activity in a
discriminated escape-avoidance paradigm. Psychophysiology, 10, 8-20.
Cox, D.]., & Hosas, W. (1982). Biofeedback as a treatment for tension headaches. In L.
WHITE & B. TuRSKY (Eds.), Clinical biofeedback: Efficacy and mechanisms. New York:
Guilford.
EGGER, M. D., & MiLLER, N. E. (1963). When is a reward reinforcing?: An experimental
study of the information hypothesis. Journal of Comparative and Physiological Psychology,
56, 132-137.
ENGEL, B. T., & BLEECKER, E. R. (1974). Application of operant conditioning techniques to
the control of cardiac arrhythmias. In P. A. OaRJST, A. H. BLACK, J. M. BRENER, & L.
V. DICARA (Eds.), Cardiovascular psychophysiology: Current issues in response mechanisms,
biofeedback, and methodology. Chicago: Aldine.
GAZZANIGA, M., & LEDoux, J. P. (1978). The integrated mind. New York: Plenum Press.
GREENE, W. A., & WIRTH, H. G. (1974). Operant conditioning of the skin resistance response
with different intensities of light flashes. Bulletin of the Psychonomic Society, 4, 177-179.
HAsHER, L., & ZACKS, R. T. (1979). Automatic and effortful processes in memory. Journal
of Experimental Psychology: General, 108, 356-388.
HILGARD, E. R. (1976). Neodissociation theory of multiple cognitive control systems. In G.
E. ScHWARTZ & D. SHAPIRO (Eds.), Consciousness and self-regulation: Advances in research
(Vol. 1). New York: Plenum Press.
HULL, C. L. (1952). A behavior system. New York: Wiley.
KAMIN, L. J. (1968). "Attention-like" processes in classical conditioning. In M. R. JoNEs

(Ed.) Miami Symposium on the Prediction of Behavior, 1967: Aversive Stimulation. Coral
Gables, Fla: University of Florida Press.
I<ATKIN, E. S., & MuRRAY, E. N. (1968). Instrumental conditioning of autonomically me-
diated behavior: Theoretical and methodological issues. Psychological Bulletin, 70, 52-
68.
I<ATKIN, E. S., MoRELL, M. A. GoLDBAND, S., BERNSTEIN, G. L. & WisE, J. A. (1982). Indi-
vidual differences in heart beat discrimination. Psychophysiology, 19, 160-166.
KELLY, B., & LACROIX, J. M. (1983, June). The acquisition of heart rate control through biofeedback:
Further evidence against an afferent process. Paper presented at the annual meetings of
the Canadian Psychological Association, Winnipeg.
KIMMEL, H. D., & BAXTER, R. (1964). Avoidance conditioning of the GSR. Journal of Exper-
imental Psychology, 68, 482-485.
LACROIX, J. M. (1981). The acquisition of autonomic control through biofeedback: The case
against an afferent process and a two-process alternative. Psychophysiology, 18, 573-
587.
LACROIX, J. M., & GowEN, A. (1981). The acquisition of autonomic control through bio-
feedback: Some tests of discrimination theory. Psychophysiology, 18, 559-572.
LACROIX, J. M., & RoBERTS, L. E. (1978). A comparison of the mechanisms and some prop-
erties of instructed sudomotor and cardiac control. Biofeedback and Self-Regulation, 3,
105-132.
LACROIX, J. M., CLARKE, M. A., BocK, J. C., DoxEY, N., Wooo, A., & LAvis, S. (1983).
Biofeedback and relaxation in the treatment of migraine headaches: Comparative ef-
fectiveness and physiological correlates. Journal of Neurology, Neurosurgery, and Psy-
chiatry, 46, 525-532.
LACROIX, J. M. CLARKE, M. A., BocK, J. C. & DoxEY, N. C. S. (1984). Biofeedback and
relaxation in the treatment of muscle contraction headache: Effectiveness, predictors,
and physiological correlates. Pain, Suppl. 2, 302.
LEDoux, -.E., WILSON, D. H., & GAZZANIGA, M.S. (1979). Beyond commissurotomy: Clues
to consciousness. In M. S. GAZZANIGA (Ed.), Handbook of behavioral neurobiology: Vol.
2. Neuropsychology. New York: Plenum Press.
LEVENE, H. I., ENGEL, B. T., & PEARSON, J. A. (1968). Differential operant conditioning of
heart rate. Psychosomatic Medicine, 30, 837-845.
LEVENSON, R. W., & DITTo, W. B. (1981). Individual differences in ability to control heart
rate: Personality, strategy, physiological, and other variables. Psychophysiology, 18, 91-
100.
MILLER, N. E. (1981, August). Invited discussion at a symposium entitled New Perspectives
on Biofeedback (D. SHAPIRO, Chair), at the annual meetings of the American Psycho-
logical Association, Los Angeles.
MILLER, N. E., & BRUCKER, B. S. (1979). A learned visceral response apparently independent
of skeletal ones in patients paralyzed by spinal lesions. In N. BIRBAUMER & H. D. KIMMEL
(Eds.), Biofeedback and self-regulation. Hillside, NJ: Erlbaum.
NISBETT, R. E., & WILSON, D. T. (1977). Telling more than we can know: Verbal reports
on mental processes. Psychological Review, 84, 231-259.
0BRIST, P. A., GALOSY, R. A., LAWLER, J. E., GAEBELEIN, c. J., HOWARD, J. L., & SHANKS,
E. M., (1975). Operant conditioning of heart rate: Somatic correlates. Psychophysiology,
12, 445-455.
0FFlfiT, C., & LACROIX, J. M. (1983). Absorption: A predictor of biofeedback learning.
Psychophysiology, 20, 461.
PENNEBAKER, J. W. (1982). The psychology of physical symptoms. New York: Springer-Verlag.
QuALLS, P. J., & SHEEHAN, P. W. (1981). Role of the feedback in electromyograph feedback:
The reJevance of attention. Journal of Experimental Psychology: General, 110, 204-216.
MEcHANISMS oF BIOFEEDBACK CoNTRoL 161
QUY, R. J., & KUBIAK, E. W. (1974). A comparison between "aware" and "naive" conditions
in the suppression of GSR activity. Quarterly Journal of Experimental Psychology, 26,561-
565.
Rn.EY, D. M., & FUREDY, J. ]. (1981). Effects of instructions and contingency of reinforcement
on the operant conditioning of human phasic heart rate change. Psychophysiology, 18,
75-81.
ROBERTS, L. E. (1974). Comparative psychophysiology of the electrodermal and cardiac
control systems. In P. A. 0BRIST, A. H. BLACK, J. BRENER, & L. V. DICARA (Eds.),
Cardiovascular psychophysiology: Current issues in response mechanisms, biofeedback and
methodology. Chicago: Aldine.
RoBERTS, L. E. (1977). The role of exteroceptive feedback in learned electrodermal and
cardiac control: Some attractions of and problems with discrimination theory. In J.
BEATTY & H. LEGEWIE (Eds.), Biofeedback and behavior. New York: Plenum Press.
RoBERTS, L. E., & MARLIN, R. G. (1979). Some comments on the self-description and dis-
crimination of visceral response states. InN. BIRBAUMER & H. D. KIMMEL (Eds.) Bio-
feedback and self-regulation. Hillsdale, NJ: Erlbaum.
RoBERTS, L. E., WILLIAMS, R. J., MARLIN, R. G., FARRELL, T., & IMIOLO, D. (1984). Awareness
of the response after feedback training for change in heart rate and sudomotor lat-
erality. Journal of Experimental Psychology: General, 113, 225-255.
ScHNEIDER, W., & SHIFFRIN, R. M. (1977) Controlled and automatic human information
processing: I. Detection, search, and attention. Psychological Review, 84, 1-66.
ScHOBER, R., & LACROIX, J. M. (1985). Effects of task instructions and contingency on the
development of phasic heart rate control and its correlates. Canadian Journal of Psy-
chology, in press.
Ross, A., & BRENER, J. (1981). Two procedures for training cardiac discrimination: A com-
parison of solution strategies and their relationship to heart rate control. Psychophy-
siology, 18, 62-70.
SCHWARTZ, G. E. (1972). Voluntary control of human cardiovascular integration and dif-
ferentiation through feedback and reward. Science, 175, 90-93.
SHAPIRO, D., ScHWARTZ, G. E., & TuRSKY, B. (1972). Control of diastolic and blood pressure
in man by biofeedback and reinforcement. Psychophysiology, 9, 296-304.
SHAPIRO, D., TuRSKY, B., GERSHON, E., & STERN, M. (1969). Effects of feedback and rein-
forcement on the control of human systolic blood pressure. Science, 163, 588-590.
SHAPIRO, D., TuRSKY, B., & ScHWARTZ, G. E. (1970). Differentiation of heart rate and systolic
blood pressure in man by operant conditioning. Psychosomatic Medicine, 32, 417-423.
STEINER, S. S., & DINCE, W. M. (1981). Biofeedback efficacy studies: A critique of critiques.
Biofeedback and Self-Regulation, 6, 275-288.
SmsoN, D. W., MATUS, I., & BALL, G. (1980). Relaxation and subjective estimates of muscle
tension: Implications for a central efferent theory of muscle control. Biofeedback and
Self-Regulation, 5, 19-36.
SuRwrr, R. S., & KEEFE, F. J. (1983). The blind leading the blind: Problems with the "double-
blind" design in clinical biofeedback research. Biofeedback and Self-Regulation, 8, 1-2.
TELLEGEN, A., & ATKINSON, G. (1974). Openness to absorbing and self-altering experiences
("absorption"), a trait related to hypnotic susceptibility. Journal of Abnormal Psychology,
83, 268-277.
TURK, D. C., MEICHENBAUM, D. H., & BERMAN, W. H. (1979). Application of biofeedback
for the regulation of pain: A critical review. Psychological Bulletin, 86, 1322-1338.
VAN TWYVER, H. B., & KIMMEL, H. D. (1966). Operant conditioning of the GSR with con-
comitant measurement of two somatic variables. Journal of Experimental Psychology, 72,
841-846.
WHITE, T. W., HoLMES, D. S., & BENNETI, D. H. (1977). Effects of instructions, biofeedback,
and cognitive activities on heart rate control. Journal of Experimental Psychology: Human
Learning and Memory, 3, 477-484.
WILLIAMSON, D. A., MoNGUJLLOT, J. E., HuTCHINSoN, P., }ARRELL, M. P., & BLOUIN, D.
(1981). Effect of feedback sensitivity upon learned heart rate acceleration. Psychophy-
siology, 18, 712-715.
7 Learning and Long-term
Physiological Regulation
BARRY R. DwoRKIN
A complex animal is composed of many different semiautonomous ho-

meostatic mechanisms, and compromises must frequently be negotiated
to achieve and maintain a stable overall physiological state optimal for
survival. As evidenced by general skeletal behavior (McFarland, 1971),
the brain has the integrative capacity required to resolve competing de-
mands of these mechanisms. However, the conception of the brain as
a process control computer-receiving data from an array of critically
placed transducers, comparing the measurements to established set
points, and dispatching instructions under a fixed program to a network
of switches, valves, and pumps-is, although an appealing analogy, not
tenable in the light of available data. Critical physiological variables are
regulated to within narrow limits for periods of weeks or even years.
Feedback-stabilized electromechanical regulators, such as thermostats,
can maintain steady-state conditions indefinitely, but to do so they em-
ploy physical sensors or transducers which remain calibrated indefi-
nitely. In contrast all interoceptors adapt (Chernigovsky, 1960; Mount-
castle, 1980; Widdicomb, 1974), and analogous biological control
schemes using interoceptors as sensors could not maintain regulation
for extended periods of time.
Figure 1 shows the response of a typical"slowly adapting" inter-
oceptor on the application of a constant stimulus to its receptive surface
(Mifflin & Kunze, 1982). The decline of impulse firing rate with time is
Parts of this chapter were originally presented as the Hinkle Lecture at Pennsylvania State
University and lectures given at the Max Plan<;:k Institut in Munich and in Budapest during
spring and summer of 1984.
BARRY R. DwoRKIN e Pennsylvania State University College of Medicine, Milton S. Her-
shey Medical Center, Hershey, Pennsylvania 17033. Work on this chapter was supported
by a special fund of the Chronic Pain Unit of the Department of Rehabilitation Medicine,
New York University Medical Center, the Go!dwater Memorial Hospital in New York, and
by the Deutsche Forschungsgemeinschaft (DFG) while the author was Guest Professor in
the Psychologisches Institut of TU.bingen University.
163
164 BARRY R. DwoRKIN
Slowly Adapling
l illll lli l lill ll I I i I
Bll~illlll l~ ilill llil~il llliilllliillli l ;i iilili FIG URE 1. Record of slowly adapting venous
low-pressure vagal receptors of the rat showing
the fall-off in firing rate following application and
maintenance of a stimulus. Adaptation occurs
with both high- and low-strength stimuli. The
classification as "slowly adapting" is made rela-
tive to other interoceptors, studied under similar
012345
T ime (sees)
conditions.
an established feature of all cells which have been studied. Receptors

can only be considered "nonadapting" if observation is limited to min-
utes or at very most hours; extended observation with a carefully main-
tained constant stimulus intensity always reveals a monotonically de-
clining output. Consequently, information about · the absolute
magnitude of the stimulus is progressively distorted by an interoceptor,
and in fact the stimulus may, after a sufficient time, fail to register al-
together; correlatively, a sufficiently gradual stimulus intensity change
may entirely fail to activate the receptor. Nevertheless, in healthy ani-
mals, homeostatic variables do not slowly drift without correction.
Figure 2 is a schematic diagram of the negative feedback linear con-
trol model as it is applied to biological systems. The reference signal or
set-point is the "target" value to which the output is regulated by the
intervening mechanism . A linear systems analysis produces a descrip-
tion of the functional relationships among all of the nodes in the dia-
gram, which can be used to predict the effect of a changed component
characteristic or node condition on the output variable. In the general
case, a component can transform the phase as well as amplitude between
its input and output nodes . Thus, complete analysis requires a full char-
acterization of the gain-frequency and phase-frequency functions for
each component.
Although the linear systems model has become increasingly com-
monplace in the literature of visceral regulation, actual parametric an-
alysis yielding unanticipated explanations of observed physiological
phenomena has been rare . The baroreceptor regulation of blood pres-
sure is an exception: frequency domain analysis has been used to de-
scribe the dynamic characteristics, predict impulse responses, and eval-
LEARNING AND LONG-TERM PHYSIOLOGICAL REGULATION 165
DISTURBANCE
(NOISE)
EFFECTOR OUTPUT
SET POINT VARIABLE
SYSTEM
Er ' - - - - - 1
FIGURE 2. A general systems diagram of a conventional feedback-regulated physiological

control system. G(z) represents the characteristics of the feedback path which may include
an interoceptor monitoring the status of the output variable. The output variable could
be central, such as systemic arterial blood pressure, or local, such as the pH at some point
in the splanchnic circulation. The disturbance or noise may include any load condition
which would require a change in effector activity to maintain the output variable (Eo)
constant. The crossed circles are nodes at which addition of the intersecting inputs occurs.
All such models assume that the relationship between the effector output and the con-
trolled variable is monotonic.
uate hypotheses about the mechanism of observed blood pressure

oscillations (Sagawa, 1983). These studies reveal the potential power of
the method. Respiration, thermoregulation (Satinoff, 1978), caloric bal-
ance, and several endocrine functions are among processes which have
been subjected to analysis which possibly do more honor to the concepts
than the exact methods of linear systems theory.
A general description of the system shown in Figure 2 requires
either application of the La Place transformation technique or the de-
velopment of a specific set of equations which are differential over time.
However, the steady-state or static equilibrium condition at the nodes
can be specified without reference to either time or frequency. Assuming
that the open loop gain of the effector system is relatively high, that is,
that for G(z) 0 for all z, IlEa ~ !lEi, then with the loop closed at
equilibrium:
(1)
For a visceral regulatory mechanism in which the tissue condition

is sensed or monitored by interoceptors, G(z) describes the chain of
neuronal components from the interoceptor to the transmitter release
which modulates the activity of a visceral efferent. The point of inter-
action between the afferent and efferent limbs, that is, the node at which
the transformed output or feedback is subtracted from the input, is in-

dicated in Figure 2 by the summation symbol (2:). In equation (1) G is
considered to be a function of only the output variable, but G may also
depend upon time, as would be true of an interoceptor which adapted
to a constant stimulus. For example, suppose that the adaptation func-
tion is approximated by an exponential, similar to that shown for the
intact Pacinian corpuscle (Lowenstein & Skalak, 1966):
(2) G(z) = -ze-at + K
then, if t = 0, G(z) = - z + K, where z is the tissue condition being

regulated. However, fort ~ 0, G(z) approaches K for all z. In terms of
Figure 2, G(Eo) = K for any Eo. The implication is that after some spec-
ified time the activity of the effector output is not significantly influenced
by the conditions at the output node, that is, negative feedback regu-
lation of the effector is lost. Furthermore, if the adapting interoceptor
is the only feedback component, for a very slowly changing (low fre-
quency) input the gain of the overall system will increase to the open
loop value. Thus, with an open loop gain sufficiently high to achieve
accurate closed-loop regulation of relatively rapid transient distur-
bances, the sensitivity to low frequency input noise could be excessive.
Although equation (2) may not accurately represent any particular
interoceptor, the general finding for all interoceptors is that the output
gradually decays toward the baseline value some time after stabilization
of the stimulus at the receptor surface. For some rapidly adapting in-
teroceptors mechanical high-pass filtering is an important component
of the adaptation mechanism (Lowenstein & Mendelson, 1965); how-
ever, even with the lamellae removed, the generator potential of the
naked Pacinian corpuscle has a duration of less than 100 msec. (Low-
enstein & Skalak, 1966) Other receptors, considered to be particularly
slowly adapting, such as muscle spindle, bee hair-plate, or slow crus-
tacean stretch receptors, can sustain generator potentials for only sec-
onds or minutes. Table 1, taken from Chernigovsky's extensive mon-
ograph (1960), lists the approximate time of adaptation for a number of
carefully studied interoceptors. 1
If, given the strong evidence that adaptation is an inherent property
of interoceptors, the linear systems model as ordinarily composed is not
able to account for steady-state physiological regulation over epochs of
1 Diamond (1955) reported that the output of isolated cat baroreceptors was invariant over
periods of more than an hour; but one of the two curves, which he shows, actually has
a slow but unambiguous negative slope during 80 minutes of observation. Moreover,
the adaptation or resetting of the baroreceptor over hours or days has been extensively
documented.
TABLE 1"
Time of Adaptation for Certain Interoceptors
Numerical data Characteristics of the
characterizing the speed of receptor's adaptation
Receptors adaptation of the receptor speed
IIL In teroceptors
Mechanoreceptors of the 25% of the initial impulse Very slow
carotid sinus (cat) frequency within 6 sec of
stimulation; the initial
rapid phase of adaptation
is practically absent.
Mechanoreceptors of the Complete adaptation within Very slow
aortic arch (rabbit) 5-10 min.
Mechanoreceptors of the Very slow
atrium (cat)
Mechanoreceptors of the 1. Complete adaptation Rapid
urinary bladder (dog) within several fractions of
a second (only 10-30
impulses).
2. Incomplete adaptation Very slow
within 15-25 min.
Mechanoreceptors of the Complete adaptation within Very rapid
urinary bladder (cat) fractions of a second.
Various types of stretch Not more than 5% of the Very slow
receptors of the lungs (cat) initial impulse frequency
within 10 sec.
L Decrease in impulse Slow
frequency by 10-55%
during the 2nd sec of
stimulation.
2. Complete or almost Very rapid
complete adaptation by
the end of the first sec of
stimulation. These
receptors are
distinguished by a high
threshold.
• From Chemigovsky, 1960.
weeks, months, or years, is there some plausible modification of the

model which can? Certainly there are possibilities; for example, the in-
clusion of a peak-detecting memory element in the feedback loop would
permit the storage of maximum and possibly also minimum values of
interoceptor activity for very long periods of time. Unlike the sensory
transduction process, long-term memory in particular appears to have
almost unlimited temporal stability. However, the relationship between
peak interoceptor activity and the effector output which would produce
satisfactory regulation is not obvious; any peak detection or peak av-
eraging scheme has the additional and probably more serious defect that
very slow but consistent drift in the regulated variable could entirely
fail to activate the interoceptor and eventually accumulate into a con-
siderable regulatory error. Consequently, simple integration or sum-
mation of peak interoceptor activity within G(z) could not completely
compensate for the inherent low-frequency response limitations of
interoceptors.
Another possibility is mechanical modulation of the stimulus on the
interoceptor surface. This would be analogous to the electronic instru-
mentation technique used to measure very low frequency signals (<0.1
Hz.) with high-stability capacitor or transformer coupled amplifiers. A
modulator or chopper converts the low-frequency signal into a high-
frequency signal with an amplitude equal to the full difference between
the instantaneous value of the low-frequency signal and some known
reference. The modulated signal is amplified and another circuit elim-
inates the reference phase and filters or smooths the interrupted wave-
form to reconstruct the original signal. This arrangement achieves a high
degree of amplification with minimal drift, using only circuits with short
time constants. Because of the pulsitile nature of the cardiac cycle, certain
central cardiovascular interoceptors are naturally exposed to regular me-
chanical "chopping" or modulation of the stimulus. This could prevent
adaptation of those receptors; however, because in the biological system
the modulation is incomplete, extraction of a steady-state variable, such
as intravascular volume, would require an additional second type of
receptor to measure the modulation independently. While such arrange-
ments may exist, 2 this scheme would have a restricted domain because
autonomic efferent control extends to anatomical regions in which pulse
is almost absent or poorly correlated with measurable central contractile
activity. In sum, the observed stability of many different physiological
variables taken with the relatively rapid adaptation of all known inter-
oceptors creates an internal contradiction for a linear systems model of
long-term physiological regulation, which relies upon conventional sen-
sory-neural transduction in the feedback loop.
2 In principle, the atrial A and B receptors (Gilmore, 1983) have the respective character-
istics and situations for measuring the modulation and peak-to-peak pressure amplitude.
Thus, for certain central cardiovascular receptors receiving a mechanically modulated
stimulus a relatively complicated scheme of independent modulation measurement and
compensated demodulation of the signal could extend the low-frequency response of
G(z) sufficiently to transduce accurately a slowly changing variable, such as central
volume.
LEARNING AND LoNG-TERM PHYSIOLOGICAL REGULATION 169
One possible conclusion from these data and arguments is that sen-
sory-neural transduction has no role in long-term regulation. This pos-
sibility was extensively developed by Granger and Guyton (1969), who
proposed that the constancy and integration of central physiological
variables is achieved by a kind of "whole body" autoregulation which
does not involve nervous mechanisms. Their model is an assemblage
of reasonably well documented humoral, physical, and reflex mecha-
nisms organized into a linear systems framework, emphasizing the mu-
tual interdependence of the components and the predominant role of
intrinsic properties of the peripheral circulation in long-term regulation
of central variables. In it regulation occurs at a local tissue level through
a variety of autoregulatory mechanisms responsive to p02, pC02, and
pH, as well as to specific metabolic products and vasoactive interme-
diaries; thus, each tissue's normal requirements are adjusted by intrinsic
mechanisms not directly involving reflexes (Guyton, 1977). Central reg-
ulation is a net consequence of the local processes. As the vascular bed
of a tissue dilates in response to metabolic requirements, its conductance
increases, and the supplementary venous return to the heart enhances
cardiac output enough to meet the additional requirements of the tissue.
Guyton maintains that this mechanism alone has sufficient capacity· to
regulate cardiac output over a wide range of conditions short of stren-
uous exercise. The following quotations will provide some sense of the
logic and assumptions of this theory:
When local tissues vasodilate in an attempt to supply themselves with ad-
equate blood flow, this instantly increases blood flow from the arteries into
the veins. The increased venous pressure in tum causes increased venous
return. And, finally, the heart responds to this increased venous return by
increased pumping mainly because of the Frank-Starling mechanism ....
Even a 1 to 2 mm Hg rise in right atrial pressure distends the heart sufficiently
to double the cardiac output. In this way cardiac output automatically adjusts
itself to the venous return. (Guyton, 1977, p. 763)
However, in the final analysis cardiac output is determined by the

