brain research 1534 (2013) 66–75

Available online at www.sciencedirect.com

www.elsevier.com/locate/brainres

Research Report

The gender you are and the gender you like: Sexual
preference and empathic neural responses

D. Perrya, K. Waldera, T. Hendlerb,c,d, S.G. Shamay-Tsoorya,n

a
Department of Psychology, University of Haifa, Mount Carmel, Haifa 31905, Israel
b
Functional Brain Center, Wohl Institute for Advanced Imaging, Tel Aviv Sourasky Medical Center, Weizman 6, 46239
Tel Aviv, Israel
c
Department of Psychology, Tel Aviv University, P.O. Box 39040, 69978 Tel Aviv, Israel
d
Sackler Faculty of Medicine, Tel Aviv University, P.O. Box 39040, 69978 Tel Aviv, Israel

ar t ic l e in f o abs tra ct

Article history: Background: Empathy relates to the ability to share the emotions and understand the
Accepted 21 August 2013 intentions and emotions of the other. Although it has been suggested that women have
Available online 28 August 2013 superior empathic abilities as compared to men, it is unknown whether it is the gender or
Keywords: the sexual preference of the individual that affects empathy. Given that sexual attraction
Empathy has been reported to affect social behavior, the present study explored the possibility that
fMRI sexual orientation affects behavioral measures of empathy as well as empathy related
Sexual attraction activations.
Gender Methods: Fifty two heterosexual and homosexual women and men were scanned while
TPJ performing an emotional judgment task involving emotional understanding of a protagonist.
Results: The behavioral and neuroimaging results indicate that empathy is related to the
gender as well as the sexual preference of the participant. Individuals sexually attracted to
men (heterosexual women and homosexual men) showed greater empathy than subjects
attracted to women (heterosexual men and homosexual women). Furthermore, brain imaging
data reveal that regions within the temporo-parietal junction (TPJ), showed sensitivity to the
sexual orientation of the individual, such that it was activated more in subjects attracted to
men than in subjects attracted to women while evaluating the emotional state of the other.
Moreover, the activation in the TPJ was found to be correlated with the degree to which
subjects were empathizing.
Conclusions: These results suggest that individual differences in empathy are related to the
gender as well as the sexual orientation of the subject.
& 2013 Published by Elsevier B.V.

1. Introduction in our everyday lives. Proper function in social circumstances

depends, to a great measure, on our successful understand-
Social cognition links together a vast array of abilities that ing of the people around us. Empathy is a central mechanism
help us function appropriately in interpersonal relationships of understanding the other, which helps us sense the other's

n
Corresponding author.
E-mail address: [email protected] (S.G. Shamay-Tsoory).

0006-8993/$ - see front matter & 2013 Published by Elsevier B.V.

http://dx.doi.org/10.1016/j.brainres.2013.08.040
 brain research 1534 (2013) 66–75 67