central or "background" control of blood volume by the kidney, which,
as Guyton emphasizes, is the only significant long-term determinant of
arterial pressure.
In the minds of both physicians and physiologists, arterial pressure control
is most often believed to be achieved either entirely or almost entirely through
nervous mechanisms. However, as we shall see, this is far from true. (Guy-
ton, 1977, p. 764)
And blood volume regulation by the kidney is primarily achieved

through the pressure diuresis/natriuresis phenomena:
Therefore, the overall mechanism of the blood volume system for pressure
regulation is the following: When arterial pressure rises too high, the kidneys
automatically begin to excrete fluid. Furthermore, they will not stop excreting
fluid until the arterial pressure returns to its original value. Conversely, when
the arterial pressure falls below normal, the kidneys retain fluid, and again
they will not stop retaining fluid until the pressure rises to its normal value.
(Guyton, 1977, p. 766)
The complete Guyton-Granger model (1972) is complicated and al-

though some isolated sections of the model have been experimentally
verified, little evidence for its sufficiency or quantitative plausibility has
been developed. It is particularly unclear whether parameters can be
found for its numerous interacting regulatory loops, which will produce
both dynamic stability and realistic response speed. In addition, the
model does not acknowledge the hierarchical and redundant structure
of most vertebrate regulatory mechanisms. The fact that the renal output
curve ultimately controls blood pressure is true but not very informative without
a quantitative assessment of the role of different supervisory variables in estab-
lishing the parameters of the curve. For example, Guyton emphasizes the
importance of pressure diuresis and secondarily the renin-angio-
tensin system, but stimulation of the renal nerves can also significant-
ly increase the pressure range of the kidney output curve. In fact, suffi-
ciently strong sympathetic activation, as evoked by cerebral ischemia,
can cause total shut-down. Chronic recordings in freely moving cats
(Schad & Seller, 1975) reveal the presence of a continuous background
activity in the renal nerves. This tonic efferent discharge is reduced by
elevated blood pressure, proportionally increased by different levels of
exercise, and almost completely eliminated by ganglionic blockade.
Thus, in the unanesthetized freely moving animal the kidney output
curve is almost certainly under some degree of tonic CNS control. And,
although the fact that the baroreceptors adapt has been seen correctly
by Guyton as a serious problem with regulatory theory based on reflexes:
Earlier in this paper we stated that long-term control of arterial pressure is
entirely different from acute control. There are 2 reasons for this. First, most
of the nervous mechanisms adapt with time so that they have progressively
less effect on the circulation after the 1st few minutes to the 1st few hours
of activity. For instance, the baroreceptors gradually reset (adapt) to a new
pressure level in less than 2 days. (Guyton, 1977, p. 766)
this observation does not justify the conclusion that nervous mecha-
nisms do not affect long-term blood pressure regulation.
Measurement of the individual time constants and loop gains may
eventually show that the entire regulatory jigsaw puzzle fits neatly to-
gether at a peripheral level without the necessary involvement of the
brain in long-term regulation; however, the data to build that model
have not been assembled and the fact that the system could in some
manner function without the brain (Granger & Guyton, 1969) would not
imply that it ordinarily did. Furthermore, there is now a significant body
of evidence implicating tonic CNS efferent activity in steady-state
regulation.
Experiments demonstrating shifts in long-term regulation following
nerve section or pharmacological blockade of ganglionic transmission
have provided the most direct evidence of the functional presence of
tonic autonomic control of cardiovascular function. 3 The effect of total
spinal anesthesia on blood pressure is a clear example. Loss of neuro-
genic vasomotor tone can reduce mean arterial pressure from 100 mmHg
to 50 mmHg or less, and injection of very small doses of norepinephrine
can immediately restore the original resting pressure (Guyton, 1982).
This and similar experiments confirm the normal presence of vasomotor
tone and its importance as a variable in long-term cardiovascular
regulation.
Tonic sympathetic control of the resistance vessels has implications
primarily for the regional distribution of the cardiac output; a similar
control of venous tone is more important for central cardiovascular ho-
meostasis. Since the veins have little if any myogenic tone, the sym-
pathetic efferents are the major determinants of venous compliance.
Thus, hexamethonium ganglionic blockade causes a decrease in the
mean circulatory filling pressure requiring an infusion of 6-10 ml!Kg to
restore the original level (Rothe, 1976). Several other studies report equal
or greater estimates of neurogenic venous tone (Rothe, 1983). These
results show that the total-body pressure-volume relationship, venous
return, and, thus, the cardiac output are substantially regulated by tonic
sympathetic nerve activity and, as with the sympathetic control of the
kidney output curve, underline the obligatory role of autonomic effer-
ents in long-term cardiovascular regulation. In addition to these "global"
effects of autonomic activity there are other data on the independent
distribution of sympathetic vasomotor tone to different tissues.
The vasculature of both the skin and adipose tissue are under in-
dependent tonic sympathetic control. Following therapeutic sympa-
thectomy, blood flow in the adipose tissue of human limbs increases by
approximately 100% (Henriken, 1977). The vasomotor control of the skin
has been particularly well doc~mented because of its role in thermo-
3 Sympathetic and parasympathetic tonic control of a variety of functions in different organ

systems has been demonstrated, and most of what is discussed in the remainder of this
chapter is in principle equally applicable to noncardiovascular visceral regulation. How-
ever, only for cardiovascular and thermoregulatory control has long-term regulation been
conceptualized in a sufficiently quantitative manner to permit a rigorous examination of
the assumptions and logic; and thermoregulation involves specific brain receptors for
which the primary adaptation characteristics are not yet established.
regulation. Since either vasoconstriction or vasodilatation can be in-

duced by nerve block depending on whether the subject is being heated
or cooled, and human forearm skin probably has tonically active va-
sodilator as well as constrictor mechanisms (Roddie, 1983), the steady-
state sympathetic tone is difficult to estimate. 4 thus, although cutaneous
nerve block under a thermoneutral condition produces little change in
skin blood flow, the implications of that result for tonic control are not
clear.
The tonic skeletal muscle vasoconstriction which occurs during REM
sleep can be eliminated by regional sympathectomy, but nerve block in
resting awake animals appears to have little net influence on muscle
blood flow. However, in steady-state exercise sympathetic vasoconstric-
tion significantly reduces local metabolic vasodilatation, and Shepherd
(1983, p. 352) has noted, "The role of the sympathetic nerves may be
to modulate the local dilator mechanisms to maintain the most econom-
ical ratio of blood flow to Oz extraction."
Neurophysiological data concerning the characteristics of the sym-
pathetic discharge verify the presence of a continuous background com-
ponent in sympathetic nerve activity. These studies provide both detail
about the characteristics and sources of variability of the sympathetic
outflow and, along with other experiments demonstrating significant
target-tissue effects of stimulating the same nerves, further evidence of
the continuous presence of a functional sympathetic tone. Observations
of continuous background discharges have been made from whole sym-
pathetic nerves of freely moving unanesthetized cats (Schad & Seller,
1975) and humans (Delius, Hagbarth, Hongell, & Wallin, 1972), indi-
vidual postganglionic neurons of cat cervical nerve (Mannard & Polosa,
1973; Polosa, Mannard, & Laskey, 1979), and intracellularly from cat
and rabbit superior cervical ganglia (Mirgorodsky & Skok, 1969).
Detailed statistical analysis of the background firing, that is, con-
tinual discharge in the absence of experimental provocation, of individ-
ual sympathetic cervical preganglionic neurons (Mannard & Polosa,
1973) has revealed a complex pattern of activity containing burstlike
rhythmic and continuous random components. The more regular activ-
ity is probably generated by oscillation in short-term reflex pathways,
but the steady stochastic background is largely determined by supra-
4 In general, nerve block or section experiments can lead to false negative conclusions
regarding the presence of sympathetic tone if the outflow is distributed to antagonistic
effectors, balanced within the field of observation; in contrast, finding that a block or
lesion changes the resting state of a tissue is almost incontrovertible evidence of preex-
isting tonic influence of the nerve. The fact that reciprocal innervation may be balanced
so that blocking both aspects is without net effect does not obviate the need for inde-
pendent tonic control of each aspect.
(J)
-'
< FIGURE 3. Interval histograms showing the statistical
>
a: properties of sample neurons studied by Mannard and
w
1-
~
Polosa (1973) in the cat spinal cord near T1 . Cervical
w units were identified by supramaximal antidromic stim-
:.: ulation of the ipsilateral sympathetic nerve. A is a pen-
0::
(J)
a: tobarbital anesthetized preparation with an intact neu-
w raxis, B is decerebrate, C transected at the cervical level,
1-
z and D a decentralized isolated cord segment. The bin
u.
0 widths are 64, 16, 32, and 96 msec., respectively; n.b.
a:
w
in C and D the reduced number of spikes and restricted
m variability in interspike interval following the elimina-
:;
~ 0 tion of supraspinal influences. The rate, averaged over
all units observed, was reduced to approximately 25%
of that for intact and decerebrate preparations. This
study compared the characteristics of the irregular back-
ground units shown in the figure to other rhythmical
o 1o 20 so 40 5o eo
firing units in the same region; an approximately equal
BIN NUMBER
number of the two types were found .
spinal influences. Figure 3 shows distributions of interspike intervals ob-

served in stochastic-type postganglionic neurons. The reduced varia-
bility and net activity observed after chord section or segmental isolation
indicates the contribution of supraspinal inputs to the background ac-
tivity. These changes are thought to reflect both modified firing patterns
and reduced numbers of active units following decentralization. This
result generally resembles that for alpha-motorneurons and underlines
the similarity between the sympathetic efferent and skeletal motor
outputs.
In sum, the notion that the nervous system participates only in short
term-dynamic adjustments and not long-term or steady-state regulation
has gradually become less tenable, and with it the idea of "whole-body"
autoregulation. Nevertheless, Guyton, Coleman, and Granger's iden-
tification (1972) of the autoregulatory process as having the capability
of providing a stable reference level-independent of neural adapta-
tion-was an extremely valuable insight. Local autoregulatory systems
which depend only on chemical or mechanical feedback can both re-
spond rapidly to transient perturbations and have long-term stability.
In normal tissue, constant pH is possibly the most important condition
maintained by autoregulation. The local autoregulatory apparatus has
many different components; fundamental among them are the chemical
buffering systems of the intravascular, interstitial, and intracellular
spaces. These molecular-level mechanisms vary in complexity from the

hemoglobin binding of carbon dioxide to the bicarbonate/carbonic acid
equilibrium system and the simple bulk affinity of protein for hydrogen
ions. At a higher level of integration there are specific hemodynamic
mechanisms which respond to elevations in metabolic product concen-
tration with relaxation of sphincters and changes in vessel wall tension.
Other mechanisms maintain a relatively constant regional blood flow
independent of variations in central blood pressure through local hu-
moral mechanisms and nonlinear mechanical compliance. There is,
thus, a hierarchical structure in each tissue bed responsible for main-
taining a constant local environment compatible with normal cell func-
tion (Johnson, 1964; Morff & Granger, 1982). This system responds rap-
idly and can maintain a nonadapting reference level indefinitely, as it
is independent of sensory-neural transduction or in some instances even
receptor-ligand interactions.
However, these local mechanisms by themselves can do little to
assure the efficient allocation of resources among the various homeo-
static mechanisms present in an organism. At times the cardiac output
must be simultaneously divided among muscular, digestive, thermo-
regulatory, CNS, renal, and other requirements-with at least some
structures necessarily receiving a less than adequate share. The impor-
tance of instrumental or Type II learning in controlling the distribution
of the resources of skeletal behavior among competing homeostatic re-
quirements is well established (Baum, 1974; Herrnstein, 1961; Sibby &
McFarland, 1964; Weardon & Burgess, 1982). The learned regulation of
eating and drinking has been studied for years by behavioral psychol-
ogists; an analogous role of learned visceral responses for autonomic
function by regulation of the tonic sympathetic background would be
phylogenetically parsimonious. But, assuming that Type II learning is
applicable to visceral responses (Dworkin & Miller, 1977; Dworkin, 1980;
1984), how could it participate in physiological regulation given the rapid
adaptation and limited specificity of interoceptors?
Ultimately the stabilization of pH through autoregulation depends
upon adequate blood flow, and flow depends upon pressure. Autore-
gulatory sufficiency of a tissue exists when blood pressure is at a value
that permits it a maximum range of deviation without affecting blood
flow and blood flow can vary maximally without affecting pH. Under
these conditions local pH is well stabilized against metabolic or central
perturbations and an interoceptor sensitive to pH is infrequently or
never activated; consequently, it will adapt and become silent. Thus,
small regulatory wrinkles, for example, in blood pressure or local met-
abolic activity, will not be registered by the nervous system. For a pH
receptor to be activated either individual blood pressure or metabolic
CENTRAL
BLOOD
PRESSURE c D
A
·················
~~~~[--------------------------------
FLOW
pH[:--------------------------------------
~~~~~~J 0 R-------rlrl-rl-----r----+-------1---~I~IHI~II~IU~III~IQ~I---r--1---
TIME
>
FIGURE 4. These hypothetical curves are constructed from well-established mechanisms
(Johnson, 1964) and receptor characteristics (Mountcastle, 1980). They illustrate how re-
ceptor adaptation and autoregulation probably interact in a perfused metabolically active
block of tissue. Because of autoregulation, the blood flow through the section remains
constant despite a pressure transient at A. But from B to C there is a gradual drift in central
pressure, and the limit of blood flow autoregulation is approached; subsequently, another
pressure transient, beginning at C, exceeds the limit (dotted line) and local flow falls; a
relatively rapid accumulation of metabolic waste (reflected in the pH) triggers the inter-
oceptor, which is sensitive to falling pH.
transients must be imposing enough to swamp the autoregulatory ap-

paratus, or a chronic state of imbalance must gradually evolve, which
compromises the regulatory compliance by moving the operating point
out of the autoregulatory and into the linear or high gain portion of the
regulation-transfer function. With autoregulation thus strained, even a
small perturbation could cause interoceptor activation. Figure 4 shows
more specifically how this might work: an interoceptor is embedded in
a block of tissue with a blood supply supporting local metabolic activity
under steady-state conditions. This may be muscle in a state of mixed
oxidative and anerobic metabolism--producing constant quantities of
carbon dioxide and lactic acid, which are steadily transported out of the
tissue by blood flowing through the capillary bed. The interoceptor is
sensitive to the pH of the tissue; if the pH remains constant for an
extended time, it will adapt and cease to fire. The dotted line represents
the lowest systemic arterial pressure for which intrinsic tissue mecha-
nisms can compensate. If blood pressure remains above this limit then
the local autoregulatory apparatus is able to buffer fluctuations in pres-
176 BARRY R. DWORKIN
sure, and changes in pressure, such as at point A in the slide, will have
no appreciable effect on either blood flow or pH. In contrast, when the
pressure gradually drifts so that the local autoregulatory apparatus even-
tually becomes incompetent, although the drift may be so slow as not
to be directly registerable by an interoceptor, occasional transient per-
turbations will periodically exceed the dynamic range of the regulator
and cause fluctuations which stimulate the interoceptor. At point C the
pressure has drifted down sufficiently for the dynamic limit of autore-
gulation to be approached, and the perturbation between C and D,
which is no larger than at A, causes a decrease in flow. The temporary
reduction in flow changes the metabolic balance in the tissue, reducing
oxygen delivery and allowing for a rapid accumulation of acidic waste,
which overwhelms the buffers and lowers the pH of the surface of the
interoceptor. If the rate of change of pH exceeds the rate of adaptation
of the interoceptor, it will fire. Given this analysis, the information trans-
mitted to the brain by the local interoceptor is that the conditions in the
tissue are getting close to the autoregulatory limit or "ragged edge."
However, simply saying that information is transmitted to the brain is
relatively meaningless without describing the mechanism that accom-
modates and uses the information. The remainder of this paper will
attempt to explain how Type II visceral learning could be that
mechanism.
The fundamental process of Type II learning is the association in
time between a response and a subsequent reinforcing stimulus (Woody,
1982, p. 46 ff.). Similarly, the association between an autonomic re-
sponse and interoceptor activity, signaling autoregulatory insufficiency,
would be the crucial feature of Type II visceral learning. For visceral
learning to modify the subsequent response probability and refine au-
tonomic regulation it is necessary only that increased interoceptor firing
closely follow the response, repeatedly, so that the nervous system can
detect the correlation against the background of random events.
In Figure 5 each group of interoceptors is connected to only a single
functionally related central efferent in an arrangement which resembles
a conventional reflex regulator but works with adapting interoceptors.
The tissue receives an autonomic efferent A from the CNS which either
directly controls a specific local function, such as vasoconstriction, or
sets the sensitivity of an autonomic reflex or intrinsic regulatory mech-
anism (Shepherd, 1983). The interoceptor I is sensitive to a specific con-
dition of the tissue such as pH. Local autoregulation maintains the pH
constant over a wide range of conditions, including considerable random
variation in the firing rate of efferent nerve A. Thus, ordinarily the in-
teroceptor is in a state of adaptation. However, some firing rates of A
are incompatible with local requirements and cause the limit of auto-
Memory*
p(AJ TYPE II (a)

VISCERAL LEARNING
t1 A
t2 A *Learning rule:
Interoceptor fires==:>
t3 )\_ Decrement the distribution

at the current value
ln A Specific
Interoceptor
Autonomic
Efferent I
A
Tissue
p(IJA 1 )=1
p(IIA,)-.;;1
FIGURE 5. This diagram shows how long-term regulation can be achieved with adapting
interoceptors and Type II visceral learning. The interoceptor (I) is specific in the sense that
it can modify only a single structurally predetermined efferent system. Firing of the in-
teroceptor causes the concurrently present efferent activity level (A;) subsequently to have
a reduced probability p(Ai)· As indicated by the conditional probabilities, interoceptor
firing occurs when A = A L, but not at other values; consequently, repeated events, ti-
tn, result in a cumulative reduction of the probability of additional occurrences of the
particular activity level, A. In addition to involving a memory process, this scheme differs
from a conventional negative feedback linear control model because monotonicity is not
assumed for the relationship between the tissue condition and efferent activity level, Aj.
regulation to be exceeded, the pH to change, and the interoceptor to be

activated. When this happens the Type II learning mechanism decre-
ments or reduces the probability of the most recent value (Ai) of the
activity of the efferent nerve. This is repeated with each firing of the
interoceptor, and successive events of this kind eventually produce a
cleft in the distribution of p(Ai) which eliminates the incompatible value.
Since the distributions are of probabilities, the total area remains con-
stant even when a specific value is decremented and this feature can
obliterate a cleft as well as form one: decrementing all except certain
values is equivalent to incrementing those values. An anatomical sub-
strate for this mechanism is not difficult to imagine and the analog for
skeletal behavior is extinction of the learned response.
In this model the requirement of long-term stability is repositioned
from the sensory receptor to the CNS, and the physiological reference
level or set point is repositioned from the CNS to the autoregulated
region surrounding the receptor. The result in both cases is consonant
with contemporary physiological data and eliminates the paradox of
receptor adaptation; however, there are some further implications of the
involvement of learning in physiological regulation. In addition to stor-
ing information learning organizes behavior. The animal is a collection
of regulatory mechanisms or feedback loops; each may have several
inputs and outputs interacting with one another at spatially separated
points. If the resources of the system are limited so that at least some
of the variables must be chronically displaced from their optimal set
point and if multiple criteria exist for overall system performance, an
imposing problem of optimization emerges. Solving this problem is nec-
essary to the survival of all complex animals. Earlier, I suggested that
Type II learning had a well-established role in the matching of the ho-
meostatic requirements and behavior of animals, through eating, drink-
ing, and other skeletal responses. A mechanism for the optimization of
multiple physiological variables through Type II visceral learning can
be developed from a relatively direct extension of the model proposed
for single-variable steady-state regulation. Figure 6 is identical to Figure
5 except that two more efferent systems, B and C, have been added.
The probability distributions for these efferents are affected by recent
interoceptor firing exactly as the distribution for A is; however, as the
situation is depicted variation of activity in these efferents does not affect
the status of the tissue sufficiently to disturb regulation and cause the
interoceptor to fire. Consequently, interoceptor firing and the efferent
activity on B or C are uncorrelated. Therefore, the value decremented
in distributions p(Bi) and p(C) is different each time the interoceptor
fires, and since the area under a probability distribution is constant,
each operation cancels out the others without having a net effect. In
contrast, since interoceptor firing is evoked by a certain activity level on
efferent A, distribution p(Ai) will be repeatedly decremented at that
same value, eventually forming a cleft at the activity level incompatible
with autoregulation.
If conditions changed so that fluctuations in C also began to strain
the autoregulatory compliance of the tissue, a correlation would emerge
between the interoceptor firing and the newly incompatible values of
C. These new conditions could be due either to a change in local re-
quirements in the tissue, so that previously satisfactory values of C be-
came inadequate, or to a shift in the distribution of p(C;) produced by
some central disturbance or input. Under the learning rule the distri-
bution of p(C) would now be decremented at the values associated with
interoceptor firing. Thus, this arrangement is a mechanism for both iden-
Memory*
p(A;) p(B;) p(C;) TYPE II (b)

VISCE RAL LEARNING
tl
I
_/,\_ A A
t2 A A /\l *L earning rule:
A A Y\._
lntero ceptor fires =:>
Dec rem ent all distributions
t3 at th e current value
tn A A A 14-
I
Nonspecific
Interoceptor
Autonomic Efferents
A 8 c ~
Tissue
_ ~
1
\p(I)A 1 )"'1 p(IIB,Hl P<rlc, >~1
[ Local
Autoregulation j
FIGURE 6. If the effect of interoceptor firing is not limited to a single efferent but distributes
over several, under the Type II learning rule only the probability distributions of those
efferents which significantly influence the tissue condition will be modified. The other
distributions will be decremented as frequently (see Figure 5), but randomly, and the net
effect will not alter their shape. With this arrangement, Type II learning could control the
allocation of resources among different anatomical structures and also select the particular
regulatory mechanisms which minimized net homeostatic imbalance. The behavioral foun-
dations have been discussed in more detail elsewhere (Dworkin, 1980; Dworkin, 1984;
McFarland, 1971).
tifying autonomic variables that assume inappropriate values and ad-

justing them.
Autonomic regulation based on visceral learning lends itself natu-
rally to negotiation among the competing requirements of individual
organ systems and tissues of a complex animal. For example, suppose
that blood flow is required in both the mesenteric bed for digestion and
in skeletal muscle for maintaining contraction, and that cardiac output
is insufficient to satisfy both requirements completely. Optimal regional
distribution, given the constraints, could be accomplished if the net ac-
tivity of intestinal and skeletal muscle interoceptors were able, through
Type II learning, to modify the autonomic efferents controlling shunts
in both tissues_ When regulatory compliance was exceeded in either
tissue, its interoceptors would fire, reducing the probability of reoc-
currence of the particular activity levels present on the mesenteric and
skeletal vascular efferents at that moment. Interoceptor activity would
be greatest when vasoconstriction was excessive in both circulations and

least when both were autoregulating. If the mesenteric circulation was
receiving more flow than necessary at the expense of skeletal muscle,
the surplus would be wasted, because interoceptor activity cannot be
reduced below that value present when flow is minimally sufficient for
autoregulation. But the net interoceptor activity level would be higher
than at an optimal adjustment because the skeletal bed, being deprived,
would be in the linear portion of the regulation transfer function, and
consequently the complementary deficit in flow would be reflected in
proportionally increased metabolic imbalance and more frequent inter-
oceptor firing. If interoceptor activity is additive, this process could reg-
ulate many different regions and kinds of tissue simultaneously without
reliance on unrealistic and undemonstrated peripheral neuroanatomical
specificity. Either greater density or heightened sensitivity of interocep-
tors in a critical tissue could establish differential priorities at a peripheral
level.
Finally, it is a logical consequence of this scheme that the reference
level or set point of a variable is not a centrally represented parameter
but is partially preserved in each individual tissue by its requirement
for maintaining an ability to compensate for perturbations through an
autoregulatory response. If that ability is maintained, the interoceptors
in the tissue will not be exposed to stimulus fluctuations of sufficient
magnitude to cause impulse generation-they will remain silent and
not signal the CNS. Functional demands in the tissue interact with cen-
tral variables to determine the compliance limits for autoregulation. For
example, if a region is metabolically active it will require a higher min-
imum blood flow and correspondingly higher minimum central pressure
to maintain acid-base balance and adequate oxygen tension than are-
gion which is quiescent. Given this perspective, a set point is not a
central parameter, but a "consensus" among multifarious local auto-
regulatory regions.
I have described a theory of autonomic regulation differing signif-
icantly from the current paradigm, which depends either upon un-
realistic interoceptor characteristics or ignores the role of the tonic sym-
pathetic outflow in long-term regulation. Although somewhat
unorthodox because it interweaves behavioral and physiological prin-
ciples, the new theory incorporates only well-established mechanisms
with a single exception: it depends upon the applicability of Type II
learning to visceral responses. Thus, demonstrating the existence of vis-
ceral learning is a necessary first step in verifying the theory and ad-
mitting the possibility that learning could participate in some autonomic
regulatory paradigms. This would require that learning be considered
as a possible mechanism in well-designed physiological experiments.
However, actual proof that visceral learning is significant in physiolog-

ical regulation would require specific analytical experiments in which,
for example, the homeostatic consequences of natural visceral regulatory
responses were experimentally manipulated (Dworkin & Miller, 1977).
Although not categorical, the demonstration in a paralyzed animal
of Type I learning or classical conditioning of a visceral response, such
as a change in regional blood flow, and of Type II learning of a skeletal
response such as firing a motor nerve, but not Type II learning of the
same visceral response, would be strong evidence against a physiolog-
ically significant Type II visceral learning, as well as against the regu-
latory hypothesis presented in this paper.
REFERENCES
BAUM, W. M. (1974). On two types of deviation from the matching law: Bias and under-
matching. Journal of the Experimental Analysis of Behavior, 22, 231-242.
CHERNIGOVSKY, V. N. (1960). Interoceptors. Moscow: State Publishing House of Medical
Literature.
DELIUS, W., HAGBARTH, K. E., HoNGELL, A., & WALLIN, B. G. (1972). General characteristics
of sympathetic activity in human muscle nerves. Acta Physiologica Scandinavica, 84, 65-
81.
DIAMOND, J. (1955). Observations on the excitation by acetylcholine and by pressure of
sensory receptors in the eat's carotid sinus. Journal of Physiology, 130, 513-532.
DwoRKIN, B. R. (1984). Operant mechanisms in physiological regulation. InT. ELBERT, B.
RocKSTROH, W. LUTZENBERGER, & N. BIRBAUMER (Eds.), Self regulation of the brain and
behavior. Berlin: Springer-Verlag, pp. 296-309.
DwoRKIN, B. R. (1980). The role of instrumental learning in the organization and main-
tenance of physiological control mechanisms. In G. ADAM, I. MESZAROS, & E. I. BANYAI
(Eds.), Advances in physiological sciences: Brain and behavior (Vol. 17) Budapest: Aka-
demiai Kiado, pp. 169-176.
DwoRKIN, B. R., & MILLER, N. E. (1977). Visceral learning in the curarized rat. In G.
ScHWARTZ & J. BEATTY (Eds.), Biofeedback: Theory and research. New York: Academic
Press, pp. 221-242.
GILMORE, J. P. (1983). Neural control of extracellular volume in the human and nonhuman
primate. In J. T. SHEPHERD & F. M. ABBOUD (Eds.), The handbook of physiology: The car-
diovascular system (Vol. 3, Sec. 2). Baltimore: Williams & Wilkins, pp. 885-915.
GRANGER, H., & GUYToN, A. C. (1969). Autoregulation of the total systemic circulation
following distruction of the central nervous system in the dog. Circulation Research,
25, 379-388.
GUYTON, A. C. (1977). An overall analysis of cardiovascular regulation. Anesthesia and An-
algesia, 56, 761-768.
GUYToN, A. C. (1982). Textbook of medical physiology (6th ed.). Philadelphia: W. B. Saunders.
GUYTON, A. C., CoLEMAN, T. G., & GRANGER, H. J. (1972). Circulation: Overall regulation.
Annual Review of Physiology, 34, 13-46.
HENRIKEN, 0. (1977). Local sympathetic reflex mechanism in regulation of blood flow in
human subcutaneous tissue .. Acta Physiologica Scandinavica, Suppl., 101, 1-48.
HERRNSTEIN, R. J. (1961). Relative and absolute strength of response as a function of fre-
quency of reinforcement. Journal of the Experimental Analysis of Behavior, 4, 267-272.
JoHNSON, P. C. (1964). Review of previous studies and current theories of autoregulation.

Circulation Research (supplement 1), 15, 2-9.
LowENSTEIN, W. R., & SKALAK, R. (1966). Mechanical transmission in a pacinian corpuscle:
An analysis and a theory. Journal of Physiology, 182, 346-378.
LowENSTEIN, W. R., & MENDELSON, M. (1965). Components of receptor adaptation in a
pacinian corpuscle. Journal of Physiology, 177, 377-397.
McFARLAND, D. J. (1971). Feedback mechanisms in animal behavior. London and New York:
Academic Press.
MANNARD, A., & PowsA, C. (1973). Analysis of background firing of single sympathetic
preganglionic neurons of cat cervical nerve. Journal of Neurophysiology, 36, 398-408.
MIFFLIN, S. W., & KuNZE, D. L. (1982). Rapid resetting of low pressure vagal receptors in
the superior vena cava of the rat. Circulation Research, 51, 241-249.
MIRGORODSKY, V. N., & SKoK, V.I. (1969). Intracellular potentials recorded from a tonically
active mammalian sympathetic ganglion. Brain Research, 15, 570-572.
MoRFF, R. L & GRANGER, H. J. (1982). Autoregulation of blood flow within individual
arterioles in the rat cremaster muscle. Circulation Research, 51, 43-55.
MouNTCASTLE, V. (1980). Medical physiology (14th ed.). St. Louis: Mosby.
PoLOSA, C., MANNARD, A., & LASKEY, W. (1979). Tonic activity of the autonomic nervous
system: Functions, properties, origins. In C. BROOKS, K. KOIZUMJ, & A. SATO (Eds.),
Integrative functions of the autonomic nervous system. Tokyo: University of Tokyo Press,
pp. 342-354.
RoDDIE, I. C. (1983). Circulation to skin and adipose tissue. In J. T. SHEPHERD & F. M.
ABBOUD (Eds.), The handbook of physiology: The cardiovascular system (Vol. 3, sec. 2).
Baltimore: Williams & Wilkins, pp. 397-452.
ROTHE, C. F. (1983). Venous system: Physiology of the capacitance vessels. InJ. T. SHEPHERD
& F. M. ABBOUD (Eds. ), The handbook of physiology: The cardiovascular system (Vol. 3, sec.
2). Baltimore: Williams & Wilkins, pp. 285-317.
RoTHE, C. F. (1976). Reflex vascular capacity reduction in the dog. Circulation Research, 39,
705-710.
SAGAWA, K. (1983). Baroreflex control of systemic arterial pressure and vascular bed. In J.
T. SHEPHERD & F. M. ABBOUD (Eds. ), The handbook of physiology: The cardiovascular system
(vol. 3, sec. 2). Baltimore: Williams & Wilkins, pp. 453-496.
SATINOFF, E. (1978). Neural organization and evolution of thermal regulation in mammals.
Science, 201, 16-22.
ScHAD, H., & SELLER, H. (1975). A method for recording autonomic nerve activity in un-
anesthetized, freely moving cats. Brain Research, 100, 425-430.
SHEPHERD, J. T. (1983). Circulation to skeletal muscle. In]. T. SHEPHERD & F. M. ABBOUD
(Eds.), The handbook of physiology: The cardiovascular system (Vol. 3, sec. 2). Baltimore:
Williams & Wilkins, pp. 319-370.
SIBBY, R., & McFARLAND, D. (1964). A state-space approach to motivation. In D. J. McFAR-
LAND (Ed.), Motivational control systems analysis. London: Academic Press, pp. 1-84.
WEARDON, J. H., & BuRGESS, I. S. (1982). Matching since Baum. Journal of the Experimental
Analysis of Behavior, 38, 339-348.
WmmcOMB, J. G. (1974). Enteroceptors. In J. J. HuBBARD (Ed.), The peripheral nervous system.
New York: Plenum Press, pp. 455-485.
8 Progressive Relaxation
Then and Now
Does Change Always Mean Progress?
PAuL M. LEHRER, RoBERT L. WooLFOLK,

AND NINA GoLDMAN
I. INTRODUCTION
Edmund Jacobson died on January 7, 1983, at the age of 94. A psy-

chologist and physician by training, he made important early contri-
butions to the fields of psychophysiology, psychosomatics, and bio-
electronics. He devised one of the most widely used self-control
techniques in the overlapping fields of behavior therapy, behavioral
medicine, and self-regulation: progressive relaxation. This chapter was
stimulated by reflecting on Jacobson's contributions and evaluating
some of his ideas and empirical contributions in light of current concerns
and controversies in the various fields in which he worked. In it we
shall j::Ontrast Jacobson's original progressive relaxation technique with
some of the "revised" progressive relaxation techniques that have been
developed over the years. We shall argue that some of these methods
are so fundamentally different from Jacobson's that they require entirely
different rationales, and we shall hypothesize that they have very dif-
ferent effects.
Most of the revised progressive relaxation techniques have been
devised by individuals who identify themselves as "behavior therapists"
(e.g., Bernstein & Borkovec, 1973; Paul, 1966; Wolpe & Lazarus, 1966).
It is probable that the changes were prompted, at least in part, by the
populations with which the behavior therapists were working. Thus,
when the first author once approached Jacobson to take part in a sym-
PAUL M. LEHRER e UMDNJ-Rutger' s Medical School, Piscataway, New Jersey 08854. RoB-
ERT L. WooLFOLK AND NINA GoLDMAN e Rutgers-The State University, New Brunswick,
New Jersey 08903.
183
184 PAuL M. LEHRER, RoBERT L. WooLFOLK, AND NINA GoLDMAN
posium on relaxation treatment of anxiety, Jacobson declined, saying

that he considered himself to 'be an internist, not a psychiatrist. The
early behavior therapists, on the other hand, worked primarily in the
fields of psychiatry and clinical psychology and tended to concern them-
selves less with the field of psychosomatics than with problems such as
phobias, behavioral disorders, and social skills problems. Consistent
with this, we shall hypothesize that Jacobson's original method is more
appropriate for use with persons suffering from stress-related somatic
disorders, whereas the revised techniques may be more appropriate in
treating those whose problems might be described as purely emotional.
The innovations in technique by the early behavior therapists may have
been prompted by the needs of their particular clients, and the advan-
tages brought by these innovations may not generalize to all persons in
need of relaxation therapy. This consideration is particularly important
in light of changes that have occurred recently in the field of behavior
therapy. Many behavior therapists are now working with nonpsychiatric
physical disorders in the relatively recent field of behavioral medicine,
thus blurring the distinction between Jacobson's applications of pro-
gressive relaxation and their own.
Although most behavior therapists working in the area of behavioral
medicine are still trained in the revised technique, we predict that they
would have better clinical success if they switched to Jacobson's original
technique. This may also be true for therapists who treat anxiety dis-
orders in which somatic symptoms predominate (e.g., hyperventilation,
panic attacks). Despite Jacobson's disclaimer about being an internist
rather than a psychiatrist, many of his published cases are of persons
suffering from various anxiety disorders Oacobson, 1938, 1964, 1970;
McGuigan, 1984); and one of his major clinical books includes the word
anxiety in its title Oacobson, 1964). The book contains numerous studies
of anxious individuals whose symptoms were markedly improved by
progressive relaxation. Conversely, we also hypothesize that adherents
of Jacobson's original method might improve upon their technique
if, for certain patients, they adopted some components of the revised
techniques, particularly when cognitive or behavioral symptoms
predominate.
Our views have been influenced by the currently widely held theory
that somatic, cognitive, and behavioral symptoms of tension are, in part,
independent of each other and are most efficiently treated by modality-
specific therapies-that is, cognitive symptoms by cognitively oriented
techniques, somatic symptoms by somatically oriented techniques, and
behavioral symptoms by behaviorally oriented techniques (Davidson &
PROGRESSIVE RELAXATION THEN AND Now 185
Schwartz, 1976; Lang, 1971; Lazarus, 1981). 1 This contrasts with Jacob-
son's theory that progressive relaxation (a somatic therapy) is the path
of choice for all tension-related problems; with the theories of the cog-
nitive and dynamic therapists that cognitive-insight therapies are best
for all such problems; and with the theory of Benson (1975) and his
followers who assert that all relaxation methods (if not all stress-reduc-
tion methods in general) are equivalent and that therefore the fastest
and simplest methods are universally preferable.
Let us return to the differences between Jacobson's technique and
the various modified progressive relaxation techniques used by most
behavior therapists. We shall contrast them on the dimensions upon
which we believe they differ fundamentally, justifying our belief that
they are, in fact, different techniques, containing different procedures
and having different effects (Woolfolk & Lehrer, 1984a). The dimensions
are as follows:
1. Pedagogy versus psychotechnology. We see Jacobson's technique as
being oriented more toward teaching a life skill, and the revised tech-
niques more as requiring something to be done to the client in order to
produce relaxation, either through a psychological technique or elec-
tronic gadgetry.
2. Emphasis on muscular learning versus on direct modification of cog-
nitions. Jacobson taught progressive relaxation entirely as a muscular
skill. Although dramatic cognitive changes would often occur, Jacobson
argued that cognitive and behavioral changes would occur automati-
cally, as individuals reduced their levels of neuromuscular tension,
thereby having more energy and a better ability to think clearly about
what is rational and what is in their own best interests. The behavior
therapists, on the other hand, give specific behavioral and cognitive
prescriptions to their patients and attempt to modify directly that which
Jacobson saw only as indirect targets of therapy. In Jacobson's mind,
the most direct goal of therapy was to teach self-control of muscle ten-
sion, and to enable his patients to approach a level of absolutely no
tension in the skeletal muscles. Then, he said Oacobson, 1970, xii), they
1 The reader should note that this theory does not postulate complete independence of
the three psychobiological systems (cognitive, somatic, and behavioral) in the etiology
and/or treatment of stress. The systems undeniably interact with and affect each other,
and how they interact with each other is an important avenue of current research and
theory (Lang, 1977; Sheehan, 1983; Tyrer, 1976). We are merely asserting that the three
modalities are partially independent, and, although symptoms in each of these areas may
be affected to some extent by almost any form of treatment, nevertheless each problem
will respond best to modality-specific treatment.
186 PAUL M. LEHRER, RoBERT L. WooLFOLK, AND NINA GoLDMAN
would become free and energetic enough to see new alternatives in their
lives and to change their cognitions and behaviors accordingly.
3. The use of suggestion. This category actually overlaps the other
two. The behavior therapists use suggestion during progressive relax-
ation instructions both as a method of producing relaxation automati-
cally (i.e., the psychotechnology approach) and as a method of enlisting
the mental and cognitive faculties in the relaxation process. Jacobson
rejected suggestion as a vehicle for relaxation training, because he felt
that it interfered with the fundamental task of learning exquisite control
of skeletal muscle activity. With suggestion, he cautioned, a person
might believe relaxation to be present even when it is not.
The relative merits of the various progressive relaxation techniques
can of course be decided only by empirical investigation. To date, the
empirical data on this issue are sparse. Two studies have directly com-
pared Jacobson's original technique with various of the revised tech-
niques, and both of them contain substantial design problems. Turner
(1978) compared Jacobson's technique with various other procedures,
including the progressive relaxation techniques described by Paul, 1966,
and by Wolpe and Lazarus, 1966, EMG biofeedback, and various control
methods. Turner found a small but significant advantage for Jacobson's
technique on one paper-and-pencil measure of anxiety and advantages
of borderline significance on other paper-and-pencil and physiological
measures. A similar study was done by Snow (1977), who found no
significant differences between Jacobson's method and other methods.
The major problem in both studies is that the authors used normal col-
lege students as subjects rather than clinical populations and therefore
the clinical implications are at best inferential. Indeed, there is some
evidence that the physiological effects of learning a relaxation technique
may not be noticeable among a nonanxious population because of a
"floor" effect and that therefore even substantial differences in the
power of the various techniques might have been obscured in the above
studies on college students. To illustrate this problem, Lehrer (1978)
found marked physiological differences between progressive relaxation
and a no-treatment control among anxiety neurotics, but almost no phys-
iological differences among normal subjects. The results were inter-
preted as indicating that nonanxious individuals are able to relax to a
great extent without special training, even under conditions of mild
stress, thus obscuring the effects of progressive relaxation training when
it is compared to a no-treatment or self-relaxation control. In a more
dramatic illustration of this, Lehrer (1972) administered a series of grad-
ually more intense electric shocks to paid college student volunteers in
order to test the effect of muscular relaxation on habituation of the skin
potential response elicited by shock. Although subjects in the relaxation
PRoGRESSIVE RELAXATION THEN AND Now 187
group did habituate more quickly than subjects in a group that had been
instructed to tense their muscles and a group that had been instructed
to hold their levels of muscle tension constant, there were no differences
in habituation rate between subjects in the relaxation group and subjects
in a group that was given no instruction at all. Indeed, in the latter
group, the experimenter had some difficulty in keeping subjects awake
during the entire procedure. One can easily conclude from this study
that normal, nonanxious college students usually have no difficulty re-
laxing when they are reclining in a comfortable chair in a dimly lit room
and are told to relax and that studies of subtle but potentially important
differences in relaxation training technique must be done on a symp-
tomatic population.
Thus studies involving direct contrasts between Jacobson's pro-
gressive relaxation technique and the revised techniques have until'now
bee.J;l inconclusive. Perhaps future investigation will, some day, tell us
whether the differences are clinically significant. We offer the following
conceptual analysis as a preliminary step in the design and execution
of this research.
II. PEDAGOGY VERSUS TECHNOLOGY
Often one hears the verb to relax used transitively, meaning some-
thing that a therapist does to someone as in the sentence, "The therapist
first relaxed the client and then began the desensitization process." In
conjunction with this usage the phrase "relaxation exercises" is often
used, meaning the method by which the therapist "relaxes" someone.
These phrases are alien to Jacobson's method but are not infrequently
heard from practitioners of the revised progressive relaxation proce-
dures. According to Jacobson's philosophy, the therapist should do
nothing to the client other than teach him or her a muscular skill. If the
client practices the skill dutifully, he or she will voluntarily be able to
relax the skeletal muscles to an extraordinary degree and thereby ex-
perience various beneficial physiological and psychological effects. If the
client chooses not to use the skill, no such effects will be felt. Indeed,
having the client feel relaxed during a training session is, at most, a
secondary goal, useful only as a motivator for further practice. Jacobson
considered countertherapeutic any attempt to "prompt" clients to feel
relaxed in any way other than by learning to recognize very low levels
of skeletal muscle tension and by learning to "'switch off" the tension
at will. He argued that any attempt to make a person feel relaxed in the
absence of actual muscle relaxation would make the client dependent
on the therapist for the beneficial feelings and would interfere with
motivation to learn the relaxation technique. Thus, after having exper-