feelings and emotions (Rogers, 1957). Empathy is a multi- Roth-Hanania et al., 2011), some functional imaging studies
dimensional construct and comprises the cognitive as well as have found evidence for sex related differences in empathy
the emotional reactions of individuals to events experienced (Schulte-Ruther et al., 2008; Derntl et al., 2010; Chou et al., 2011).
by other individuals. Thus, empathy may involve reactions For example, Singer et al. (2006) showed that empathy-related
such as emotion recognition, perspective taking and emo- activations are differently modulated in men and women by
tional contagion which may occur simultaneously or sepa- learned preferences, which are gained through social inter-
rately (Shamay-Tsoory, 2011). actions. In line with this finding, it has also been found that
Empathy and its various constructs have been shown to be women tend to recruit areas containing mirror neurons, such
sensitive to individual differences, particularly to gender as the inferior frontal region, more prominently than men
related differences (Schulte-Ruther et al., 2008; Yang et al., (Schulte-Ruther et al., 2008). Hence the question whether
2009; Derntl et al., 2010; Pavlova et al., 2010). Indeed, several gender influences empathy skills is an open one which needs
studies have supported the view of female superiority in more exploration.
empathy-related tasks, such as decoding non-verbal commu- However, no study to date has examined sexual prefer-
nication, picking up subtle nuances from tone of voice or ence of the subject relates to different empathic abilities.
facial expression, and judging a person's character (Klein and Although sexual preference has been shown to be critical to
Hodges, 2001; Hall and Matsumoto, 2004). Although several social behavior (Liu et al., 2011), little is known about the
studies failed to find gender differences in empathy (Bandstra relationship between sexual orientation and empathy and its
et al., 2011; Roth-Hanania et al., 2011) it has been reported underlying neural mechanism.
that women score higher than men on self-report measures Sexual orientation is defined as the degree of sexual
of empathy (Davis, 1994; Baron-Cohen and Wheelwright, attraction to either men or women (Ellis and Ames, 1987).
2004). The findings regarding sex differences in empathy are Studies on sexual preference have focused mainly on genes
in accordance with the fact that various psychiatric disorders, (Liu et al., 2011), prenatal hormones, and brain neuroanatomy
such as autism spectrum disease, conduct disorder, and (Bem, 1996; Rahman, 2005). It has been suggested that in
antisocial personality disorder, which are often characterized homosexual men, the interstitial nucleus 3 of the human
by impaired empathy, are far more common among men anterior hypothalamus is more female-like (Le Vay, 1991) and
(Chakrabarti and Baron-Cohen, 2006). that in homosexual women, the grey matter in the perirhinal
Recent meta-analyses have shown that empathy related cortex, a region associated with social bonding, displays a
activations include a series of different brain regions includ- more male-like structural pattern (Ponseti et al., 2007).
ing the anterior cingulate, supplementary motor area and Recently, Savic and Lindstrom (2008) demonstrated differ-
bilateral anterior insula (Fan et al., 2011; Lamm et al., 2011). ences in brain asymmetry and connectivity related to sexual
On the other hand more cognitive aspects of empathy, preference. They found that homosexual men showed similar
including metalizing related tasks have repeatedly shown to amygdala connectivity patterns to heterosexual women (i.e.,
activate the medial prefrontal cortex (MPFC), the temporo- connections with the caudate, putamen, and prefrontal
parietal junction (TPJ), and the superior temporal sulcus (STS) cortex), while the pattern of connectivity of homosexual
(Frith and Frith, 2003; Vogeley and Fink, 2003; Decety and women resembled that of heterosexual men (i.e., connection
Lamm, 2007; Schulte-Ruther et al., 2008; Van Overwalle, 2009). with the cingulate cortex).
Specifically, the TPJ is thought to be important in various In accordance with these anatomical findings, a consider-
social behaviors. The TPJ has a critical role in empathic able amount of personality and behavioral findings have
processing (Cheon et al., 2011) and seems to be engaged in demonstrated that homosexual men are more feminine than
self-reference thoughts (Johnson et al., 2002), mentalizing heterosexual men and that homosexual women are more
about the other (Lombardo et al., 2010) as well as in emotion masculine than heterosexual women (Haslam, 1997). Homo-
evaluation (Zysset et al., 2002; Winston et al., 2003), and has sexual and heterosexual individuals are also found to differ
been shown to respond atypically in autism during mentaliz- on measures of masculine instrumentality and feminine
ing (Saxe and Wexler, 2005; Lombardo et al., 2011). A common expressiveness (Lippa, 2000). Furthermore, investigations of
denominator underlying these behaviors may be the integra- cognitive skills have shown that homosexual men demon-
tion and comparisons between self and other mental state. strate female-typical responses in cognitive-related tasks,
Thus, it might be speculated that the TPJ is a key region such as spatial location memory, recall of spatial landmarks
where the decoupling of self versus other representations during navigation, and phonological and semantic fluency
occurs. While the TPJ seems to be responsible for self–other (Rahman, 2005), indicating similarities between the brains of
decoding, the MPFC appears to be involved in more abstract heterosexual women and homosexual men.
inferences about self and others. It has been shown that this Collectively, these findings support the possibility that
region is recruited in self-knowledge, person perception, and individual differences in empathy may be related not only
mentalizing processes, all of which underlie empathy (Decety to gender, but also to sexual preference. If indeed women
and Chaminade, 2003; Gallagher and Frith, 2003). show higher levels of empathy and homosexual men show
To date, few behavioral and neuroimaging studies (Canli similar neural and behavioral patterns, then it is possible that
et al., 2002; Bandstra et al., 2011; Lamm et al., 2011; Mercadillo individuals attracted to men (homosexual men, heterosexual
et al., 2011; Roth-Hanania et al., 2011; Chun et al., 2012; women) would show higher levels of empathy than indivi-
Decety and Svetlova, 2012) have investigated gender differ- duals attracted to women (homosexual women, heterosexual
ences in empathy. While several studies did not find any men). To the best of our knowledge, no brain imaging study
gender differences (Geangu et al., 2010; Bandstra et al., 2011; has investigated sexual preference differences in social
68 brain research 1534 (2013) 66–75