imented with them in his own practice, he eschewed the use of such
aids as tape recorders and biofeedback instrumentation because, he
found, they created dependence on an external machine rather than on
internal proprioception and muscle skill.
A. Number of Sessions
Jacobson's method is often characterized as unworkable because of
the large number of sessions needed to teach the technique. Indeed, the
number of sessions required by Jacobson's method is the reason almost
universally cited by those who opt for the adaptations of Jacobson's
technique rather than the original. It is true that various treatment man-
uals written by Jacobson (1964, 1970) do indicate that the technique can
only be taught in upwards of twenty sessions. An examination of Ja-
cobson's own cases, however, indicates that he himself often did not
adhere to these guidelines. In some instances he reported therapeutic
effects in as few as two hours of treatment Gacobson, 1964, pp. 200-
201). Although he believed that thorough training is the preferred treat-
ment for everyone, he understood that, for many practitioners, thor-
oughness must be sacrificed for clinical expediency. Thus, although on
the average most practitioners of the revised procedures may spend less
time than Jacobson did in training an individual to relax, this is not a
necessary difference between the two approaches.
The research literature is unclear about the importance of lengthy
training. In their literature review, Borkovec and Sides (1979) concluded
that multiple sessions of training were necessary to insure that pro-
gressive relaxation training has a measurable physiological effect and
they found a significantly greater number of sessions devoted to training
among studies that found measurable physiological effects as opposed
to those that did not. On the other hand, Lehrer (1982) reviewed a
voluminous amount of more recent literature and did not find significant
differences on this measure, thus casting in doubt the importance of
length of training. Even in the review by Borkovec and Sides, however,
the number of training sessions in the "effective" studies was not large
(with a mean of less than 5).
B. One Muscle at a Time versus All at Once
A characteristic that does differentiate the two techniques quite

clearly is the number of muscle groups that are given training in each
session. Jacobson tended to restrict each session to training of a single

muscle group, or at most to the muscles of a single area of the body
(viz., the arms, or the legs, the trunk, the neck, the speech region, or
the eye region). Virtually all of the revised procedures involve asking
the client to tense and release muscles throughout the body in each
session. The difference stems from the contrasting goals of the two tech-
niques: teaching a muscle skill as opposed to having the individual ex-
perience relaxation during each training session. There have been no
systematic empirical studies of this component of the progressive re-
laxation technique. Our own clinical experience leads us to predict that
focusing on fewer muscle groups per session should produce a greater
ability to decrease muscle tension in the particular muscle group(s),
whereas presenting training in all muscle groups might lead to a greater
subjective sense of relaxation, particularly during early training sessions.
One recent empirical study bears on this issue. Sime and DeGood (1977)
found that multiple sessions of taped progressive relaxation training of
the frontalis muscle alone did produce significant decreases in frontalis
EMG. Of the many studies reviewed by Lehrer (1982) of taped pro-
gressive relaxation training, this was one of only two that showed a
significant physiological effect for this mode of treatment compared with
a control group. In the other study (Parker, Gilbert, & Thoreson, 1978)
significant results were obtained because of increased arousal in the con-
trol group. We take the Sime and DeGood study as supporting the theory
that repeated focus on an isolated muscle does produce greater ability
to relax specific muscles than does simultaneous training in all muscles
(which had been offered to subjects in virtually all of the other studies
of taped relaxation that Lehrer reviewed).
C. Live versus Taped Relaxation
As just mentioned, Borkovec and Sides (1979) and Lehrer (1982)

have both concluded, after extensive literature reviews, that taped pro-
gressive relaxation training is ineffective as a method of training people
to control physiological activity. There was a subtle difference in the
conclusions reached in the two reviews, however. By the time of Lehrer's
more recent review, a number of studies had appeared showing, in fact,
a physiological effect for taped relaxation training, although the pro-
portion of studies showing at least one significant physiological effect
was still greater with live training than with taped training. The supe-
riority of live training was most obvious when g~!J..fiJli{'{tE}_Q_'} of the relax-
ation skill outside the training session was examined. The generaliza-
bility of relaxation is, of course, critical for its clinical effectiveness.
Nevertheless, a number of studies have found positive effects of taped

progressive relaxation training on self-report measures (e.g., Borkovec,
Grayson, & Cooper, 1978; Carrington et. al. 1980; Edelman, 1970; Gilbert,
Parker, & Oairbom, 1978; Haynes, Griffen, Mooney, & Parise, 1975;
Russell, Sipich, & Knipe, 1976; Sheridan, Vaughan, Wallerstedt, &
Ward, 1977; Townsend, House, & Addario, 1975). It is thus possible
that taped training may have important therapeutic use in treating con-
ditions for which self-report or, perhaps, overt behavior is clinically more
relevant than is physiological activity. It is also possible that self-report
measures are more influenced by placebo or expectancy effects than are
physiological measures.
D. Conditioned Relaxation
The induction of relaxation by classical conditioning methods is for-
eign to Jacobson's method but often used in the revised procedures.
This is another form of "inducing" relaxation by external means rather
than teaching it as a skill, and Jacobson eschewed all such approaches.
Included among techniques for inducing conditioned relaxation are im-
agining a pleasant scene and "cue-controlled relaxation," a procedure
in which the individual learns to say a silent cue word (e.g., "relax")
while in the relaxed state, and to use it later when under stress to induce
the relaxed state.
Proponents of relaxing imagery have theorized that pleasant and
calm feelings are evoked by the associations produced by the relaxing
images. Jacobson, on the other hand, theorized that the process of de-
liberately thinking about a pleasant scene necessarily involves increased
tension in the muscles in and around the eyes because, he found, visual
imagery is always accompanied by such increases in muscle tension
(Jacobson, 1931). By its very nature, according to Jacobson, relaxation
is a completely passive experience. 2 Arguing in favor of the use of re-
laxing imagery, Lazarus (1981) has proposed that it has its greatest ben-
eficial effects on problems involving anxiety-producing imagery. Jacob-
son observed that when complete muscle relaxation is achieved,
particularly involving the muscles of the eyes, no imagery takes place.
He thus felt that deliberate use of relaxing imagery is counterproductive.
No empirical comparisons have yet been reported between Lazarus's
2 In some ways Jacobson's view of relaxation is similar to descriptions of the deeper stages
of Zen or yogic meditation (Patel, 1984; Suzuki, 1979). Actually, however, Jacobson dis-
liked this comparison, feeling that his method was entirely scientific and not at all myst-
ical, as are the Eastern meditative disciplines.
"modality-specific" argument and Jacobson's argument for sticking to

muscle reeducation.
A number of studies have appeared evaluating cue-controlled re-
laxation and comparing it to progressive relaxation. Grimm (1980) ex-
tensively reviewed the empirical literature on cue-controlled relaxation
and concluded that the method had no advantages over progressive
relaxation. Indeed, there is some evidence that progressive relaxation
training alone has superior effects (e.g., Holstead, 1978). Although Ja-
cobson never wrote about this technique, it is probable that he would
have disapproved of it for reasons similar to those used against relaxing
imagery. Imagining a word, he found, necessarily involves tension in
the speech area and, being a deliberate muscular act, it necessarily de-
creases relaxation of the musculature. This reasoning could explain the
otherwise seemingly paradoxical findings that one component of a re-
laxation technique (i.e., progressive muscle relaxation, which is a com-
ponent of cue-controlled relaxation) might be more powerful than the
combined technique. Another possible explanation for this anomaly is
that the time demands involved in cue training necessarily detract from
the intensity of the progressive relaxation training given during the early
stages of this technique, thus diluting its possible beneficial effects.
E. The "Pendulum Effect"

Jacobson's method differs starkly from the revised progressive re-
laxation methods in its use of deliberate tension and in the theoretical
justification for such tension. In Jacobson's method, deliberate tensing
of muscles was done during training sessions in order to familiarize
people with the sensations of muscle tension and to illustrate how these
sensations are produced by doing things with particular muscles. Jacob-
son advised taking great pains to differentiate the feelings of active mus-
cle flexion with other sensations in the muscles and tendons that do not
directly reflect active "doing," such as the sensations of passive muscle
stretching that occurs when an opposing muscle is tensed and a limb
moved, and the sensations in tendons, joints, and ligaments that may
occur when a nearby muscle is flexed. The sensations of muscle flexion
are extraordinarily subtle, often considerably more so than the sensa-
tions of passive stretching or joint sensations, and people who are not
very sensitive to muscle sensations often have great difficulty feeling
them. Jacobson insisted that these particular sensations be recognized
at extraordinarily low levels of flexion, because he found that in many
cases of anxiety and tension disorders the degree of muscle tension in
the body is nowhere near the level of maximal tension that a person can
produce. Although Jacobson did advise his patients to tense each muscle
quite severely when first becoming acquainted with its function and with
the sensations associated with its flexion, he did not advise extreme
tension when attempting to relax. Extreme tension was used only di-
dactically, when first pointing out the things that individuals must refrain
from doing in each muscle when they want to relax. Thereafter Jacobson
advised using the "method of diminishing tensions" (Jacobson, 1938,
p. 53), in which the trainee successively tenses as little as possible while
still remaining aware of the "control" sensations (i.e., the sensations of
tension that tell us when we are exerting control over the actions of part
of the body, by use of the particular muscle). In each training session,
the individual tenses a muscle only a few times and spends most of the
time "switching off" the muscle-but each incidence of tension should
require progressively less muscle contraction to produce readily per-
ceived control sensations. The goal of training is to enable the individual
to develop such exquisite sensitivity to, and control over, the skeletal
muscles that he or she can switch off each muscle at any moment, with-
out tensing at all.
In contrast to Jacobson's method, some of the revised progressive
relaxation procedures unabashedly advise tensing muscles as a method
of inducing relaxation (cf. Bernstein & Borkovec, 1973, p. 20). Individuals
are often asked to tense several muscle groups simultaneously, with
maximal effort, and then to release them, each time the individual
wishes to relax. In the latter case, the sudden release of extreme tension
is theorized to produce a decrease in baseline muscle tension, much as
a pendulum swings back in the opposite direction past its resting point
when it has been lifted and then released.
Jacobson's research specifically argued against the existence of a
pendulum effect. He found (Jacobson, 1934, a, b) that in many individuals
tension remains high, often for prolonged periods, after release from high
levels of deliberate tension. Individuals who have received training in
progressive relaxation tend to be able to return to low levels much more
quickly than others; and athletes (who presumably devote considerable
time and concentration to their muscles and their muscle sensations,
and therefore have developed greater sensitivity and control than most
people) show a post-tension release that is intermediate between the
levels achieved by other individuals who are trained and those untrained
in progressive relaxation (Jacobson, 1934b).
Jacobson's results would actually suggest that an egg timer would
be a better analogy than a pendulum in describing the effect of tension.
When the sand inside is moved to the top, it reaches bottom rather slowly
and gradually, and much of it may remain elevatecl for considerable
time. With athletes and those trained in relaxation, the sand in the timer
may flow more swiftly, and with untrained individuals suffering from
neuromusclular tension the sand may flow more slowly, but in no in-
stance does it reach levels below those at which it had started before
the timer was turned over.
We have recently reported the results of a study on this issue from
our laboratory. We performed it because, upon reviewing Jacobson's
studies of EMG levels following tension release, we noticed that Jacob-
son did not use tension procedures that are analogous to those used in
the revised progressive relaxation procedures. In the latter, tension takes
place only for 10-20 seconds, and tension-release cycles are repeated
five or more times; whereas in Jacobson's experiments subjects tensed
their muscles for one minute or more, and the tension took place only
once. In our study (Lehrer, Batey, Woolfolk, Remde, & Garlick, 1985),
we attempted to replicate more closely the conditions used in the revised
progressive relaxation procedures. We repeatedly asked subjects to
tense for 20 seconds and then to release their frontalis and dominant
forearem extensor muscles, alternately, and in balanced order. Simul-
taneously, we measured EMG levels from both sites and asked subjects
to rate their levels of tension in the muscles that they were tensing at
the time. We found little evidence for a pendulum effect. In both muscle
groups, EMG levels remained elevated for a minute or more after sub-
jects were told to release their tension. Although frontal EMG levels did
eventually reach levels lower than baseline after tension release, forearm
EMG levels appeared to have remained elevated. Self-reported feelings
of tension appeared to dissipate relatively slowly in both muscle groups.
An additional objection to the tense-release method of training used
in the revised methods was raised by Sime (1982), who pointed out that
most of these methods involve tensing several muscle groups at once
(e.g., the entire fist and arm at once) and therefore produce so many
sensations that the trainee cannot possibly isolate the subtle control sen-
sations. For example, tightening the fist and arm together produces mul-
tiple sensations from tendons and ligaments (the feelings of nails digging
into the palms, etc.), all superimposed upon each other in a way that
makes them almost impossible to discriminate from each other. In most
cases, this completely masks the sensations of muscle contraction, which
often are much more subtle. If any relaxation is achieved by this method,
it certainly would not be by training individuals to make subtle muscle
discriminations. Beneficial effects from this procedure, if they occur at
all, probably result more from suggestion and expectation, issues that
will be discussed in detail below.
Note that, although we call into doubt the procedure of inducing
relaxation through tense-release cycles, we are pointedly not denying
the usefulness of tensing and releasing muscles during the process of
learning to discriminate tension. Indeed, in the treatment of insomnia

Borkovec, Kaloupek, and Slama (1975) found that relaxation using ten-
sion-release cycles is more useful than simple suggestions to relax (with-
out any deliberate tension, but therefore also without any instruction
in recognizing muscular tension). Similarly, the only study that found
EMG biofeedback to be more powerful than live progressive relaxation
as a method of reducing EMG activity utilized progressive relaxation
instructions which consisted only of suggestions to relax each muscle
and descriptions of the feelings that should be obtained-without tense-
release instructions (Canter, Kondo, & Knott, 1975). These findings sup-
port our opinion that tension and release do serve an important role in
progressive relaxation training, even if tension-release cycles are not
appropriate as methods of inducing relaxation. In our view, as in Jacob-
son's, their proper use is as a training vehicle, not as a form of relaxation
technology designed to produce automatic results.
F. The Use of Biofeedback

Actually, Jacobson was one of the first to experiment with EMG
biofeedback as an aid to teaching people to relax. According to Mc-
Guigan (1984), he rejected the method in the 1950s because he felt it
tended to make people dependent on the machines. Jacobson's goal was
to teach people to recognize tension in their own muscles. Reliance on
an external machine, he felt, was counterproductive.
Parametric studies of relaxation and biofeedback have found that,
although not harmful (as predicted by Jacobson), EMG biofeedback is
generally not so powerful as progressive relaxation and seems, on the
average, to add little to Jacobson's method. In a review of the literature
on progressive relaxation and EMG biofeedback, Lehrer (1982) con-
cluded that EMG biofeedback does not add incremental effectiveness
when combined with live relaxation instructions and that, when the two
techniques are compared with each other, live relaxation instructions
are, on the average, superior. More recently, Janssen (1983) found the
combination of live relaxation training and biofeedback to be more po-
tent than biofeedback alone but three other recent studies (Nicassio,
Boylan, & McCabe, 1982; Pollard & Ashton, 1982; Wilson & Bird, 1981)
found no differences between live progressive relaxation training and
biofeedback.
When compared and/or combined with taped relaxation instruction,
the picture is quite different. On the average, EMG biofeedback tends
to be more powerful than taped progressive relaxation instruction (d.
reviews by Lehrer, 1982; Lehrer & Woolfolk, 1984; and recent studies
by Bennick, Holst, & Renthem, 1982; Daly, Donn, Galliher, & Zimmer-
man, 1983; Stout, Thornton, & Russell, 1980; Walsh, Dale, & Anderson,
1977). Also, although the combination of taped relaxation instructions
and biofeedback does yield superior results to taped relaxation instruc-
tions alone (d. review by Lehrer, 1982), taped relaxation training prob-
ably does not add anything to biofeedback. Two recent studies (Kappes,
1983; Walsh, Dale, & Anderson, 1977) found no differences between
biofeedback alone and the combination of biofeedback and taped relax-
ation instructions.
Nevertheless, EMG bjofeedback might still be useful to some in-
dividuals. Prager-Decker (1978) found that individuals who score high
on a paper-and-pencil measure of external control tend to do better with
EMG biofeedback than do other individuals. Two recent studies from
Blanchard's laboratory (Blanchard et. al., 1982; Neff, Blanchard, & An-
drasik, 1983) also suggest that some individuals who cannot be helped
by progressive relaxation might be helped by biofeedback. They found
that administering biofeedback produced significant improvements in
symptoms among individuals suffering from headaches who did not
respond to prior relaxation instruction. Several investigators have ar-
gued that biofeedback might be a useless technology because of the high
cost of equipment and because it does not appear to be more powerful
than relaxation therapy (d. Furedy, 1984a,b; White & Tursky, 1982).
However, Lehrer (1983, 1984a,b) has argued that the price of biofeedback
equipment had declined significantly enough to warrant consideration
as an inexpensive home aid to relaxation training, or, at times, even a
substitute-especially when the alternative is superficial or taped
instruction.
III. DIFFERING VIEws oN CoGNITIVE AND SoMATIC FAcToRs IN

EMOTION AND IN TREATMENT
In both his clinical approach and in his theoretical view of the nature
of mind and body, Jacobson emphasized the importance of the periphery
of the body. Emotions and other maladaptive mental processes, he felt,
might most effectively be modified by changing a person's use of the
skeletal muscles. The post-Jacobson relaxation therapists have tended
to give more emphasis to direct modification of cognitions; indeed, the
whole field of behavior therapy in recent years has emphasized the
importance of cognitive mediation of behavior and of direct treatment
of cognitions. Even such a bedrock "somatic" technique as biofeedback
has come to be interpreted as being mediated by cognitions and having
profound effects upon cognitions (Lazarus, 1977; Holroyd, 1979). One
practical corollary of this theoretical direction is the devotion of a rela-

tively greater amount of therapy time to cognitive interventions and
rather less to physical relaxation strategies in modern behavior therapy.
Jacobson's theory of emotion is quite similar to that of Gellhorn.
Indeed, the two men had great influence on each other. Gellhorn (1958;
Gellhorn & Loofbourrow, 1963, Chapter 17) devoted considerable at-
tention to devising a theory explaining the effectiveness of Jacobson's
techniques; Jacobson reciprocated by adopting, in large part, Gellhorn's
theory Oacobson, 1967, pp. 149-158). Following Hess (1925), Gellhorn
postulated two emotional systems: the ergotrophic and the tropho-
trophic. These systems are often mutually inhibitory, with ergotrophic
activation associated with sympathetic reactivity and the trophotrophic
response associated with parasympathetic reactivity. They also, how-
ever, work homeostatically, such that overactivation in one system can
produce increased reactivity in the opposing system. This can occur
through reflexes elicited by blood pressure fluctuations, mediated
through baroreceptors in the carotid sinus and aortic arch, and often
modulated by cerebral (chiefly hypothalamic) and neurohumoral mech-
anisms (Gellhorn, 1957; Lacey, 1967). According to Gellhorn (1958), pro-
gressive relaxation decreases ergotrophic dominance directly by de-
creasing proprioception from the striated musculature. He notes that a
high proportion of the innervation of the ascending reticular system
emanates from skeletal muscle proprioceptors, so that diminished mus-
cular tension would be expected to produce diminished autonomic and
emotional reactivity through decreased stimulation of the posterior hy-
pothalamus and cortex from the ascending reticular system. He sup-
ported this theory by citing evidence that, in cats, administration of
curare (which was thought to block muscular activity at the neuromus-
cular juncture) produced decreased sympathetic reactivity and slow-
wave EEGs similar to those observed in a state of somnolence.
Jacobson believed that anxiety and anxious ideation-indeed, any
ideation-are impossible when the skeletal musculature is completely
relaxed Oacobson, 1938, pp. 164-189). In his final book, The Human Mind,
Jacobson (1982) argued that thought depends upon proprioception, such
that the perception of any thought or feeling would necessarily differ
among any organisms that differed physically in the periphery of the
body as well as in the brain.
Davison (1966) and Campbell, Sanderson, and Laverty (1964) ar-
gued against this peripheralist theory of emotion. They criticized Gel-
lhorn's reliance upon animal studies of curare in order to justify his
assertion that muscle relaxation produces relaxation and somnolence,
and they cited some human curare studies indicating that people whose
muscles are experimentally rendered flaccid by curare can nevertheless
remain extremely anxious. McGuigan (1978), in justifying Jacobson's

position, however, noted that the humans who allowed themselves to
be experimentally curarized were never as deeply curarized as the an-
imals in the studies cited by Gellhom.
Following Davison's earlier conclusions, Davidson and Schwartz
(1976) hypothesized that the positive affective state generated by pro-
gressive relaxation and the inhibitory effect of diminished efferent motor
activity are the elements in progressive relaxation therapy responsible
for its anxiolytic effect. By emphasizing these central effects of relaxation
therapy, Davison established the theoretical rationale for the develop-
ment of cognitive behavior therapy. If relaxation therapy for anxiety has
primarily central effects, then it would be quite natural to presume that
other procedures affecting the fundamental operation of the central ner-
vous system (e.g., direct modification of cognition) might be expected
to have effects that are just as profound as muscular relaxation or even
more so.
Although peripheralist theories of emotion have fallen into disre-
pute, there is a large body of evidence indicating the importance of
peripheral events in the generation of emotion. It has been shown that
distension of the carotid sinus affects electrocortical activity (Bonvallet,
Dell, & Hiebel, 1954). Lacey (1967) argued that these data suggest a
mechanism whereby changes in blood pressure might have a direct ef-
fect on cortical activity. A study of individuals with spinal cord injuries
(Hohmann, 1966) has suggested that the loss of autonomic feedback is
associated with a lessened capacity to experience emotion. Tomkins
(1970) has hypothesized that feedback from the facial musculature to
subcortical centers may cause the latter to influence emotional experi-
ence. Some empirical support for this hypothesis was generated in a
number of studies (Kraut, 1982; Laird, 1974; Lanzetta, Cartwright-Smith,
& Kleck, 1976).
In ascertaining whether a more cognitive approach to relaxation has
effects that are superior or inferior to those of Jacobson's progressive
relaxation-or simply different kinds of effects-it is instructive to ex-
amine the literature on the effects of cognitive therapy. All forms of
cognitive therapy have in common the attempt to restructure or modify
the client's thoughts, beliefs, or imagery so as to alleviate emotional
distress. Although there is some overlap among the different forms of
cognitive therapy, each has its distinctive emphasis. Meichenbaum' s
(1977) method tends to emphasize the systematic alteration of clients'
self-verbalizations and the replacement of worry and anxiety-generating
self-talk by salutary thinking. Beck's (1976, 1984) approach stresses the
identification and alteration of faulty thinking and the cognitive styles
that produce anxiety and arousal. This occurs through a "Socratic" dia-
logue between therapist and patient. Ellis's (1962) approach is a didactic,