Fig. 1 – A schematic design of the fMRI task.

behavior and empathy. Therefore, the aim of the present this index than individuals attracted to women (homosexual
study was to examine differences in the neural empathic women, heterosexual men). Main effect for gender (F(1,48)¼
responses of individuals attracted to men versus individuals 3.97, p¼ 0.05) indicated higher empathy scores for men over
attracted to women. Given that gender differences have been women, and an interaction was found between gender and
reported for the neural correlates of emotional perspective attraction (F(1,48)¼4.15, p ¼0.047). A follow-up one-way
taking (Schulte-Ruther et al., 2008), in the current study we ANOVA for sexual orientation (homosexual men, heterosex-
used an emotional mentalizing task in which subjects rated ual men, homosexual women, heterosexual women) was
the emotional magnitude of events happening to another carried out to further explore differences in empathy levels
protagonist (“OTHER” condition) or to themselves (“SELF” among homosexual and heterosexual women and men. This
condition, Fig. 1). It was speculated that individuals attracted analysis indicated a significant sexual orientation effect
to men would show higher behavioral and neural empathic (F(3,48)¼3.89, p¼0.014). Follow-up post-hoc analysis (Duncan's
responses toward the protagonist as compared to subjects MRT) revealed that the group of homosexual women had
attracted to women. Specifically, it was hypothesized that significantly lower empathy index scores (Mean¼
1.15,
both the gender and the sexual orientation of the subjects SD¼ 0.23) as compared to the three other groups (po0.05).
would interact so that heterosexual women will show the In order to further explore the role of sexual orientation in
highest empathy followed by homosexual men, homosexual empathy, a correlation analysis was conducted between the
women and heterosexual men. Finally, we predicted that the empathy index and the KSOG score. This analysis revealed
levels of sexual orientation tendencies, as assessed by the that for women, a trend was found between the empathy
Klein Sexual Orientation Grid (KSOG) (Klein, 1993) would scores and the KSOG scores, indicating that higher hetero-
predict the levels of empathy in the task. sexual tendencies in women were associated with higher
levels of empathy (r¼
0.41, p¼ 0.06). No significant correla-
tion was found for men (r¼ 0.015, p ¼0.95).
2. Results
2.2. Neuroimaging results
2.1. Behavioral results
2.2.1. Gender and sexual attraction
To examine the empathic abilities of the participants an To determine whether the often-reported gender differences
empathy index was calculated as the ratio between the in empathy are due to the sexual attraction of the subjects,
ratings for SELF and OTHER protagonists. For purposes of we first examined the interaction between the two between
convenience, the index was multiplied by
1, indicating that subject factors (gender and sexual attraction). This analysis
higher index scores represent higher empathy and vice versa. revealed that 0.003% of 106,720 map voxels were activated
The index was calculated for each subject. both for the attraction and the gender main effects for F(1,5)¼
A univariate ANOVA of the empathy index, with gender 7.5, po0.007. That is, no overlapping brain regions were
(women, men) and attraction (attracted to men, attracted to activated for attraction and gender factors.
women) as the between-subject factors was performed. This To investigate the effects of gender on empathy related
analysis revealed a significant attraction effect (F(1,48)¼ 4.08, activations, group fixed effects general linear model analysis
p¼ 0.049), indicating that individuals attracted to men (het- was conducted. A 2  2 ANOVA (gender by protagonist con-
erosexual women and homosexual men) scored higher on dition) was carried out, with gender (women, men) as the
 brain research 1534 (2013) 66–75 69

Table 1 – Peak of activation obtained from whole-brain contrast. Talairach coordinates of regions extracted from the SELF–
OTHER contrast (random effect, FDR corrected GLM). (a) SELF4OTHER. (b) OTHER4SELF.

Region BA Voxel of peak activation (x,y,z) F p # of voxels

a
L hippocampus
24,
19,
11 6.41 0.01  10
7 336
L MPFC 10
3,53,10 14.9 0.01  10
14 43,804
R STG 21 48,
25,
5 5.77 0.01  10
7 1603
R cuneus 18 24,
79,19 5.18 0.000004 1082
R cerebellum 27,
76,
26 5.79 0.01  10
7 2468
L cingulate 24 0,
16,40 7.9 0.01  10
9 5454
L posterior cingulate 30
6,
58,7 6.95 0.01  10
9 4246
L superior occipital 19
39,
82,28 7.33 0.01  10
9 2778
L MTG 21
57,
49,7 8.17 0.01  10
11 11,848

b
R posterior cingulate cortex 31 12,
64,28 9.57 0.01  10
10 11,605

Stereotactic coordinates and t values are provided for local voxel maxima. L¼ left; R ¼ right; M ¼middle. MTG – Medial Temporal Gyrus,
STG – Superior Temportal Gyrus, MPFC – Middle Prefrontal Cortex.