persuasive effort directed toward the modification of the content of dys-
functional cognitions. It aims to identify and to eradicate "irrational"
beliefs and attitudes (especiallly those having to do with moralism and
perfectionism) and to replace them with a more "rational" (in Ellisonian
terms) outlook on life.
Cognitive techniques often are applied in conjunction with somatic
relaxation methods and within a specified sequential program of ther-
apeutic activities. Stress inoculation training Garemko, 1979; Meichen-
baum & Cameron, 1973) is one such cognitive-behavioral package. Its
three stages are described at (1) client preparation, (2) skills training,
and (3) application training. In the phase of client preparation the patient
is taught the relationship between maladaptive thinking and behavior.
The client learns a cognitive-appraisal model of stress (cf. Lazarus, 1966)
and a rationale that explains the efficacy of cognitive procedures. In the
second stage, the client learns and rehearses a variety of skills including
cognitive restructuring, coping imagery, modification of self-verbaliza-
tion, and physical relaxation (usually a version of Jacobson's technique).
In the third stage, the client practices and applies the newly learned
skills for coping with stress by confronting a graded sequence of stres-
sors, in one or a combination of several ways: imaginally presented, by
role playing with the therapist or in vivo.
Other forms of combined cognitive and somatic interventions are
those related to the widely used technique of systematic desensitization
(Wolpe, 1958). This involves slowly exposing oneself to phobic stimuli,
either in vivo or in imagination, while in a deeply relaxed state. This
technique is widely used in the treatment of phobias. Although Wolpe
described it as a form of classical conditioning, Wilkins (1971) and Locke
(1971) argued that cognitive mechanisms were probably more important
than conditioning in producing therapeutic success with the technique.
Although this view has not been universally accepted (e.g., Davison &
Wilson, 1972), it is unquestionable that some cognitive processes are
directly involved in the desensitization technique (e.g., the requirement
that the client must build a hierarchy of anxiety-provoking scenes, thus
making cognitive judgments about the amount of anxiety that is aroused
by each scene). When it is done as described originally by Wolpe, this
technique thus involves both somatic (relaxation) and cognitive com-
ponent. A similar technique is stress management training, which is a
combination of visuomotor behavior rehearsal and a technique called
anxiety management training. Anxiety management training takes place
in three phases. In the first, clients are given brief training in one of the
modified progressive relaxation techniques. During the second phase,
they imagine anxiety-producing scenes, in a manner analogous to sys-
tematic desensitization, and they are taught to identify the specific phys-
iological or muscular sensations associated with the anxiety they per-
ceive. In the third phase, they learn to utilize a cue word, as is done in
the technique of cue-controlled relaxation, and to use it to induce a
relaxed state when confronted with an anxiety-provoking situation. In
the visuomotor rehearsal phase of stress management training, clients
imagine themselves engaging in behaviors that are alternatives to those
that generate life stress. Following imaginal rehearsal, these behaviors
are practiced in vivo.
Cognitive therapies have been proven to be effective treatments for
a variety of disorders, including depression, asthma, anxiety, and head-
aches. For most of these dysfunctions, at least at their present level of
conceptualization and measurement, the self-report of distress is the
most important and even the defining element of the difficulty. For other
syndromes (e.g., hypertension and other somatic complaints) we have
little evidence supporting the efficacy of cognitive therapy. The literature
to date on cognitive approaches to physiological problems has been
sparse. Few studies have evaluated the impact of cognitive interventions
on somatic systems. The available literature does suggest that cognitive
treatments may carry less impact on behavioral and physiological re-
sponse systems than do treatments that target these systems more di-
rectly. A study by Kremsdorf, Kochanowicz, and Costell (1981) found
that cognitive therapy reduces self-reported headache activity but does
not reduce actual muscle tension. Conversely, they found, EMG bio-
feedback appears to reduce muscle tension but not self-reports of head-
ache. Similarly, Biran and Wilson (1981) found that cognitive treatment
is ineffective relative to an exposure-based treatment in ameliorating
avoidance among phobic patients.
Cognitive therapy has most often been found to be effective with
those disorders for which the individual client's phenomenology is
either sufficient to its diagnosis or the most salient indication of its pres-
ence. Holroyd and Andrasik (1982) found some evidence for superiority
of cognitive therapy over EMG biofeedback in the treatment of tension
headache. The effectiveness of cognitive techniques in the treatment of
clinical pain resulting from various disorders has been demonstrated
repeatedly (Turk, Meichenbaum, & Genest, 1983). Cognitive approaches
have also been found effective in the treatment of test anxiety (Mei-
chenbaum, 1972), social anxiety (Goldfried, 1979), depression (Beck,
Rush, Shaw, & Emery, 1979), and chronic anger (Novaco, 1975).
We recently (Lehrer & Woolfolk, 1984) reviewed the results of em-
pirical studies comparing cognitive with somatic treatments for various
stress-related problems, and concluded that cognitive therapy has more
powerful effects than somatic relaxation procedures on cognitive symp-
toms, including cognitive assessments of somatic problems (self-ratings

of pain, insomnia, etc.). A recent comparative study by Woolfolk and
McNulty, 1983, of sleep-onset insomnia found that self-report of sleep
latency showed more improvement at a six-month follow-up among
subjects who had been given visual imagery training (with or without
progressive relaxation) than among those who had been given pro-
gressive relaxation alone. Also, Langosch et al. (1982) studied patients
at a cardiac rehabilitation center and found that, immediately after train-
ing, subjects receiving stress management training showed greater de-
creases in feelings of job responsibility than subjects in the relaxation
group. At a six-month follow-up subjects in the stress management
training group reported less emotional lability and a greater ability to
reduce stress than subjects in the relaxation group, who reported a de-
crease in the latter ability. Although subjects in the relaxation group
showed a greater decrease in cardiac complaints than subjects in the
stress management training and control groups immediately after train-
ing, they also showed the greatest post-test to follow-up increase in this
measure. The authors interpreted the results as indicating the superi-
ority of stress management training, but the interpretation is clouded
by the fact that subjects chose which treatment to receive rather than
having been assigned randomly.
Several other recent studies have found greater effects for combined
relaxation and cognitive therapies than for relaxation therapy alone, al-
though the combination of techniques may not be superior to cognitive
therapy alone when cognitive measures are used to assess outcome.
Osberg (1981) studied speech-anxious subjects and found that practicing
public speaking, with or without concurrent relaxation training, pro-
duced greater decreases in self-rated anxiety than did relaxation training
alone. Furthermore, the combined technique produced greater de-
creases in observer ratings of anxiety than did relaxation training alone.
Although perhaps more of a "behavioral" than a cognitive technique,
the speech practice condition did include some imaginal and cognitive
components, so we classify it as a cognitive technique. Similarly, Miller
and DiPilato (1983) compared systematic desensitization with relaxation
therapy in treatment of nightmares and found, 25 weeks after the be-
ginning of their six-week treatment program, that subjects reported
greater decreases in intensity of nightmares in the desensitization group
than in the relaxation group. At the 15th week after therapy onset, how-
ever, there had been no differences between the two treatments, and
ratings of nightmare frequency did not differ between the two groups
at either time period. Finally, Peal, Handa!, and Gilner (1981) found
greater decreases in two measures of death anxiety after desensitization
than after relaxation among death-anxious undergraduates. This is con-
sistent with a large body of earlier literature indicating the systematic

desensitization is more effective than progressive relaxation alone in
treatment of phobias (cf. Cooke, 1968; Davison, 1968; Lamont & Ed-
wards, 1967). (In these studies self-report and behavioral approach
measures were used almost exclusively to assess outcome.)
Three recent studies found no differences on various cognitive
measures between relaxation therapy and combined relaxation and cog-
nitive therapies. White, Gilner, Handal, and Napoli (1983) found that
systematic desensitization produced the same decreases in death anxiety
among high death-anxious undergraduates as did relaxation training
alone. Similarly, Philips and Hunter (1981) found that the combination
of relaxation therapy and calming images did not differ from relaxation
therapy alone in improving pain behavior and self-rated pain experience
among psychiatric patients suffering from headaches. Finally, Van Has-
sel, Bloom, and Gonzales (1982) studied schizophrenic individuals who
were given either relaxation training or anxiety management training.
They found that both treatments, compared to a control group, pro-
duced improvements in self-ratings of generalized anxiety and in ther-
apist ratings of a number of behaviors related to anxiety, emotion, and
overall psychiatric status.
The only study comparing cognitive therapy and progressive re-
laxation therapy using a physiological measure as an outcome variable
was reported by Kaplan, Metzger, and Jablecki (1983). They studied 40
adult male patients while they were undergoing a rather painful clinical
electrornyographic study in a Veterans Administration hospital and corn-
pared the effects of four interventions: cognitive reappraisal therapy,
prpgressive relaxation, a combined technique called cognitive behavior
modification, and an attention control group. In a post hoc analysis, they
found lower heart rates in the groups that received relaxation training
than in the other two groups. On various self-report measures of emo-
tional distress and on a behavioral rating of stress and discomfort, pa-
tients who received training had lower scores than control subjects, but
no differences were found among the various treatment conditions. Al-
though not directly relevant to progressive relaxation therapy, dramatic
findings from a study comparing cognitive therapy with EMG biofeed-
back deserve mention because they illustrate the somatic-cognitive treat-
ment-symptom specificity so well. Schandler and Dana (1983) studied
students with high scores on paper-and-pencil tests of nervousness,
hostility, and depression. They compared frontalis EMG biofeedback
with taped training in guided imagery and cognitive desensitization
(using relaxing images in place of muscle relaxation in systematic de-
sensitization) and with a self-relaxation control. Only guided imagery
produced reliable therapeutic changes in a wide array of self-report
measures (including anxiety, hostility, and depression), but only bio-

feedback produced reliable decreases in frontalis EMG and heart rate.
Our laboratory, which is one of the few active centers using Ja-
cobson's original relaxation technique, has produced some additional
evidence for the somatic specificity of progressive relaxation therapy.
Lehrer (1978) found that among anxiety neurotics progressive relaxation
training produces decreases in psychophysiological reactivity, but not
necessarily in self-report of anxiety-a measure which may be more
refractory to change in this population. Also, a recent study of relaxation
therapy for asthma found greater effects for relaxation therapy in a so-
matic measure of asthma (bronchial constriction caused by inhalation of
methacholine) than in various self-report measures of asthmatic symp-
toms (Lehrer, Hochron, McCann, Swartzman, & Reba, in press), al-
though in this study various self-report measures of anxiety did show
improvement. It is possible that longer follow-up in both of these studies
might have shown cognitive changes as well, and future research may
prove that the discrepancies we observed resulted from a cognitive lag.
Both of these studies were of populations with marked chronic prob-
lems, and self-perception of improvement in such conditions may not
occur as rapidly as physiological improvement. For the short run, how-
ever, data from our research does suggest that progressive relaxation
has more reliable somatic than cognitive effects.
In summary, there is considerable evidence for somatic specificity
for progressive relaxation and cognitive specificity for cognitive therapy.
Relaxation therapy produces more reliable effects on somatic measures
and cognitive therapy produces more reliable effects on cognitive meas-
ures. Adding cognitive therapy to relaxation therapy tends to strengthen
the cognitive effects of the latter, whereas there is evidence from one
study that adding relaxation therapy to cognitive therapy does not ap-
pear to strengthen such effects. There is as yet no systematic evidence
that these therapies have additive effects on physiological outcome
variables.
IV. SuGGESTION
Edmund Jacobson was unalterably opposed to the notion that sug-

gestion could be used effectively in teaching people to relax. The danger
of suggestion, he said, is that it may make the individual feel that re-
laxation is taking place even when it is not. The perception of relaxation
is not so important as actual physical relaxation, according to Jacobson.
Therefore, suggestion may be deleterious because a person may stop
devoting the time and concentration necessary to learn relaxation if he
or she feels relaxed already. Jacobson went to great lengths to eliminate
suggestion from his technique. In his clinical technique, when asking a

person to tense a particular muscle in order to detect the exact location
of muscle tension sensations, Jacobson would not even tell the trainee
where the sensations should be felt. If the trainee was having difficulty
"finding" the sensation of tension in a particular muscle,Jacobson sim-
ply asked the trainee to prolong or to intensify the tension, or he applied
pressure with his hand to oppose the muscle movement he was asking
the trainee to make, in order to intensify the sensations of tension. If
the trainee discovered the source of the "control sensations" (i.e., the
sensations produced by the muscles when they produce some form of
active movement or control over part of the body) without being told
where to look for them, Jacobson felt more certain that proprioceptive
learning was, in fact, taking place.
In Jacobson's mind, the processes involved in progressive relaxation
are radically different from the processes involved in hypnosis, the quin-
tessential form of suggestion. In his classic work, Progressive Relaxation,
Jacobson (1938, pp. 303-308) enumerated 32 ways in which progressive
relaxation differs from hypnosis. Included are the following most no-
table ways:
1. In the method of relaxation, no technical suggestions are given. For in-
stance, the physician would never suggest, "Now your arm is becoming
limp!" or "This will help you be quiet!" He simply directs the patient in the
same manner as when prescribing diet or exercise.
2. The physician does not hesitate to interrupt the patient at a moment when
he is failing in the attempt to relax and to criticize him vigorously. Such
practice is foreign to all suggestive procedures.
3. Relaxation has to be learned step by step with various details of success
and failure. It is a learning process by the method of trial and error. It requires
the cultivation of the observation skill of the patient, which largely depends
upon practice apart from the physician. Hypnosis does not have to be learned
at all.
4. The therapeutic effects of suggestion and hypnosis, while brilliant, are
likely to be ephemeral. Symptoms disappear to be succeeded by other ones,
for the basic nervous hypertension is not treated. On the other hand, re-
educative treatment by relaxation tends to be relatively thorough-going and
permanent in effect.
5. The individual is generally awakened from hypnosis by some suggestive
signal. Occasionally the awakening proves difficult. In contrast with this,
there is no difficulty and no special signal of arousal from profound relaxation.
If the individual has fallen asleep, he awakes in every respect as from a
natural sleep.
6. In suggestion, the doctor designates the symptoms which are to disappear.
After learning to relax, the patient often reports disappearance of symptoms
never before mentioned.
Jacobson's view of the sharp distinction between skill learning and

suggestion is not universally accepted. Edmonston (1981, p. 210) takes
the relatively extreme position that hypnosis and relaxation are identical.
He says that "traditional" hypnotic procedures all either specifically use

relaxation inductions or have phenomenological and physiological ef-
fects that are very similar to relaxation procedures. He cites studies in-
dicating that direct comparisons between hypnosis and various relax-
ation procedures reveal no important differences between the two
procedures. Where differences are found, he explains them away by
idiosyncratic problems in individual studies.
Even among those who believe that there are important differences
between relaxation and hypnosis, few will doubt that they both involve
a number of similar characteristics. Both involve the focusing of atten-
tion, and both can produce profound effects on thought processes and
levels of physiological arousal. Both can be used therapeutically to help
people to control aspects of themselves that appear, in ordinary circum-
stances, to be uncontrollable. The degree to which these characteristics
and effects are emphasized in the various techniques does differ, how-
ever. Some hypnotic inductions do not use relaxation at all and seem
to depend for their effects almost entirely on the active focus of attention;
some relaxation techniques, such as biofeedback and Jacobson's pro-
gressive relaxation method, use very little direct suggestion of any men-
tal effects. Also, reasonable people can and do differ about the meaning
of various empirical studies comparing relaxation and hypnosis. For
example, despite Edmonston's dismissal of the studies finding differ-
ences between relaxation and hypnosis, we (Lehrer & Woolfolk, 1984)
have interpreted them as valid and important.
Expanding on this view, others have shown that it is almost im-
possible to eliminate suggestion from any form of psychological treat-
ment. Even before an individual arrives at the office, he or she has
already invested the therapist and the therapist's method with the power
to influence and change important aspects of personal functioning. The
decision to seek out a professional and to pay a fee to have oneself
changed already predisposes the individual to change. Frank (1961) has
stated that such influence is a hallmark of all psychotherapies. Orne
(1982, p. 429) has pointed out that the "placebo" effect is present in all
treatments, including drug treatments. Indeed, he could find only one
drug study that evaluated the behavioral effects of drugs independently
of a placebo effect. In virtually all the other studies, he found, the sub-
jects knew that they were taking a drug; therefore these studies actually
evaluate the combination of the drug and the placebo effect, rather than
just the drug effect. Virtually no studies of psychological treatment have
completely eliminated the placebo effect. So it is with Jacobson's own
work; although he tried to minimize suggestion in his technique, he
coul<;i not totally eliminate it. The most ingenious approach to elimi-
nating the placebo effect in relaxation research was devised by Borkovec
and his colleagues (cf. Borkovec & Hennings, 1978), who used "coun-
terdemand" instructions: they told their subjects that no therapeutic
effects might be expected until after the first four sessions, so they could
measure the therapeutic effects of the technique devoid of placebo con-
tribution before the fourth session. This methodology, unfortunately, is
much more adaptable to studying the revised techniques than to study-
ing Jacobson's original technique, because the former present training
in all the muscle groups right from the beginning, whereas Jacobson's
method does not; therefore Jacobson's method may actually have fewer
therapeutic effects during early sessions. Unless a suitable method can
be found to eliminate suggestion even further fromJacobon's technique,
the most we can hope for is comparisons between various relaxation
techniques that differ in the amount of suggestion that is deliberately
included in the technique, compared with the amount of specific phys-
iological training that takes place.
·An approach that comes closer to our own was recently articulated
by Clarke and Jackson (1983), who interpret hypnosis primarily as a
cognitive technique and progressive relaxation primarily as a somatic
technique. In a review of the comparative outcome literature, we con-
cluded that the effects of hypnosis are greater on mental function and
on overt behavior than on levels of physiological arousal, and that ex-
actly the opposite is the case with such somatically oriented techniques
as progressive relaxation and biofeedback (Lehrer & Woolfolk, 1984).
We found that the most dramatic results among studies of hypnosis
were obtained on self-report measures or on overt voluntary behavior.
Thus, even in studies of ostensibly physiological events such as pain,
insomnia, smoking, drug craving or withdrawal, and overeating-all of
which are found to yield to hypnotic treatment-the most relevant out-
come measure consists of self-report or overt behavior, both of which
are directly manipulated by hypnotic suggestion. Although physiolog-
ical effects also are occasionally obtained with hypnosis, they tend to
be weaker, perhaps because the physiological effects are indirect. Exactly
the opposite tends to happen with such presumably somatically oriented
techniques as progressive relaxation and biofeedback.
Another approach to this problem is to define the hypnotic and
relaxation procedures operationally and from an atheoretical perspective
to examine the comparative and combined effects of the two forms of
intervention. Barber (1984) takes this course and hypothesizes that com-
bining suggestion with relaxation may help both techniques. We, on
the other hand, predict that the advantage of adding hypnotic sugges-
tion to relaxation would be seen only when an outcome measure is of
a sort that directly reflects cognitive or overt behavioral function. Con-
sistent with this, Barber states that "the goal (of relaxation) is not mus-
cular relaxation per se but a new level of being which is characterized

by peace of mind." He thus focuses his prescription of a combined tech-
nique to conditions under which the mental state associated with re-
laxation is the primary goal of therapy. This certainly would not be the
case in treating such disorders as hypertension, ulcers, or epilepsy,
when a physiological measure would be the logical measure of thera-
peutic outcome (cf. Jacobson's self-description as a specialist in internal
medicine, not psychiatry!).
A study by Friedman and Taub (1977) gives some support to our
position. This study found the combination of biofeedback and hypnotic
relaxation suggestions to have worse results than either technique alone
in reducing blood pressure. The interpretation of this study is somewhat
questionable because it used nonrandom assignment of subjects to
groups; but if these results are borne out by future studies we would
be forced to conclude that in the case of a physiological measure like
blood pressure the two techniques should not be combined. Previously
we (Lehrer & Woolfolk, 1984) explained these results by hypothesizing
that "the unrelenting truth of the biofeedback machine may undermine
the more flexible 'suggested truth' that must be believed if one is to
enter a hypnotic trance; and the hypnotic set may detract from learning
the biofeedback task." On the other hand, the fact that the biofeedback
and hypnosis groups did not differ from each other does run counter
to the prediction that hypnosis should have weaker physiological effects.
Other studies also produced results consistent with our position. Stan-
ton (1975) found hypnotic relaxation suggestions to be more effective
than progressive relaxation in improving self-reports of sleep latency
and awakenings among insomniacs. Conversely, Paul (1969a, 1969b)
found progressive relaxation to have greater effects than hypnotic sug-
gestion to relax in lowering various indices of physiological arousal.
Holroyd, Nuechterlein, Shapiro, and Ward (1982) reported some results
that at first glance appear to give our position some difficulty. In both
high and low hypnotizable subjects, faster reductions in EMG levels
were achieved with EMG biofeedback than with hypnosis, whereas
faster reductions in blood pressure and skin conduction took place with
hypnosis than with blood pressure biofeedback. A possible explanation
for the latter finding is the ineffectiveness of the blood pressure bio-
feedback condition. Some increases in arousal occurred in this condition.
We are not, of course, claiming that the cognitive and physiological
levels of functioning are completely independent of each other. One
would then be hard-pressed to explain the effects of progressive relax-
ation on such obviously mental or cognitive processes as anxiety and
depression (Lehrer & Woolfolk, 1984), or the known physiological effects
of suggestion (Barber, Spanos, & Chaves, 1974, p. 70; Barber, 1984). Our
argument now is only that hypnotic suggestion has a relatively greater

direct effect on thought and overt action, and that Jacobson's progressive
relaxation method has a relatively greater direct effect on somatic
function.
The voluminous literature on the technique of autogenic training
(Luthe, 1969), does appear to produce some problems for our theory
about the contrasting natures of hypnosis and relaxation. Autogenic
training is a self-hypnotic technique in which individuals give them-
selves self-hypnotic instruction to affect various physiological functions.
The empirical outcome literature shows rather conclusively that it can
have profound effects on various psychosomatic problems and that it
can affect various measures of somatic function rather substantially.
From our perspective, however, autogenic training is much more than
simple suggestion, just as a hypnotic technique that employs progressive
relaxation as part of the hypnotic induction is more than simple sug-
gestion. Autogenic training involves learning to perceive subtle changes
in physiological activity and to control them voluntarily. It is often com-
bined with various forms of biofeedback (cf. Norris & Fahrion, 1984).
In our opinion, the most important aspect of the autogenic technique
for achieving somatic effects is training in self-monitoring and in un-
derstanding of various physiological functions-rather than the sugges-
tive aspects of the technique. In order to evaluate our hypotheses the
autogenic training must be "dismantled" using a component control
design, (cf. Borkovec, Johnson, & Block, 1984) so that the effects of
various aspects of the technique can be studied separately. There are as
yet no such dismantling studies, but in our review of studies comparing
progressive relaxation with autogenic training (Lehrer & Woolfolk, 1984)
we found relatively greater muscular effects for progressive relaxation
and relatively greater autonomic effects for autogenic training. This is
consistent with the muscular emphasis of progressive relaxation and the
relatively greater emphasis on autonomic activity in autogenic training.
Clinically, our review found progressive relaxation (often accompanied
by EMG biofeedback) to be more effective with problems of muscular
origin than autogenic training (often accompanied by surface temper-
ature biofeedback); but for autonomic problems the two techniques ap-
pear to be equivalent. This remains consistent with Jacobson's (1938)
theory of the central importance of skeletal muscle tension in affecting
general physiological arousal. Reducing muscle tension to extremely low
levels appears to have profound effects on the autonomic nervous sys-
tem; however, the combination of suggestion and training in autonomic
control does not produce effects that generalize to the musculoskeletal
system.
Examining the relationship between hypnotizability and individual
differences in response to relaxation therapy is still another way to ex-