between subject fixed effect factor, condition (SELF, OTHER) In order to further examine the main effects in the brain
as the within-subjects fixed effect factor, and subject (N¼ 52) structures activated by the attraction factor, a MANOVA was
as the random effect factor. An interaction effect (po0.0006, conducted for the beta weights derived from clusters acti-
FDR corrected) between protagonist (SELF, OTHER) and gen- vated in the attraction  protagonist interaction effect, for
der (men, women) was evident in several key empathy OTHER condition, with attraction as the between subject
related brain regions (see Table 2), including the MPFC, factor. Main effects for attraction were found in the TPJ for
posterior cingulate and MTG. the OTHER condition (F(1,50)¼11.9, p¼ 0.001). The rest of the
To further examine the main effects in the brain struc- effects did not survive the corrections for multiple analyses.
tures activated by the gender factor, a MANOVA was con- Follow up t-tests revealed that individuals attracted to men
ducted. This analysis was performed on beta weights derived showed higher beta weights than individuals attracted to
from clusters (each of them regarded as a dependent variable) women in the in the TPJ for OTHER condition (t(50) ¼3.45,
activated in the gender  protagonist interaction effect, for p¼ 0.001). In order to further explore the effects of gender and
OTHER condition, with gender as the between subject factor. sexual orientation on empathy-related activations, a one-way
Main effects for gender were found in the MPFC (BA 11) for ANOVA was conducted for the TPJ for OTHER condition, with
the OTHER condition (F(1,50)¼ 14.69, po0.0001). The rest of sexual orientation as the between-subjects factor. This ana-
the regions did not show significant effects following correc- lysis indicated significant differences between the groups
tion for multiple analyses. Follow up t-tests revealed that (F(3,48)¼4.13, p ¼0.01, see Fig. 2). Post hoc Duncan's MRT
women showed higher beta weights than men the in the reveals that the heterosexual women were significantly differ-
MPFC (BA 11) for OTHER condition (t(50) ¼3.82, po0.0001). ent from the heterosexual men and homosexual women.
Finally, to explore the role of sexual orientation, a one-way Furthermore, the heterosexual men are not significantly differ-
ANOVA was conducted for the MPFC (BA 11) for OTHER ent from the homosexual women (po0.05).
condition, with sexual orientation as the between-subjects To rule out the possibility that the handedness of subjects
factor. This analysis indicated significant differences between had an effect on the results (10 of the participants were left
the groups (F(3,48)¼8.69, po0.0001). Post-hoc Duncan's MRT handed), the ANOVA was performed for the TPJ for OTHER
test indicated that betas for heterosexual women were the condition with sexual orientation and handedness as the
highest and significantly different from the three other between-subjects factors. This analysis indicated that the
groups (po0.05). difference between groups remained significant (F(7,44)¼2.75,
In order to identify the effects of sexual preference p¼ 0.2), ensuring that handedness did not affect the reported
(attraction to men and attraction to women) on empathy group differences.
related activations, a second-level group analysis was con-
ducted using a fixed-random mixed effects general linear
model. A 2  2 ANOVA (attraction by protagonist condition) 2.3. Brain and behavior
was carried out, with attraction (attracted to men, attracted
to women) as the between subject fixed effect factor, condi- Following the finding that the left TPJ showed sensitivity to
tion (SELF, OTHER) as the within-subjects fixed effect factor, the OTHER emotional state, its role was further explored
and subject (N¼ 52) as the random effect factor. using a correlation analysis between this region and the
An interaction effect between protagonist (SELF, OTHER) empathy index. This analysis revealed a significant correla-
and attraction (attracted to men, attracted to women) was tion between the beta weights for the TPJ OTHER and the
evident in several key empathy related brain regions, includ- behavioral empathy index (r¼0.373, p¼ 0.006), indicating that
ing the left TPJ (BA 40) and the MPFC (BA 24, BA 32) (po0.0008, the higher the emotional ratings of the other, the more active
FDR corrected, see Table 3). the TPJ was while thinking about the other. The same
70 brain research 1534 (2013) 66–75

Table 2 – Peak of activation obtained from whole-brain analysis. Talairach coordinates of regions extracted from the 2  2
ANOVA (gender by protagonist condition), with gender as the subject fixed effect factor, condition (SELF, OTHER) as the
within-subjects fixed effect factor, and subject (N¼ 52) as the random effect factor. Clusters taken from the interaction effect
between protagonist (SELF, OTHER) and gender (men, women), random effect, FDR corrected GLM.

Region BA Voxel of peak activation (x,y,z) F p # of voxels

R posterior cingulate 23 11,
38,23 1.8 0.01  10
8 316

L MPFC 10
6,64,1 2.24 0.01  10
11 1801
L MPFC 11
44,36,
14 5.04 0.01  10
9 739
L MTG 21
55,1,
10 3.46 0.01  10
8 1490

Stereotactic coordinates and t values are provided for local voxel maxima. L¼left; R ¼right; M ¼middle. MTG – Medial Temporal Gyrus,
MPFC – Middle Prefrontal Cortex.

Table 3 – Peak of activation obtained from whole-brain analysis. Talairach coordinates of regions extracted from the 2  2
ANOVA (attraction by protagonist condition), with attraction as the subject fixed effect factor, condition (SELF, OTHER) as
the within-subjects fixed effect factor, and subject (N¼ 52) as the random effect factor. Clusters taken from the interaction
effect between protagonist (SELF, OTHER) and sexual attraction (attracted to men, attracted to women), random effect, FDR
corrected GLM.

Region BA Voxel of peak activation (x,y,z) F p # of voxels

R MPFC 32 26,37,10 6.84 0.01  10
7 3607

L MPFC 32
21,29,19 8.01 0.01  10
7 2986
L SFG 24
10,63,23 6.21 0.01  10
7 607
R precuneus 7 20,
52,42 5.7 0.01  10
7 262
L MPFC 32
12,45,10 5.98 0.01  10
7 522
L TPJ 40
45,
29,23 6.21 0.01  10
7 261
L cuneus 19
25,
82,28 5.86 0.01  10
7 632
L medial occipital gyrus 19
43,
73,3 5.64 0.01  10
7 1105

Stereotactic coordinates and t values are provided for local voxel maxima. L ¼left; R ¼ right; M ¼middle. SFG – Superior Frontal Gyrus,
MPFC – Middle Prefrontal Cortex, TPJ – Temporo-Parietal Junction.

analysis was carried out with the MPFC region and did not rest and OTHER4rest revealed brain activations in fronto
yield any significant result (r¼ 0.054, p ¼0.7). parietal and occipital and mid brain regions.
To examine the relationship between sexual orientation
and TPJ activity, a correlation analysis was conducted
between the beta weights derived from the TPJ OTHER and 3. Discussion
the subjects' KSOG scores. Correlations were calculated for
each gender separately. Results showed a trend so that The present study investigated the behavioral and neural
among men a trend of higher ratings of homosexuality were characteristics of empathy as manifested by subjects from
related to higher left TPJ activity(r¼ 0.39, p¼ 0.06). No signifi- different sexual orientation groups. Brain-related empathy
cant result was found for women (r¼
0.33, p¼ 0.12). The activations were measured in women and men with a clear
difference between the two correlation coefficients was homosexual or heterosexual preference who were under-
tested and found to be significant (z ¼
2.56, p¼ 0.005). going a mentalizing task comprising of emotional mentaliz-
ing about the self versus another protagonist.
Our data show that at the behavioral level, individuals
attracted to men (homosexual men and heterosexual
2.3.1. SELF versus OTHER women) show more empathy than subjects attracted to
To determine empathy related activations across all groups a women (heterosexual men and homosexual women). More-
whole-brain contrast was performed between the SELF and over, we found that empathy levels are significantly higher in
OTHER protagonist conditions. As shown in Fig. 3, this heterosexual women than in homosexual women. Interest-
analysis revealed a wide network showing increased activa- ingly, our data show that men score higher on empathy than
tion during emotional judgment of the self as compared to women. This may be explained by the use of four equal
judgment of others (SELF4OTHER) including the MPFC, hip- homogenous sexual orientation groups. Thus, the extremely
pocampus and cingulate cortex (po0.01  10
7, FDR corrected, low empathy scores of the homosexual women on one hand
see Table 1a). Increased activation during emotional judg- and the high empathy scores of the homosexual men may
ment of others (OTHER4SELF) involved a network including account for the unexpected superiority of men in the empa-
the posterior cingulate cortex (PCC) (po0.01  10
10, FDR thy task reported here. The brain imaging data reveal
corrected, see Table 1b). Conjunction analysis between SELF4 increased activations in self related areas such as the MPFC
 brain research 1534 (2013) 66–75 71