amine the importance of the hypnotic component in relaxation therapy.
One study from our laboratory found no correlation between hypno-
tizability and response either to progressive relaxation or autogenic
training among anxious individuals (Lehrer, Atthowe, & Weber, 1980),
whereas another found a negative correlation between hypnotizability
and response to progressive relaxation among asthmatics (Infantino et
al., 1983). Other studies have also found a lack of relationship between
hypnotizability and response to biofeedback (Engstrom, 1976; Frischholz
& Tryon, 1980; Holroyd et al., 1982). One study found a positive rela-
tionship between hypnotizability and self-reported improvement in mig-
raine headaches (Andreychuk & Shriver, 1975), whereas another study
found that high hypnotizable subjects do more poorly than low hyp-
notizable subjects in increasing EEG alpha during biofeedback training
(Dumas, 1980).
A measure that has been used in relaxation research that is closely
related to hypnotizability is that of absorption. Tellegen and his col-
leagues define absorption as the capacity to engage in episodes of "total"
attention, and they have designed a scale to measure the capacity for it
(Tellegen & Atkinson, 1974). Neff et al. (1983) reported some rather in-
triguing results indicating an interaction between response to relaxation
versus biofeedback and the capacity for absorption. In a study of head-
ache treatment using either progressive relaxation or EMG biofeedback,
they found that "vascular headache patients low in absorption were
significantly improved after biofeedback training. Tension headache pa-
tients low in absorption did not respond significantly to either form of
treatment, while those high in absorption responded significantly to
biofeedback training." Qualls and Sheehan (1979) similarly found
greater reductions among high absorption subjects in frontal EMG with
a no-feedback self-instructed relaxation procedure than with biofeed-
back, whereas low absorption subjects tended to do better with EMG
biofeedback. Roberts, Schuler, Bacon, Zimmerman, and Patterson (1975)
found no relationship between absorption and response to skin tem-
perature biofeedback.
All we can conclude from this literature is that the relationship be-
tween absorption, hypnotizability, and the ability to relax is quite com-
plex. There is some suggestion from the literature that biofeedback may
produce better results among low hypnotizable subjects and relaxation
methods (which generally included suggestion in their administration)
produce better results with high hypnotizable subjects; but there is some
inconsistency in the results, and no definitive conclusion can be made
at this time.
We have taken a rather long digression from the original topic of
the contribution of suggestion to progressive relaxation. What does the

empirical literature say about the facilitative or detrimental effects that
suggestion may have on the progressive relaxation technique, if any?
We present our conclusions as speculations and hypotheses, rather than
as proven fact. As with autogenic training, the progressive relaxation
technique has not been systematically and empirically dismantled so
that the effects of muscle training and suggestion can be separately eval-
uated. Indeed, almost no studies actually used Jacobson's original tech-
nique, minimizing suggestion; therefore, the effects of the two are con-
founded in almost all the progressive relaxation research. Our
prediction, however, is that adding suggestion to the progressive re-
laxation technique will be shown to have beneficial effects only on self-
report measures and that the muscle training component will be shown
to be more important for obtaining somatic effects.
V. CoNCLUSION
Jacobson's original progressive relaxation technique differs from the

types of progressive relaxation used by many current practitioners in a
number of fundamental respects. Jacobson emphasized relaxation as a
method of learning to control one's excess muscle tension 24 hours per
day. In his mind, progressive relaxation was not a method by which
something is done to a person. Rather, it is a method by which the
individual learns to control his or her own body. Jacobson, therefore,
rejected the use of suggestion and of various biofeedback instruments
and conditioning techniques that may induce relaxation during a train-
ing session. Empirical evaluations of most elements of the two pro-
gressive relaxation techniques have not yet been done. Thus, although
many studies have compared progressive relaxation with a number of
hypnotic, cognitive, and combined somatic-cognitive techniques, no one
has dismantled the progressive relaxation technique Jacobson's or mod-
ified versions, in order to study the exact contribution of suggestion or
cognitive interventions to the modified progressive relaxation technique,
or of teaching one muscle at a time. The evidence reviewed above, how-
ever, does lead us to hypothesize that Jacobson's original technique
would be relatively more effective in producing lasting somatic changes,
whereas the revised technique might be more effective in producing
cognitive changes or even short-term somatic changes. If these hy-
potheses are borne out, we predict that for many applications in be-
havioral medicine Jacobson's original technique will be found to be pref-
erable. This will be especially true for those disorders which cannot be
assessed by asking the patient how he or she feels but must be evaluated
physiologically (e.g., hypertension and various cardiac arrythmias,

where the patient may sometimes even feel worse when the problem
is controlled than when it is not).
In the "big picture" of therapy, of course, the distinctions between
the two techniques may be overshadowed by such overriding issues as
whether relaxation therapy is even relevant for the individual. We have
extensively discussed this issue elsewhere (Woolfolk & Lehrer, 1984b),
but we reemphasize here that we see relaxation training as a specific
method for overcoming definable problems and not as a panacea nor
as a way of life. Nevertheless, we believe that the various approaches
to the progressive relaxation technique are sufficiently different, both
in practice and in philosophy, that we would do well to evaluate these
differences in a rigorous fashion.
AcKNOWLEDGMENTS
The authors are indebted to John Atthowe for his helpful sugges-
tions in the section about hypnosis.
REFERENCES
ANDREYCHUK, T., & SKRIVER, C. (1975). Hypnosis and biofeedback in the treatment of
migraine headache. International Journal of Clinical and Experimental Hypnosis, 23. 172-
183.
BARBER, T. X. (1984). Hypnosis, deep relaxation, and active relaxation: Data, theory, and
clinical applications. In R. L. WooLFOLK & P.M. LEHRER (Eds.), Principles and practice
of stress management. New York: Guilford.
BARBER, T. X., SPANOS, N. P., & CHEVES, }. F. (1974). Hypnosis, imagination, and human
potentialities. New York: Pergamon.
BECK, A. T. (1976). Cognitive therapy and the emotional disorders. New York: International
Universities Press.
BECK, A. T. (1984). Cognitive approaches. In R. L. WooLFOLK & P. M. LEHRER (Eds.),
Principles and practice of stress management. New York: Guilford.
BEcK, A. T., RusH, A. J., SHAW, B. F., & EMERY, G. (1979). Cognitive therapy of depression.
New York: Guilford.
BENNICK, C. D., HoLST, L. L., & BENTIIEM, T. A. (1982). The effects of EMG biofeedback
and relaxation training on primary dysmenorrhea. Journal of Behavioral Medicine, 5,
329-341.
BENSON, H. (1975). The relaxation response. New York: Morrow.
BERNSTEIN, D. A., & BoRKOVEC, T. D. (1973). Progressive relaxation training: A manual for the
helping professions. Champaign, IL: Research Press.
BIRAN, M., & WILSON, G. T. (1981). Treatment of phobic disorders using cognitive and
exposure methods: A self-efficacy analysis. Journal of Consulting and Clinical Psychology,
49, 886-899.
BLANCHARD, E. B., ANDRASIK, F., ARENA. J., AHLES, T. A., }URJSH, s. E., PALLMEYER, T. P.,
BARRON, K. D., & RoDICHOK, L. D. (1982). Biofeedback and relaxation training with
three kinds of headache: Treatment effects and their prediction. Journal of Consulting
and Clinical Psychology, 50, 562-575.
BONVALLET, M., DELL, P., & HIEBEL, G. (1954). Tonus sympathique et activite electrique
corticale. Electroencephlography and Clinical Neurophysiology, 6, 119-144.
BoRKOVEC, T. D., & HENNINGS, B. L. (1978). The role of-physiological attention-focusing
in the relaxation treatment of sleep disturbance, general tension, and specific stress
reaction. Behaviour Research and Therapy, 16, 7-19.
BoRKOVEC, T. D., GRAYSON, J. B., & CooPER, K. M. (1978). Treatment of general tension:
Subjective and physiological effects of progressive relaxation. Journal of Consulting and
Clinical Psychology, 46, 518-528.
BoRKOVEC, T. D., JoHNSON, M. C., & BLocK, D. L., (1984). Evaluating experimental designs
in relaxation research. In R. L. WooLFOLK & P.M. LEHRER (Eds.), Principles and practice
of stress management. New York: Guilford.
BoRKOVEC, T. D., KALDUPEK, D. G., & SLAMA, K. (1975). The facilitative effect of muscle
tension-release in the relaxation treatment of sleep disturbance. Behavior Therapy, 6,
301-309.
BoRKOVEC, T. D., & SIDES, K. (1979). Critical procedural variables related to the physio-
logical effects of progressive relaxation: A review. Behaviour Research and Therapy, 17,
119-126.
CAMPBELL, D., SANDERSON, R. E., & LAVERTY, S. G. (1964). Characteristics of a conditioned
response in human subjects during extinction trials following a single traumatic con-
ditioning trial. Journal of Abnormal and Social Psychology, 68, 627-639.
CANTER, A., KoNDo, C. Y., & KNorr, S. R. (1975). A comparison of EMG biofeedback and
progressive muscle relaxation in anxiety neurotics. British Journal of Psychiatry, 127,
470-477.
CARRINGTON, P., CoLLINGS, G. H., }R., BENSON, H., RoBINSON, H., Wooo, L. W., LEHRER,
P. M., WooLFOLK, R. L., & CoLE,}. W. (1980). The use of relaxation-meditation tech-
niques for the management of stress in a working population. Journal of Occupational
Medicine, 22, 221-231.
CLARKE, J. C., & jACKSON, J. A. (1983). Hypnosis and behavior therapy: The treatment of anxiety
and phobias. New York: Springer.
CooKE, G. (1968). Evaluation of the efficacy of the components of reciprocal inhibition
psychotherapy. Journal of Abnormal Psychology, 73, 464-467.
DALY, E. J., DoNN, P. A., GALLIHER, M. J., & Zimmerman, J. S. (1983). Biofeedback ap-
plications to migraine and tension headaches: A double blinded outcome study. Bio-
feedback and Self-Regulation, 8, 135-152.
DAVIDSON, R. J., & ScHWARTZ, G. E. (1976). Psychobiology of relaxation and related states.
In D. MosTOFSKY (Ed.), Behavior modification and control of physiological activity. Engle-
wood Cliffs, NJ: Prentice Hall.
DAVISON, G. C. (1966). Anxiety under total curarization: Implications for the role of mus-
cular relaxation in the desensitization of neurotic fears. Journal of Nervous and Mental
Disease, 143, 443-448.
DAVISON, G. C. (1968). Systematic desensitization as a counterconditioning process. Journal
of Abnormal Psychology, 73, 91-99.
DAVISON, G. C., & WILSON, G. T. (1972). Critique of "Desensitization: Social and cognitive
factors underlying the effectiveness of Wolpe's procedure." Psychological Bulletin, 78,
28-31.
DUMAs, R. A. (1980). Cognitive control in hypnosis and biofeedback. International Journal
of Clinical and Experimental Hypnosis, 28, 58-63.
EDELMAN, R. I. (1970). Effects of progressive relaxation on autonmnic processes. Journal of
Clinical Psychology, 26, 421-425.
EDMONSTON, W. E. (1981). Hypnosis and relaxation: Modern verification of an old equation. New
York: Wiley.
ELLIS, A. (1962). Reason and emotion in psychotherapy. New York: Lyle Stuart.
ENGSTROM, D. R. (1976). Hypnotic susceptibility, EEG alpha, and self-regulation. In G. E.
ScHWARTZ & D. SHAPIRO (Eds.), Conciousness and self-regulation: Advances in research.
New York: Plenum Press, pp. 173-221.
FRANK, J. D. (1961). Persuasion and healing. New York: Shocken Books.
FRIEDMAN, H., & TAUB, H. A. (1977). The use of hypnosis and biofeedback procedures for
essential hypertension. International journal of Clinical and Experimental Hypnosis, 27,
335-347.
FRISCHHOLZ, E. J., & TRYON, W. W. (1980). Hypnotizability in relation to the ability to learn
thermal biofeedback. American journal of Clinical Hypnosis, 23, 53-56.
FUREDY, J. (1984a). Booers, beware of biofeedback boosting. Contemporary Psychology, 29,
524.
FUREDY, J. (1984b). Specific vs. placebo effects in scientific vs. snake oil medicine. Contem-
porary Psychology, 29, 599.
GELLHORN, E. (1957). Autonomic imbalance and the hypothalamus. Minneapolis: University of
Minnesota Press.
GELLHORN, E. (1958). The physiological basis of neuromuscular relaxation. Archives of In-
ternal Medicine, 102, 392-399.
GELLHORN, E., & LooFBURROW, G. N. (1963). Emotions and emotional disorders. New York:
Harper & Row.
GILBERT, G. S., PARKER, J. C., & CLAIRBORN, C. D. (1978). Differential mood changes in
alcoholics as a function of anxiety management strategies. Journal of Clinical Psychology,
34, 229-232.
GOLDFRIED, M. R. (1979). Anxiety reduction through cognitive-behavioral intervention. In
P. C. KENDALL & S.D. HoLON (Eds.), Cognitive-behavioral interventions: Theory, research,
and procedures. New York: Academic Press, pp. 117-152.
GRIMM, L. G. (1980). The evidence for cue-controlled relaxation. Behavior Therapy, 11, 283-
293.
HAYNES, S. N., GRIFFIN, P., MoONEY, D., & PARISE, M. (1975). Electromyographic biofeed-
back and relaxation instructions in the treatment of muscle contraction headaches.
Behavior Therapy, 6, 672-678.
HEss, W. R. (1925). On the relations between psychic and vegetative functions. Zurich: Schwabe.
HoHMAN, G. W. (1966). Some effects of spinal cord lesions on experienced emotional
feelings. Psychophysiology, 3, 143-156.
HoLROYD, K. A. (1979). Stress, coping, and the treatment of stress-related illness. In J. R.
McNamara (Eds.), Behavioral approaches to medicine: Application and analysis. New York:
Plenum Press.
HoLROYD, K. A., & ANDRASIK, F. (1982). Do the effects of cognitive therapy endure? A
two-year follow-up of tension headache sufferers treated with cognitive therapy or
biofeedback. Cognitive Therapy and Research, 6, 325-333.
HoLROYD,}. C., NuECHTERLEIN, K. N., SHAPIRO, D., & WARD, F. (1982). Individual differ-
ences in hypnotizability and effectiveness of hypnosis or biofeedback. The International
Journal of Clinical and Experimental Hypnosis, 30, 45-65.
HoLSTEAD, B. N. (1978). Cue-controlled relaxation: An investigation of psychophysiological
reactions to a stressor film (doctoral dissertation, University of Mississippi) Disserta-
tions Abstracts International, 1978, 39, 29878. (University Microfilms No. 78-05,763)
INFANTINo, A. T., SwARTZMANN, L. C., KARLIN, R. A., LEHRER, P.M., McCANN, B., & HocH-
RON, S. (1984). Hypnotizability: Its relationship to outcome in the behavioral treatment of
asthma. Paper presented at the annual meeting of the Association for the Advancement
of_ Behavior Therapy, Washington, DC.
JACOBSON, E. (1931). Electrophysiology of mental activities. American Journal of Psychology,

44, 677-694.
JAcOBSON, E. (1934a). Electrical measurements concerning muscular contraction (tonus)
and the cultivation of relaxation in man: Studies on arm flexors. American Journal of
Physiology, 107, 230-248.
JACOBSON, E. (1934b). Electrical measurements concerning muscular contraction (tonus)
and the cultivation of relaxation in man: Relaxation times of individuals. American
Journal of Physiology, 108, 573-580.
JACOBSON, E. (1938). Progressive relaxation. Chicago: University of Chicago Press.
JACOBSON, E. (1964). Anxiety and tension control: A physiologic approach. Philadelphia:
Lippincott.
JACOBSON, E. (1967). Biology of emotions. Springfield, IL: Charles C Thomas.
jACOBSON, E. (1970). Modern treatment of tense patients. Springfield, IL: Charles C Thomas.
JACOBSON, E. (1982). The human mind. Springfield, IL: Charles C Thomas.
JANSSEN, K. (1983). Differential effects of EMG-feedback vs. combined EMC-biofeedback
and relaxation instructions in the treatment of tension headache. Journal of Psychoso-
matic Research, 27, 243-253.
}AREMKO, M. E. (1979). A component analysis of stress inoculation: Review and prospectus.
Cognitive Therapy and Research, 3, 35-48.
KAPLAN, R. M., METZGER, G., & }ABLECKJ, C. (1983). Brief cognitive and relaxation training
increases tolerance for a painful clinical electromyographic examination. Psychosomatic
Medicine, 45, 155-162.
KAPPES, B. M. (1983). Sequence effects of relaxation training, EMG, and temperature bio-
feedback on anxiety, symptom report, and self-concept. Journal of Clinical Psychology,
39, 203-208.
KRAUT, R. E. (1982). Social presence, facial feedback, and emotion, Journal of Personality
and Social Psychology, 42, 853-863.
l<REMSDORF, R. B., KocHANOWICZ, N. A., & CasTELL, S. (1981). Cognitive skills training
versus EMG biofeedback in the treatment of tension headaches. Biofeedback and Self-
Regulation, 6, 93-102.
theory. In M. H. APPLEY & R. TRUMBULL (Eds.), Psychological stress. New York: Appleton
Century Crofts.
LAIRD, J. D. (1974). Self-attribution of emotion: The effects of expressive behavior on the
quality of emotional experience. Journal of Personality and Social Psychology, 29, 475-
486.
LANG, P. J. (1971). The application of psychophysiological methods to the study of psy-
chotherapy and behavior change. In A. E. BERGIN & S. L. GARfiELD (Eds.), Handbook
of psychotherapy and behavior change: An empirical analysis. New York: Wiley
LANG, P. J. (1977). Imagery in therapy: An information processing analysis of fear. Behavior
Therapy, 8, 862-886.
LANGOSCH, W., SEER, P., BRODNER, G., KALLINKE, D., KuucK, B., & HElM, F. (1982). Behavior
therapy with coronary heart disease patients: Results of a comparative study. Journal
of Psychosomatic Research, 26, 475-484.
LANZETI"A, J. CARTWRIGHT-SMITH, J., & KLECK, R. (1976). Effects of nonverbal dissimulation
on emotional experience and autonomic arousal. Journal of Personality and Social Psy-
chology, 33, 354-370.
LAZARUS, A. A. (1981). The practice of multimodal therapy. New York: McGraw-Hill.
LAZARUS, R. S. (1966). Psychological stress and the coping process. New York: McGraw-Hill.
LAZARus, R. S. (1977). A cognitive analysis of biofeedback control. In G. E. ScHWARTZ &
J. BEATTY (Eds.), Biofeedback: Theory and research. New York: Academic Press.
LEHRER, P. M. (1972). Psychophysiological effects of relaxation in a double blind analog of

desensitization. Behavior Therapy, 3, 193-208.
LEHRER, P.M. (1978). Psychophysiological effects of progressive relaxation in anxiety neu-
rotic patients and of progressive relaxation and alpha feedback in nonpatients. Journal
of Consulting and Clinical Psychology, 46, 389-404.
LEHRER, P. M. (1982). How to relax and how not to relax: A re-evaluation of the work of
Edmund Jacobson. Behaviour Research and Therapy, 20, 417-428.
LEHRER, P. M. (1983). Biofeedback boosters beware. Contemporary Psychology, 28, 824-825.
LEHRER, P. M. (1984a). Yes: A cautious (and limited) boost for biofeedback. Contemporary
Psychology, 29, 524.
LEHRER, P. M. (1984b). Biofeedback: Not penicillin but not snake oil. Contemporary Psy-
chology, 29, 599.
LEHRER, P.M. & WooLFOLK, R. L. (1984). Are stress reduction techniques interchangeable,
or do they have specific effects?: A review of the comparative empirical literature. In
R. L. WooLFOLK & P. M. LEHRER (Eds.), Principles and practice of stress management.
New York: Guilford.
LEHRER, P.M., ArrHOWE, J., & WEBER, E. S. P. (1980). Effects of relaxation and autogenic
training on anxiety and physiological measures, with some data on hypnotizability.
In F. J. McGuiGAN, W. SIME, & J. M. WALLACE (Eds.), Stress and tension control. New
York: Plenum Press.
LEHRER, P. M. BATEY, D., WooLFOLK, R. L., REMDE, A., & GARLICK, T. (1985). The effect
of repeated tense-release cycles on EMG and self-report of muscle tension. Submitted
for publication.
LEHRER, P.M., HocHRON, S., McCANN, B., SwARTZMAN, L., & REBA, P. (In press.) Relaxation
decreases asthma in asthmatic subjects with large-airway constriction. Journal of Psy-
chosomatic Research.
LocKE, E. (1971). Is "behavior therapy" behavioristic? An analysis of Wolpe's psychoth-
erapeutic methods. Psychological Bulletin, 76, 318-327.
LoMONT, J. F., & EDWARDS, J. E. (1967). The role of relaxation in systematic desensitization.
Behaviour Research and Therapy, 5, 11-25.
LUTHE, W. (1969). Autogenic therapy (Vols. 1-6). New York: Grune & Stratton.
McGuiGAN, F. J. (1978). Imagery and thinking: Covert functioning of the motor system.
In G. E. ScHWARTZ & D. SHAPIRO (Eds.), Consciousness and self-regulation: Advances in
research and theory (Vol. 2). New York: Plenum Press.
McGUIGAN, F. J. (1984). Progressive relaxation: Origins, principles, and clinical applica-
tions. In R. L. WooLFOLK & P. M. LEHRER (Eds.), Principles and practice of stress man-
agement. New York: Guilford.
MEICHENBAUM, D. (1972). Cognitive modification of test anxious college students. Journal
of Consulting and Clinical Psychology, 39, 370-380.
MEICHENBAUM, D. (1977). Cognitive behavior modification: An integrative approach. New York:
Plenum Press.
MEICHENBAUM, D., & CAMERON, R. (1973). Stress inoculation: A skills training approach to
anxiety management. Unpublished manuscript, University of Waterloo.
MILLER, W. R., & DIPILATO, M. (1983). Treatment of nightmares via relaxation and desen-
sitization: A controlled evaluation. Journal of Consulting and Clinical Psychology, 6, 870-
877.
NEFF, D. F., BLANCHARD, E. B., & ANDRASIK, F. (1983). The relationship between capacity
for absorption and chronic headache patients' response to relaxation and biofeedback
treatment. Biofeedback and Self-Regulation, 8, 177-183.
NICASSIO, P. M., BoYLAN, M. B., & McCABE, T. G. (1982). Progressive relaxation, EMG
biofeedback, and biofeedback placebo in the treatment of sleep-onset insomnia. British
Journal of Medical Psychology, 55, 159-166.
NoRRIS, P. A., & FAHRION, S. L. (1984). Autogenic biofeedback in psychophysiological

therapy and stress management. In R. L. WooLFOLK & P.M. LEHRER (Eds.), Principles
and practice of stress management. New York: Guilford.
NovACO, R. W. (1975). Anger control: The development and evaluation of an experimental treat-
ment. Lexington, MA: Heath.
ORNE, M. T. (1982). Perspectives in biofeedback: Ten years ago, today, and .... In L.
WHITE & B. TURSKY, Clinical biofeedback: Efficacy and mechanisms. New York: Guilford,
pp. 422-437.
OsBERG, J. W. (1981). The effectiveness of applied relaxation in the treatment of speech
anxiety. Behavior Therapy, 12, 723-729.
PARKER, J. C., GILBERT, G. S., & THORESON, R. W. (1978). Reduction of autonomic arousal
in alcoholics: A comparison of relaxation and meditation techniques. Journal of Con-
sulting and Clinical Psychology, 46, 879-886.
PATEL, C. (1984). Yogic therapy. In R. L. WOOLFOLK & P.M. LEHRER (Eds.), Principles and
practice of stress management. New York: Guilford.
PAUL, G. L. (1966). Insight vs. desensitization in psychotherapy. Stanford, CA: Stanford Uni-
versity Press.
PAUL, G. L. (1969a) Inhibition of physiological response to stressful imagery by relaxation
training and hypnotically suggested relaxation. Behaviour Research and Therapy, 7, 249-
256.
PAuL, G. L. (1969b) Physiological effects of relaxation training and hypnotic suggestion.
Journal of Abnormal Psychology, 74, 425-437.
PEAL, R. J., HANDAL, P. J., & GILNER, F. H. (1981). A group desensitization procedure for
the reduction of death anxiety. Omega, 12, 61-69.
PHILIPS, C. & HUNTER, M. (1981). The treatment of tension headache:-11. EMG "normality"
and relaxation. Behaviour Research and Therapy, 19, 499-507.
PoLLARD, G., & ASHTON, R. (1982). Heart rate decrease: A comparison of feedback mo-
dalities and biofeedback with other procedures. Biological Psychology, 14, 245-257.
PRAGER-DECKER, I. J. (1978). The relative efficacy of progressive muscle relaxation, EMG
biofeedback and music for reducing stress arousal of internally versus externally con-
trolled individuals (Doctoral dissertation, University of Maryland). Dissertations Ab-
stracts International, 1979, 39, 3177B. (University Microfilms No. 79-00,924.
QuALLS, P. J., & SHEEHAN, P. W. (1979). Capacity for absorption and relaxation during
electromyograph biofeedback and no-feedback conditions. Journal of Abnormal Psy-
chology, 88, 652-662.
RoBERTS, A. H., SCHULER, J., BACON, J. G., ZIMMERMANN, R. L., & PATTERSON, R. (1975).
Individual differences and autonomic control: Absorption, hypnotic susceptibility, and
the unilateral control of skin temperature. Journal of Abnormal Psychology, 84, 272-279.
RussELL, R., SIPICH, J., & KMPE, J. (1976). Progressive relaxation training: A procedural
note. Behavior Therapy, 7, 566-568.
ScHANDLER, S. L. & DANA, E. R. (1983). Cognitive imagery and physiological feedback
relaxation protocols applied to clinically tense young adults: A comparison of state,
trait, and physiological effects. Journal of Clinical Psychology, 39, 672-680.
SHEEHAN, D. (1983). The anxiety disease. New York: Scribner.
SHERIDAN, C. L., VAUGHAN, K. S., WALLERSTEDT, M. J., & WARD, L. B. (1977, March).
Electromyographic biofeedback and progressive relaxation compared: Interactions with gender
and type of dependent measure. Paper presented at the annual meeting of the Biofeedback
Society of America, Orlando, Florida.
SIME, W. (1982). Review of Principles and practice of progressive relaxation by E. jACOBSON &
F. J. McGuiGAN. Stress and Pain Manager Newsletter, 3, 1,3-4.
SJME, W. E., & DEGooD, D. E. (1977). Effect of EMG biofeedback and progressive muscle
relaxation training on awareness of frontalis muscle tension. Psychophysiology, 14, 522-

530.
SNow, W. G. (1977). The physiological and subjective effects of several brief relaxation
training procedures. (Doctoral dissertation, York University, Canada, 1977). Disser-
tations Abstracts International, 1984, 40, 3458B.
STANTON, H. E. (1975). The treatment of insomnia through hypnosis and relaxation. Terpnos
Logos, 3, 4-8.
STOUT, C. C., THORNTON, B., & RussELL, H. L. (1980). Effects of relaxation training on
students' persistence and academic performance. Psychological Reports, 47, 189-190.
SuJNN, R. M. (1975). The cardiac stress management program for Type A patients. Cardiac
Rehabilitation, 5, 13-15.
Suzui<J, S. (1979). Zen mind, beginner's mind. New York: Weatherhill.
TELLEGEN, A., & ATKINSON, G. (1974). Openness to absorbing and self-altering experience
("absorption"), a trait related to hypnotic susceptibility. Journal of Abnormal Psychology,
83, 268-277.
ToMI<JNS, S. S. (1970). Affect as the primary motivational system. In M. B. ARNoLD (Ed.),
Feelings and emotions. New York: Academic Press.
ToWNSEND, R. E., HousE, J. F., & ADDARIO, D. (1975) A comparison of biofeedback-mediated
relaxation and group therapy in the treatment of chronic anxiety. American Journal of
TuRNER, P. E. (1978). A psychophysiological assessment of selected relaxation strategies
(Doctoral dissertation, University of Mississippi). Dissertations Abstracts International,
1978, 39, 1010B. (University Microfilms No. 78-24,063).
TURK, D. C., MEICHENBAUM, D. H., & GENEST, H. (1983) Pain and behavioral medicine: A
cognitive-behavioral perspective. New York: Guilford.
TYRER, P. (1976). The role of bodily feelings in anxiety. London: Oxford University Press.
VAN HASSEL, J. H., BLOOM, L. J., & GoNZALEZ, A.M. (1982). Anxiety management with
schizophrenic outpatients. Journal of Clinical Psychology, 38, 280-284.
WALSH, P., DALE, A., & ANDERSON, D. E. (1977). Comparison of biofeedback of pulse wave
velocity and progressive relaxation in essential hypertensives. Perceptual and Motor
Skills, 44, 839-843.
WHITE, L., & TURSKY, B. (1982). Clinical biofeedback: Efficacy and mechanisms. New York:
Guilford.
WHITE, P. D., GJLNER, F. H., HANDAL, P. J., & NAPOLI,}. G. (1983). A behavioral intervention
for death anxiety in nurses. Omega, 14, 33-41.
WILI<JNS, W. (1971). Desensitization: Social and cognitive factors underlying the effective-
ness of Wolpe's procedure. Psychological Bulletin, 76, 311-327.
WILSON, V. E., & BIRD, E. (1981). Effects of relaxation and/or biofeedback training upon
hip flexion in gymnasts. Biofeedback and Self-Regulation, 6, 25-34.
WoLPE, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University
Press.
WoLPE, J., & LAZARUS, A. A. (1966). Behavior therapy techniques. New York: Pergamon.
WooLFOLK, R. L., & LEHRER, P. M. (1984a). Clinical stress reduction: An overview. In R.
L. WooLFOLK & P.M. LEHRER (Eds.), Principles and practice of stress management. New
York: Guilford.
WooLFOLK, R. L., & LEHRER, P.M. (1984b). Clinical applications. In R. L. WooLFOLK & P.
M. LEHRER (Eds.), Principles and practice of stress management. New York: Guilford.
WooLFOLK, R. L., & McNuLTY, T. F. (1983). Treatment of insomnia: A component analysis.
Journal of Consulting and Clinical Psychology, 51, 495-503.
Author Index
Abrams, S., 125 Basbaum, A. I., 88, 89, 94 Blouin, D., 158
Ach, N., 2 Basmajian, J. V., 43 Bock, J. C., 142, 144
Adair, J. R., 58 Batey, D., 193 Boden, M., 6
Adams, J. E., 88, 94 Battit, G., 74 Bodnar, R. J., 89, 90, 92,
Addario, D., 190 Baum, W. M., 174 100
Ader, R., 97, 98 Baxter, R., 146 Bonvallet, M., 58, 197
Akaike, A., 94 Beahrs, J. 0., 149 Boon, C., 130
Akil, H., 88, 89, 90, 93 Beck, A. T., 197, 199 Borkovec, T. D., 183, 188,
Alexander, F., 131 Beck, E. C., 74 189, 190, 192, 193, 204-
Allport, D. A., 13, 14 Bekesy, G. V., 119 205, 207
Amir, S., 89, 90, 92 Bellman, K., 4 Botwinick, J., 67
Amit, Z., 89, 90, 92 Bennett, D. H., 148 Boylan, M. B., 194
Anderson, D. E., 194, 195 Bennett, G. ]., 88, 90 Branch, B. J., 97
Andrasik, F., 195, 199 Bennick, C. D., 194 Brener, J., 68, 138-139,
Andreychuk, T., 208 Benson, H., 185 140, 141, 142, 153, 156
Anisman, H., 100 Bentov, I., 80 Brown, B. B., 30
Antonis, B., 13, 14 Bering, E. A., 80 Brown, H. 0., 21
Antrobus, J. S., 31 Berka, C., 67 Brucker, B. S., 147, 152,
Antrobus, T. S., 30, 42 Berman, J. S., 130 157
Appleton, J., 21 Berman, W. H., 140 Buchsbaum, M., 69
Archer, R. L., 130 Bernstein, D. A., 183, 192 Buck, R. W., 127
Ascough, J. C., 146 Bernstein, G. L., 141 Burgess, A. W., 128
Ashton, R., 194 Bhatnagar, R. K., 97 Burgess, I. S., 174
Atkinson, G., 157, 208 Binder, R., 128 Buss, A., 116
Atthowe, J., 208 Biran, M., 199
Azami, J., 94 Birch, L., 21 Cacioppo, J. T., 49, 50,
Bird, E., 194 63, 64, 68
Birren, J., 67, 77 Cain, M. P., 44
Baars, B., 16 Birzis, L., 61 Callaway, E., 67, 69
Bacon, J. G., 208 Bissell, L. E., 69 Cameron, R., 198
Bain, A., 19, 24 Black, A. H., 68 Campbell, D., 196
Baker, ]., 138 Blackburn, R., 126 Campos, J., 67
Ball, G., 153 Blanchard, E., 148, 156, Cannon, J. T., 89, 09, 92,
Barber, T. X., 205, 206 195 93
Barchas, J. D., 89, 90 Bleecker, E. R., 147, 157 Cannon, W., 55, 56, 57
Bardo, M. T., 97 Block, D. L., 207 Canter, A., 194
Barrett, J., 33 Bloom, F. E., 93 Cardon, P., 67
Barrios, B. A., 109 Bloom, L. J., 201 Carmon, A., 74
Bartoshuk, A. K., 38 Bloomfield, H. H., 44, 45 Carrington, P., 190
217
218 AuTHOR INDEX
Cartwright-Smith, J., 27, Davison, G. C., 196, 197, Fattuson, H. F., 56

197 198, 201 Fechner, G., 25
Castiglioni, A. ]., 88 Deckert, G. H., 30 Fenigstein, A., 116
Chan, B., 89, 90 DeDoux, ]. E., 77 Festinger, L., 118
Chance, W. T., 89, 90, 92, DeGood, D. E., 189 Fields, H. L., 88, 94
93,94 Delfini, L., 67 Folkman, S., 132
Chang, K.-J., 92 Delius, W., 172 Forgy, C., 4
Chaves, J. F., 206 Dell, P., 58, 197 Forster, A., 59
Chernigovsky, V. N., Dennis, S. G., 88, 90 Fowles, D. C., 121, 125
163, 166, 167 Denny-Brown, D., 10 Frank, J., 130, 204
Chesher, G. B., 89, 90 Derouesne, ]., 9 Freud, S., 120, 131, 132
Chew, C., 122 Dewey, W. L., 93 Fridlund, A. J., 19, 50
Choiniere, M., 88 Diamond, B., 67 Friedman, H., 206
Chudler, E. H., 92 Diamond,]., 166nl Friesen, W., 19
Clairborn, C. D., 190 DiCara, L. V., 68 Frischholz, E. ]., 208
Clanton, C. H., 89 Diliberto, E. J., Jr., 92 Furedy, J. ]., 137, 195
Clarke, J. C., 205 Dince, W. M., 142 Fuster, J. M., 10
Clarke, M. A., 142, 144 DiPilato, M., 200
Cleckley, H., 125 Ditto, W. B., 147 Gadzhiev, 9
Clemens, W. ]., 138, 146 Donchin, E., 74 Gahery, Y., 59
Cobb, S., 109, 118 Donn, P. A., 194 Galambos, R., 75, 77
Cobelli, D. A., 92, 94 Doxey, N., 142, 144 Galliher, M. J., 194
Cohen, B. H., 19, 24, 34, Dumas, R. A., 208 Garlick, T., 193
38,50 Duncan, L 13 Gazzaniga, M. A., 149,
Cohen, M. J., 146 Dunlap, K., 21 150
Cole, R. A., 48 Dustman, R. E., 74 Gebhart, G. F., 97
Coleman, T. G., 173 Dworkin, B. R., 163, 174, Geisser, S., 31
Coleridge, H. M., 59 179, 181 Gellhorn, E., 196, 197
Coleridge,]. C. G., 59 Dyk, R. B., 56 Genest, H., 199
Conrad, M., 131 Gershon, E., 146
Cooke, G., 201 Edelman, R. I., 190 Gibson, J. J., 120
Coon, D. L 90 Edmonston, W. E., 203, Giesler, G. J., 88, 90
Cooper, K. M., 190 204 Gilbert, G. S., 189, 190
Corner, M. A., 80 Edmundson, E. D., 148 Gilmore, J. P., 168n2
Costell, S., 199 Edwards, J. E., 201 Gilner, F. H., 200, 201
Cox, D.]., 140 Egger, M. D., 152 Glass, D., 116
Crane, L. A., 94 Ehrlichman, H., 33 Glass, G. V., 130
Crowley, W. R., 100 Ekman, P., 19 Glick, S., 94
Crowne, D. P., 116, 126 Elliot, R., 67 Glusman, M., 90
Curzon, G., 94 Ellis, A., 197 Goldband, S., 141
Cuthbert, B. N., 33-34 Ells, J. G., 13 Goldfried, M. R., 199
Ellsworth, P., 19 Goldman, Nina, 183
Goldstein, K., 9
Dale, A., 194, 195 Emery, G., 199
Goleman, D., 116
Daly, E. J., 194 Engel, B. T., 146, 147, 157
Gombus, G. M., 74
D'Amour, F. E., 90 Engstrom, D. R., 208
Gonzales, A.M., 201
Dana, E. R., 201 Eriksen, C. W., 126
Goodell, J., 108
D'Andrade, R., 16 Evarts, E. V., 21
Goodenough, D. R., 56
Davidson, L. A., 56 Goodman, L. S., 21
Davidson, R. J., 116, 126, Fahrion, S. L., 207 Gough, H. G., 125
184-185, 197 Farnell, S. K., 57 Gowen, A., 141, 142, 146,
Davies, M., 126 Farrell, T., 141 147
AUTHOR INDEX 219
Graf, V., 67 Holst, L. L., 194 Kaiser, D. N., 67, 68

Graham, K. R., 30 Holstead, B. N., 191 Kaloupek, D. G., 193
Granger, H., 169, 170, Holt, V., 92 Kamin, L. J., 152
171, 173, 174 Hongell, A., 172 Kaplan, R. M., 201
Gray, J., 120-121, 129 Hoover, C. W., 107 Kappes, B. M., 195
Grayson, J. B., 190 Horowitz, M. J., 130 Karp, 5. A., 56
Greene, W. A., 146 Hosobuchi, Y., 88 Katkin, E. 5., 138, 141
Greenhouse, S. W., 31 House, J. F., 190 Keefe, F. J., 140
Greenwald, A. G., 5 Howard, J. L., 39 Keele, S. W., 13, 16
Griffen, P., 190 Hughes,}., 89 Kelley, H. H., 120
Grimm, L. G., 191 Hull, C. L., 119, 155 Kelly, B., 146
Grudin, J., 16 Humphrey, D. R., 60 Kelly, D. D., 90
Guillemin, R., 91, 92, 93 Hunter, M., 201 Kennedy, J. L., 80
Guyton, A. C., 169, 170, Hutcheson, J. 5., 59 Kibler, G., 30
171, 173 Hutchinson, P., 158 Kilpatrick D. G., 129
Hutson, P. H., 94 Kimmel, H. D., 146, 147
Kissen, D. M., 126
Hagbarth, K. E., 172 Ikemi, Y., 61 Kleck, R. E., 127, 197
Handa!, P.}., 200, 201 Iker, H., 110 Klein, M. V., 96
Hansen, J. C., 76 Imiolo, D., 141 Klivington, K. A., 77
Hardyck, C. D., 47 Inanaga, K., 61 Knipe, J., 190
Hare, R. D., 63, 125 Infantino, A. T., 208 Knott, S. R., 194
Hasher, L., 151 Ingvar, D. H., 61, 69 Kobasa, 5., 109
Hayes, R. L., 89, 90, 94 Isenhart, R., 61 Koch, E., 57
Haynes, S. N., 190 Izard, C. E., 19, 50 Kochanowicz, N. A., 199
Hazum, E., 92
Kokka, N., 97
Hebb, D. 0., 23, 24 Jablecki, C., 201 Komisaruk, B. R., 100
Hefferline, R. F., 23 Jackson, J. A., 205 Kondo, C. Y., 194
Helson, H., 11 Jackson, R. L., 90 Konarski, J., 120
Hennings, B. L., 204-205 Jacobson, E., 22, 23, 44, Krantz, D., 116
Henriken, 0., 171 45, 52, 183, 184, 185, Krausz, H. I., 75
Hentall, I. D., 94 186, 187, 188, 189, 190, Kraut, R. E., 197
Hernandez-Peon, R., 119 191, 192, 193, 194, 195, Kremsdorf, R. B., 199
Herrnstein, R. J., 174 196, 197, 202, 203, 204, Krynock, G. M., 89, 93
Herz, A., 92 205, 206, 207, 209 Kubiak, E. W., 148
Hess, W. R., 196 Jaffe, D. T., 44 Kunze, D. L., 163
Hiebel, G., 58, 197 James, W., 2, 15, 25, 31, Kuramoto, 5., 61
Hilgard, E. R., 149 39, 40, 55, 56, 57, 80
Hillyard, 5. A., 75, 76 Janssen, K., 194
Hinton, G., 16 Jaremko, M. E., 198 Lacey, B. C., 39, 60, 63,
Hirst, W., 13 Jarrell, M. P., 158 67
Hobbs, W., 140 Johnson, M. C., 207 Lacey, J. I., 39, 60, 63, 67,
Hochron, 5., 202 Johnson, P. C., 174, 175 196, 197
Hoffer, J. L., 59 Jonas, C., 34 Lacroix, J. M., 137, 140,
Hohmann, G. W., 197 Jones, H. E., 126 141, 142, 143, 144, 145,
Hi:ikfelt, T., 89 Jourard, 5. M., 120, 130 146, 147, 148, 153, 156,
Holmes, D. S., 148 Joy, M. D., 58 157
Holmes, T., 109, 110 Judy, W. V., 57 Laird, J. D., 127, 197
Holmstrom, L. L., 128 Jutai, J. W., 74 Lang, P. J., 33-34, 185
Holroyd, B. N., 208 Lange, A. F., 30
Holroyd, J. C., 206 Kagan, J., 63 Langosch, W., 200
Holroyd, K. A., 195, 199 Kahneman, D., 2, 5 Lanzetta, J. T., 127, 197
220 AUTHOR INDEX
Laskey, W., 172 Mandler, G., 16 Morell, M. A., 141