and the hippocampus when contrasting emotional judgment behavioral findings was found. Empathy-related structures,
of self versus other. Furthermore, we found that particularly such as the MPFC and the left TPJ, show sensitivity to the
the MPFC was sensitive to the gender of the subject such that grouping of the subjects based on their sexual preference.
women tended to show more activity in this region as These regions have been shown to be critical for self–other
compared to men. The MPFC has been shown repeatedly to decoding (Denny et al., 2012), which enables self–other
mediate perspective taking (D’Argembeau et al., 2007), men- awareness, a vital component of empathy (Decety and
talizing (Assaf et al., 2009; Perry et al., 2011; Denny et al., 2012) Jackson, 2004). Further analysis indicated that particularly
and empathy (Meyer et al., 2012), which suggest that women the TPJ is differentially activated in individuals attracted to
tend to use this region more than men when empathizing. men versus individuals attracted to women. Specifically, a
Moreover, sexual orientation, a similar pattern to the significant effect for sexual orientation was found in the left
TPJ when the subjects were judging the emotional state of the
other, with its activation being the lowest for homosexual
women as compared to the three other groups. Furthermore,
the activation in this region was significantly associated with
the level of empathy toward the other, such that the more
active the TPJ, the more the subjects were empathizing.
It has been suggested that the TPJ is a part of a network
which allows for the mental separation of one's own per-
spective from that of another person and thus enables us to
disentangle our own feelings from those observed in other
people (Schulte-Ruther et al., 2008). The sensitivity of the TPJ
to the emotional mental state of the other observed here
corresponds to other studies that show activation in this
region in tasks that involve empathy and mental states
attribution (Young et al., 2010). Though not part of the core
empathy network (Fan et al., 2011; Lamm et al., 2011), it has
been demonstrated that TPJ may be selectively involved in
representing and inferring the mental states of others (Saxe,
2006) and that left TPJ lesions impair mental state inference
(Samson et al., 2004). Here also we found that left lateralized
TPJ activation predicts levels of empathy in the task.
The right TPJ and the posterior part of the Superior Temporal
Sulcus have been repeatedly reported to be involved in the
representations of others' behavior (Blakemore and Frith, 2003;
Decety and Jackson, 2004; Decety and Grezes, 2006), studies of
agency (Decety and Grezes, 2006) as well as in mirror neuron
activity (Chong et al., 2008). It is important to note here that in
this study the activated region within the TPJ was the left BA 40,
an area which lies on the border of the inferior parietal lobule
Fig. 2 – (a) Left TPJ region (encircled in yellow) revealed by and the TPJ, not quite the typical TPJ activation found in theory
the main effect for “attraction” (uncorrected). (b) Mean left of mind studies. We suggest that this region has a pivotal role in
TPJ beta weights for OTHER condition across sexual empathic processes and that its activity is affected by one's
orientation groups. sexual preference.

Fig. 3 – Whole-brain activation patterns obtained for the contrast “SELF–OTHER” superimposed on sagittal, axial and coronal
anatomical slices (N¼ 52, random effect, FDR corrected GLM analysis). Orange represents SELF related activations and in blue –
OTHER related activations.
72 brain research 1534 (2013) 66–75