Laudenslager, M. L., 100 Mannard, A., 172 Morff, R. J., 174
Laverty, S. G., 196 Manning, J. W., 58 Moss, H., 63
Lavis, S., 144 Marks, D. F., 33 Mountcastle, V., 163
Layne, R., 67, 69 Marley, N. ]., 89 Murray, E. N., 138
Lazarus, A. A., 183, 185, Marlin, R. G., 141, 153 Mycielska, K., 2, 12
186, 190 Marlowe, D., 116, 126
Lazarus, R. S., 56, 120, Marsden, D. C., 10 Nakazawa, T., 94
132, 195, 198 Mason, K. A., 98 Napoli, J. G., 201
LeDoux, J. E., 149, 150 Matin, E., 30 Naranjo, C., 44
Lee, W., 16 Matin, L., 30 Neff, D. F., 195, 208
Lehrer, P. M., 183, 185, Matus, 1., 153 Neisser, U., 13, 120
186, 188, 189, 193, 194, Maudsley, H., 19 Nelson, H., 9
195, 199, 202, 204, 205, Mayer, D. ]., 88, 89, 90, Nelson, L. R., 87, 97, 98
206, 207, 208, 210 92, 94 Newlon, P. G., 89, 90
Lenox, J. R., 20 McCabe, T. S., 194 Newsome, H. H., 92
LeShan, L. L., 110 McCann, B., 202 Nicassio, P. M., 194
Leuba, C., 21 McCanne, T. R., 57, 63, Nicholson, R. A., 130
Levene, H. 1., 146 67 Nisbett, R. E., 118, 149
Levenson, R. W., 127, McCarthy, 9 Nolan, J. M., 64
146-147 McClelland, J. L., 3 Norman, D. A., 1, 2, 3,
Levy, S. M., 98 McCubbin, J. A., 59 12, 27-28
Lewis, J. W., 87, 88, 89, McDermott,]., 4 Norris, P. A., 207
90, 91, 92, 93, 94, 95, McFarland, D. ]., 163, Notarius, C. 1., 127
96, 98, 99 174, 179 Novaco, R. W., 199
Lewis, K., 13 McFie,]., 9 Neuchterlein, K. N., 206
Lhermitte, F., 9 McGivern, R. F., 100
Libby, W. L., Jr., 39, 63 McGuigan, F. J., 21, 23,
Obrist, P. A., 39, 59, 68,
Liebeskind, J. C., 87, 88, 28, 48, 50, 184, 194, 196
147
89, 90, 91, 92, 93, 94, McLean, J. P., 11
Obrist, W. D., 69
95, 96 McLeod, P. D., 13, 16
Offutt, C., 157
Light, K. C., 59 McNulty, T. F., 200 O'Halloran, ]. M., 61
Linchitz, R., 88 McPeek, B., 74 O'Heeron, R. C., 130
Liu, S. H., 97 Meichenbaum, D. H., O'Keefe, ]., 89
Livingston, R. B., 87 140, 197, 198, 199 Oleson, T. D., 89
Llewelyn, M. B., 94 Melton, A. W., 120 Oliveras, J. L., 88
Locke, E., 198 Melzack, R., 88, 90, 119 Orne, M. T., 121, 125
Lomont, J. F., 201 Mendelson, M., 166 Ornstein, R. E., 44
Long, J., 16 Metzger, G., 201 Osberg, ]. W., 200
Loofbourrow, G. N., 196 Miczek, K. A., 100
Lovaas, K. M., 96
Mifflin, S. W., 163 Pappas, B., 67, 77
Lowenstein, W. R., 166
Millan, M. ]., 92, 93 Parise, M., 190
Luria, A. R., 8, 9, 28
Miller, N. E., 142, 147, Parker, J. C., 189, 190
Luthe, W., 207
152, 157, 174, 181 Patel, 190n2
Lynn, M., 10
Miller, T. 1., 130 Patrick, R. L., 90
Miller, W. R., 200 Patterson, R., 208
MacLenna, A. ]., 92 Milner, B., 9 Paul, G. L., 183, 186, 206
Madden, J., 90, 93 Mirgorodsky, V. N., 172 Pavlov, 1., 119
Maier, S. F. 1 90, 99 Monguillot, J. E., 158 Peal, R. J., 200
Maixner, W., 93 Mooney, D., 190 Pearce, D., 30
Malmejac, ]., 57 Mordkoff, A.M., 56 Pearce, J. W., 60
AuTHoR INDEX 221
Pearson, J. A., 146 Roberts, L. E., 141, 147, Shallice, T., 1, 9, 13, 27-
Pedigo, N. W., 93 148, 153, 156, 158 28
Pennebaker, J. W., 107, Robbins, T. W., 10 Shapiro, D., 64, 146, 147,
108, 109, 110, 116, 122, Roddie, I. C., 172 206
130, 149 Rodier, W. I., III, 28 Shattock, R. J., 146
Penner, E. R., 95 Rodriguez-Sierra, J. F., Shavit, Y., 87, 98, 99, 100
Perera, T. B., 23 100 Shaw, B. F., 199
Persson, B., 59 Rogers, M. C., 74 Shaw, W. A., 22
Pert, A., 89 Rosecrans, J. A., 89, 90, Sheehan, D., 185nl
Peters, L. ]., 98 93 Sheehan, P. W., 153, 157,
Peterson, L. H., 60 Rosenthal, F., 59 208
Petrides, 9 Ross, A., 138, 141 Shepherd, J. T., 172, 176
Petrinovich, L. F., 47 Ross, N., 125 Sheridan, C. L., 190
Petty, R. E., 49, 50, 64, 68 Rossier, J., 93 Sherman, J. E., 91, 92
Pew, R. W., 8 Rothe, C. F., 171 Shiffrin, R. M., 1, 2, 40,
Philips, C., 201 Rubenstein, C., 110, 112, 151
Phillips, S., 67 115 Shriver, C., 208
Picton, T. W., 75, 76 Rudy, T. A., 89 Shulman, A. G., 5
Pillsbury, W. B., 15, 19, Ruemlhart, D. E., 3, 16 Shulman, G. L., 11
25 Rush, A. ]., 199 Shuster, L., 100
Pollard, G., 194 Russell, H. L., 194 Sibby, R., 174
Polosa, C., 172 Russell, R., 190 Sides, K., 188, 189
Ponte, JU., 61 Rylander, G., 9, 10 Signoret, J.-L., 9
Pope, K. S., 40, 41 Silver, R. L., 130
Posner, M. I., 2, 13 Sime, W., 189, 193
Saari, M., 67, 77
Prager-Decker, I.]., 195 Singer, J. L., 30, 31, 40,
Sagawa, K., 165 41,42
Prieto, G. ]., 94
Sahaikian, B., 10
Przewlocki, R., 92 Sipich, J., 190
Sanderson, R. E., 196
Purves, M. ]., 61, 74 Sipprelle, C. N., 146
Sandman, C. A., 55, 56,
Skalak, R., 166
57, 61, 63, 64, 65, 67,
Qualls, P. L 153, 157, Skinner, B. F., 120
68, 69, 71, 73, 74, 76,
208 Sklar, L. S., 100
77, 81
Quy, R. J., 148 Satinoff, E., 165
Skok, V. I., 172
Slama, K., 193
Satoh, M., 94
Schad, H., 170, 172 Smith, D. L., 90
Rahe, R., 109, 110
Schandler, S. L., 201 Smith, M. L., 130
Randich, A., 93
Reason, J. T., 2, 12 Scheier, M., 116 Smith, M. 0., 19
Reba, P., 202 Schenkenberg, T., 74 Smith, R. E., 60
Reeves, J. L., 64 Schmale, A. H., 110 Smith, S. M., 21
Remde, A., 193 Schmidt, R. A., 8 Snow, W. G., 186
Renthem, T. A., 194 Schneider, W., 1, 2, 40, Snyder, M., 116
Rescorla, R. A., 120 151 Sotaniemi, K. A., 69
Resick, P. A., 129 Schober, R., 147 Spanos, N. P., 206
Reynolds, D. V., 88 Schuler,]., 208 Speisman, J. C., 56
Reynolds, P., 13, 14 Schwartz, G. E., 116, 126, Spelke, E., 13
Ribot, T. A., 19, 26 146, 147, 184-185, 197 Spence, K. W., 119
Richardson, D. D., 88 Schwartz, J. C., 94 Sperry, R. W., 20, 21
Riley, D. M., 137 Scott, R. E., 148 Stapleton, J. M., 88
Risse, 149 Seligman, E. P., 100 Steger, J. A., 11
Roberts, A. H., 208 Seller, H., 170, 172 Steiner, S. S., 90, 142
Roberts, H. H. T., 94 Selye, H., 91, 119, 128 Stern, M., 146
222 AUTHOR INDEX
Stilson, D. W., 153 Turk, D. C., 140, 199 Weinberg, V. E., 89

Stone, T. W., 59 Turner, P. E., 186 Weinberger, D. A., 126
Stones, M. H., 130 Tursky, B., 146, 147, 195 White, A. C., 89, 93
Stout, C. C., 194 Twombly, D. A., 89 White, L., 195
Stoy, E. G., 33 Tyrer, P., 185nl White, P. D., 201
Striker, 31 White, T. W., 148
Surwit, R. S., 140 Urea, G., 89, 90 Widdicomb, J. G., 163
Sutterer, J. R., 39 Wilkins, W., 198
Suzuki, S., 190n2 Valins, S., 118 Williams, R. J., 141
Svensson, T. H., 59 Van Hassel, J. H., 201 Williamson, D. A., 158
Swartzman, L., 202 Van Twyver, H. B., 146, Willison, J. R., 77
147 Wilson, S. P., 92
Tachibana, S., 61 Vaughan, K. S., 190 Wilson, D. H., 149
Takagi, G., 94 Veronen, L. ]., 129 Wilson, D. T., 149
Taub, H. A., 206 Vigier, D., 59 Wilson, G. T., 198, 199
Taylor, A. N., 97 Visintainer, M. A., 100 Wilson, V. E., 194
Taylor, J., 126 Viveros, 0. H., 92 Winstead, C. L., JR., 23
Tellegen, A., 157, 208 Volpicelli, J. R., 100 Wirth, H. G., 146
Terman, G. W., 87, 91, Wise, J. A., 141
92, 93, 94, 95, 96 Waid, W. M., 121, 125 Witkin, H. A., 56
Terrace, H. S., 120 Wallerstein, H., 38 Wolff, H. G., 108
Thompson, F. J., 60 Walsh, K. W., 8, 9, 10 Wolpe, J., 183, 186, 198
Thompson, L. W., 67 Walker, B. B., 63, 64, 67, Wood, A., 144
Thompson, M. L., 100 68, 69, 71, 73, 77, 81 Woody, C. D., 176
Thoren, P., 59 Walker, J. M., 81 Woolfolk, R. L., 185, 193,
Thoreson, R. W., 189 Wallerstedt, M. J., 190 194, 199, 200-204, 205,
Thornton, B., 194 Wallin, B. G., 172 206, 207, 210
Todd, D., 74 Walsh, P., 194, 195
Toman, J. E. P., 21 Walter, M., 89 Yaksh, T. L., 89
Tomkins, S. S., 197 Ward, F., 206 Yanagisawa, N., 10
Torda, C., 97 Ward, L. B., 190 Yeung, J. C., 89
Tordoff, M. G., 92 Washburn, M. F., 19, 22 Young, L. D., 148
Totten, E., 23 Watkins, L. R., 88, 90, 92, Young, M., 48
Tournade, A., 57 94 Young, R. F., 88
Townsend, R. E., 190 Watson, S. J., 89, 92
Trasler, G., 125 Weardon, J. H., 174 Zacks, R. T., 151
Treisman, A. M., 2, 11 Webb, R. A., 39 Zimmerman, J. S., 194
Tricklebank, M. D., 94, 96 Weber, E. S. P., 208 Zimmerman, R. L., 208
Tryon, W. W., 208 Weimer, W. B., 20 Zorman, G., 94
Subject Index
Action, 1-2 Biofeedback, 137-162

Activation values, 6 learning theory and, 138-147
Adrenal hormones, 92 motor theory of voluntary thinking
Affect and, 47-48
consciousne ss and, 56-57 progressive relaxation and, 194-195
inhibition and, 126-127 questions in, 137-138
progressive relaxation and, 195-202 two-process theory in, 152-158
Alcohol exposure, 97-98 verbal interface, 147-152
Amphetami nes, 10 See also Feedback
Anxiety, 196 Biosyntonic therapy, 64-67
Attention, 1-18 Blood, 173-176
action selection and, 13 Blood pressure
cardiovascu lar system and, 63-64 barorecepto r system and, 59, 60, 164-
conscious control and, 14-16 165
modulating role of, 10-12 biofeedback and, 147
motor theories of, 24-26 kidney and, 169-170
motor theory of voluntary thinking spinal anesthesia and, 171
and, 38-40, 42 Brain
neuropsycho logical evidence and, 8- barorecepto r system and, 58-59
10 behavior and, 150-151
perception and, 1 cardiovascu lar system and, 69-80
tasks and, 2-3, 13-14 contention scheduling and, 10
theory and, 3-7 frontal lobe lesions, 9
Auditory evoked potentials, 74-77 information -processing systems in,
Autogenic training, 207 149
Automaticit y motor theories and, 21
conscious control and, 14-15 motor theory of voluntary thinking
definitions of, 1-2 and, 26, 40
theory and, 3 physiologica l regulation and, 163
Awareness, 1-2. See also Attention prefrontal lesions, 8-9
reality experience and, 55-56
Barorecepto r system sensory information and, 87
blood pressure and, 164-165 Bulimia, 109
cardiovascu lar physiology and, 57-63
Basal ganglia, 10 Cardiac afferent influences, 55-85
Behavior Cardiac output
brain and, 159-151 regulation factors in, 169, 174
cardiovascu lar system and, 63-67 tonic sympathetic control of, 171
learning theory and, 139 See also Heart
223
224 SUB)ECf INDEX
Cardiovascular system Eating disorders, 109

baroreceptor physiology and, 57-63 Electroencephalography (EEG)
behavior and, 63-67 heart/brain relationship, 69-74
brain and, 69-80 progressive relaxation and, 208
consciousness and, 67-68 Electromyography (EMG)
See also Heart biofeedback and, 142-144, 147
Central nervous system motor theories and, 22, 23
interoceptors and, 176-178 motor theory of voluntary thinking
physiological regulation and, 173 and, 47-48, 49-52
sensory perception and, 87 progressive relaxation and, 189, 193,
See also Brain; Spinal cord 208
Cerebral blood flow, 61 Emotion. See Affect
Chemical regulation, 173-176 Endocrine system, 91-93
Child molestation. See Pedophilia Ergotrophic emotional system, 196
Cognition Error correction, 9. See also Slips of
confiding and, 130-131 action
consciousness and, 56-57 Evoked vascular responses, 74-77
inhibition and, 120 Experience
learning theory and, 139 action and, 1-2
motor theory of voluntary thinking reality and, 55
and, 42-43 Eye. See Vision
progressive relaxation and, 184-185,
195-202 Face, 50
Conditioned relaxation, 190-191 Feedback
Confiding confiding and, 118
cognition and, 130-131 motor theories and, 21, 23
inhibition and, 128-129 motor theory of voluntary thinking
trauma and, 118-129 and, 43-44
Conflict resolution, 4, 5-6 See also Biofeedback
Conscious control Feedforward processes, 153-155
neuropsychology and, 8 Frontal lobe lesions, 9
slips of action and, 12 Frontal syndrome, 9
speed of activity and, 11
supervisory attention system and, 14- Guilty knowledge test, 108, 121-124
16
Consciousness
biofeedback and, 155 Hand temperature data, 144-145
cardiac afferent influences on, 55-85 Heart
cardiovascular system and, 67-68 cardiac output regulation, 169
cognition/emotion and, 56-57 motor theory of voluntary thinking
motor theory and, 20-21 and, 39
Contention scheduling, 5-6 See also entries under Cardiac;
action theory and, 3-5 Cardiovascular system.
neuropsychology and, 8, 10 Heart rate control, 141, 146-147
Creativity, 43 Horizontal thread
Curare studies action theory, 4
motor theory, 21 model aspects of, 8
motor theory of voluntary thinking supervisory attentional system, 6-7
and, 21, 42 Hypertension
progressive relaxation and, 196-197 biofeedback and, 147
See also Blood pressure
Dopaminergic system, 10 Hypnosis, 203-204, 205-209
SUBJECT INDEX 225
Ideo-motor acts, 2 Neurotransmitters, 93-94

Inhibition Nucleus raphe magnus, 94
contention scheduling and, 5
definition of, 119-120 Operant autonomic conditioning
failure to confide as, 128-129 biofeedback and, 138, 147, 156
personality considerations and, 124- verbal interface and, 148
127
psychodynamic views of, 131-132 Pain perception, 87-106
psychotherapy and, 130 endocrine system in stress analgesia,
Interoceptive stimuli, 139-140 91-93
Interoceptors endogenous analgesia system
central nervous system and, 176-178 activation, 89-90
physiological regulation and, 165-166, endogenous control mechanisms, 88-
167-168 89
immunosuppressive and tumor-
enhancing effects of stress, 98-99
Kidney, 169-170
Kinesthetic imagery, 22-23, 28 independent opioid forms of stress
analgesia, 95-96
neuroanatomy of stress analgesia, 94-
Language
95
brain and, 150
neurotransmitters in stress analgesia,
visu~l imagery and, 32 93-94
Learning, 163-182
opioid and nonopioid mechanisms of
Learning theory
stress analgesia, 91
biofeedback and, 138-147, 152-158
prenatal alcohol exposure and, 97-98
brain and, 149-150
spinal cord and, 87-88
physiological regulation and, 176, 178
See also Perception
Parkinson's disease, 10
Meditation, 43-45 Pedophilia, 64-67
Migraine headache, 144-145 Pendulum effect, 191-194
Motivation, 7 Perception
Motor systems, 139, 140 attention and, 1-18
Motor theory cardiova.scular system and, 63-64
attention and, 24-26 consciousness and, 56-57
cognition and, 22-24 motor theory and, 20
consciousness and, 20-21 See also Pain perception
Motor theory of voluntary thinking Physiological regulation, 163-182
(MTVT), 19-54 Pituitary hormones, 91-92
attention (spontaneous) and, 39-40 Prefrontal lesions, 8-9
attention (voluntary) and, 38-39 Progressive relaxation, 183-216
central assertion of, 45-46 behavior therapy and, 183-184, 185-
defining of, 26-38 186
motor theories and, 20-26 cognitive/somatic factors in, 195-202
research implications of, 46-52 origin of, 183
stream of thought regulation and, 40- pedagogy versus technology in, 187-
45 195
Muscle. See entries under Motor suggestion and, 202-209
See also Relaxation
Naloxone, 89, 90 Psychotherapy, 130
Natural relaxation, 41-43
Nervous system. See Brain; Central Reality, 55
nervous system; Spinal cord Refractory period, 5
226 SuB)Ecr INDEX
Relaxation Stress analgesia

biofeedback and, 142-144 activation of system in, 89-90
motor theory of voluntary thinking endocrine system and, 91-93
and, 41-45 neuroanatomy of, 94-95
See also Progressive relaxation neurotransmitters in, 93-94
REM sleep opioid and nonopioid forms of, 95-96
motor theory of voluntary thinking prenatal alcohol exposure and, 97-98
and, 42-43 two independent opioid forms of, 95-
tonic skeletal muscle vasoconstriction 96
during, 172 Stroop phenomenon, 11
Repression, 125-126 Suggestion, 186, 202-209
Response image, 139 Supervisory attentional system, 6-7
action selection and, 13
Selective attention, 11 conscious control and, 14-16
Self-disclosure. See Confiding modulating role of, 10-12
Sensory processes, 59 motor theory and, 27-28
Skin, 171-172 neuropsychological evidence of, 8-10
Skin conauctance tests, 141 Sympathectomy, 171-172
Sleep
motor theory of voluntary thinking
and, 42-43 Taped relaxation techniques, 189-190
progressive relaxation and, 200 Task competition, 13-14
tonic skeletal muscle vasoconstriction Tasks, 2-3
during, 172 Taste, 34
Slips of action, 12 Thinking. See Cognition
Smell, 34 Transcendental meditation, 44-45
Socialization Trauma, 107-136
affect and, 126 confiding and, 118-119
inhibition and, 125 health and, 109-118
Somatic factors, 195-202 Triggers, 3-5, 6
Speech, 23, 28-29 Trophotrophic emotional system, 195
Speed of performance, 11-12
Spinal cord Vasomotor control, 171-172
blood pressure and, 171 Verbal processing, 147-152
pain perception and, 87-88 Vertical thread influences, 6
sensory perception and, 87 Vision
See also Brain; Central nervous system motor theory and, 20, 23-24
Stimulation-produced analgesia, 88-89 motor theory of voluntary thinking
Stream of consciousness, 40-45 and, 29-34
Stress Volition. See Will
cognition and, 132
confiding and, 119, 131
immunosuppressive and tumor- Will, 2
enhancing effects of, 98-99 motor theory of voluntary thinking
pain perception and, 88 and, 26-28
progressive relaxation and, 198, 199, supervisory attention system and, 14-
200 16

(Advances in Research and Theory 4) Richard J. Davidson, Gary E. Schwartz, David Shapiro (Eds.) - Consciousness and Self-Regulation-Springer (1986)

Uploaded by

Copyright:

Available Formats

(Advances in Research and Theory 4) Richard J. Davidson, Gary E. Schwartz, David Shapiro (Eds.) - Consciousness and Self-Regulation-Springer (1986)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Advances in Research and Theory 4) Richard J. Davidson, Gary E. Schwartz, David Shapiro (Eds.) - Consciousness and Self-Regulation-Springer (1986)

Uploaded by

Copyright:

Available Formats

Vol.

SPRINGER SCIENCE+BUSINESS MEDIA, LLC

Main entry under title:

Consciousness and self-regulation.

Vol. 3, 4 has subtitle: Advances in research and theory.

© 1986 Springer Science+Business Media New York

All rights reserved

No part of this book may be reproduced, stored in a retrieval system, or transmitted

BARRY H. CoHEN, Department of Psychology, State University of New

GREGORY W. TERMAN, Department of Psychology, University of Califor-

In the Preface to the third volume, we described the evolution of this

The first chapter "Attention to Action: Willed and Automatic Control

diac feedback on neurobehavioral processes. For example, he examined

variations in the administration of relaxation are evaluated, such as live

1. Attention to Action: Willed and Automatic Control

2. The Motor Theory of Voluntary Thinking 19

I. Previous Motor Theories 20

B. Specific Motor Associations 28

3. Cardiac Afferent Influences on Consciousness 55

4. Intrinsic Control Mechanisms of Pain Perception

IV. Opioid and Nonopioid Mechanisms of Stress Analgesia

5. Inhibition and Cognition: Toward an Understanding

6. Mechanisms of Biofeedback Control: On the

III. The Verbal (Consciousness?) Interface 147

7. Learning and Long-term Physiological Regulation

8. Progressive Relaxation Then and Now: Does Change

Author Index 217

2 Self-Consciousness and Intentionality: A Model Based on an

3 Self-Regulation of Stimulus Intensity: Augmenting/Reducing and

4 Neodissociation Theory of Multiple Cognitive Control Systems

5 Hypnotic Susceptibility, EEG-Alpha, and Self-Regulation

6 Toward a Cognitive Theory of Self-Control

7 Physiological and Cognitive Processes in the Regulation of Anxiety

8 Dreaming: Experimental Investigation of Representational and

9 Biofeedback and the Twilight States of Consciousness

1 The Human Brain and Conscious Activity

2 Imagery and Thinking: Covert Functioning of the Motor System

3 Regulation of the Stream of Consciousness: Toward a Theory of

4 Self-Deception, Self-Confrontation, and Consciousness

5 Visceroception, Awareness, and Behavior

6 Stimulus-Bound Behavior and Biological Self-Regulation: Feeding,

7 Operant Conditioning of Autonomic Responses: One Perspective

8 Acquired Control of Peripheral Vascular Responses

9 On the Nature of Alpha Feedback Training

10 Passive Meditation: Subjective, Clinical, and Electrographic

1 A Reason for Doubting the Existence of Consciousness

2 Conscious Contents Provide the Nervous System with Coherent,

3 Event-Related Brain Potentials in the Study of Consciousness

4 Anxiety and Fear: Central Processing and Peripheral Physiology

5 Meditation: In Search of a Unique Effect

DoNALD A. NoRMAN AND TIM SHALLICE

ond, it refers to the way an action may be initiated without deliberate

5. They require the overcoming of a strong habitual response or

FIGURE 1. A horizontal thread. For well-learned, habitual tasks an autonomous, self-suf-

of action schemas constitutes a "horizontal thread." The important point.

come more specialized in their use of processing structures, reducing

B. Determination of Activation Values

C. The Supervisory Attentional System

HORIZONT Al VE RT ICAL THREA DS

That horizontal thread control of action may be viewed within a