The present study provides direct behavioral and neural belonging to their sexual attraction object group. Thus, it is
evidence supporting the hypothesis that sexual preference important to further examine the current research question
modulates empathic abilities. Notably, the results from this using matched and unmatched protagonists. Furthermore, in
study were more evident among women than among men, the current study, subjects used their imagery skills to
suggesting that the behavioral and neural differences imagine and put together their protagonist. Future studies
between heterosexual and homosexual women with regard should investigate if some of the current results can be
to empathy to the other are much more prominent than the explained by imagery capacities and not empathy per se.
differences between heterosexual and homosexual men.
While functional and structural brain differences between
men and women have been widely reported, neuroanatomical, 4. Conclusion
functional and neurochemical differences between homosexual
and heterosexual men and women have yet to be investigated. We show here that subjects differ on their empathic abilities
A few conflicting studies propose that sexual orientation and depending on their sexual preference such that subjects
preference are associated with anatomical and neurochemical attracted to men empathize to a higher degree than subjects
brain differences (Ponseti et al., 2007; Rametti et al., 2010; Liu attracted to women. These results indicate that individual
et al., 2011; Rametti et al., 2011). Kranz and Ishai (2006) have differences in empathy may be modulated by one's sexual
found that within the neural circuitry for reward, heterosexual preference. Hence, sexual orientation and preference are impor-
men and homosexual women respond more to female face tant when examining individual differences in empathy.
than male face, while heterosexual women and homosexual
men show the reverse pattern and prefer male faces. In line
with this, it has been reported that homosexual women have 5. Experimental procedures
experienced higher prenatal testosterone level which may be
assessed by the second to fourth finger lengths ratio (or 2D:4D 5.1. Participants
ratio) compared to heterosexual women. Furthermore, Chapman
et al. (2006) have found a negative correlation between empathic Participants were 52 healthy individuals, ranging in age from
abilities and prenatal testosterone levels. It has also been 22–36 (mean age: 28.46, SD¼3.4). The recruitment of subjects
demonstrated that administration of testosterone in women was carried out through an ad and was based on a screening
led to a significant impairment in their empathy, and that this which included a self-report questionnaire involving direct
effect is powerfully predicted by their 2D:4D ratio (Van Honk questions regarding the sexual orientation of the participant
et al., 2011). (heterosexual/homosexual). Following the screening, the par-
Given that empathy is considered a feminine trait (Brody, ticipants were grouped into two sexual preference groups and
1985) and that homosexual men have been reported to possess were invited to the scanning session. Subjects were included
more feminine behavior entities than heterosexual men (Rieger only if they showed a clear sexual preference to one sex
et al., 2010) it is not surprising that individuals attracted to men (bisexual individuals were excluded from the study). Exclu-
show higher empathy levels. Accordingly, Salais and Fischer sion criteria included psychiatric or neurological conditions.
(1995) report that homosexual males display significantly The group that reported being attracted to women consisted
higher levels of empathy as compared to heterosexual men. of 25 subjects, including 13 heterosexual men and 12 homo-
Our findings regarding lower levels of empathy among homo- sexual women. The group that reported being attracted to
sexual women as compared to heterosexual women may be men consisted of 27 participants, including 12 homosexual
explained in the context of the former group's similarity to men and 15 heterosexual women. Ten subjects were left-
heterosexual men (Gettelman and Thompson, 1993; handed (4 heterosexual men, 2 homosexual men, 2 hetero-
Drummond et al., 2008). sexual women, and 2 homosexual women). There was no age
Yet, an equally plausible explanation may be based on difference between the groups (F(3, 42)¼ 0.199, p ¼ 0.658).
learning theory principles. The attraction to either gender Participants had no reported history of psychiatric or neuro-
usually signifies with which gender a person shares the most logical disorders and were not using psychoactive drugs at
intimate and close relationships. Hence, it is possible that the the time of the study. In addition, all participants had normal
object to which we direct our mating strategies may affect or corrected-to-normal vision. All subjects gave informed
our empathic abilities. As such, subjects attracted to a specific consent and were paid for their participation.
gender, whether heterosexual or homosexual, might share
commonalities in their social skills based on their relation- 5.2. Instruments
ship past experiences (Deaux and Major, 1987).
A limitation of the current design is the difference in Following scanning participants completed the Klein Sexual
genders of the protagonist. Homosexual and heterosexual Orientation Grid (Klein, 1993) to further assess levels of homo-
men judged the emotional states of men while homosexual sexuality and heterosexuality. The Hebrew version of the Klein
and heterosexual women judged the emotional states of Sexual Orientation Grid (KSOG) (Klein, 1993), assesses seven
women. This design was planned to minimize variability so dimensions of sexual orientation tendencies, including sexual
that each subject would judge a protagonist from his own preference, sexual behavior, sexual fantasies, emotional pre-
gender. Yet, this may be problematic as homosexuals actually ference, social preference, self-identification, and heterosexual/
judged the emotional states of other potential sexual attrac- homosexual lifestyle. The items are rated on a seven-point
tion objects, while heterosexual subjects judged others not scale, with 1 depicting opposite sex tendencies and 7 depicting
 brain research 1534 (2013) 66–75 73

same sex tendencies. On each dimension, the subjects are men, and so forth. One character from the vignettes was
asked to mark their past attitude (prior to the last year), their chosen by the experimenter to be part of the scanning
present attitude (in the last six months), and what they believe session.
to be the ideal attitude. Using this measure, we confirmed Scanning: The stimuli consisted of 120 sentences depicting
subjects' sexual preference. The mean KSOG score of the everyday emotional events happening to either the subject
heterosexual men was 44.08, SD¼ 11.4 and of the homosexual (“SELF” condition) or another protagonist (“OTHER” condi-
men was 119.00, SD¼ 9.65. As for women, the mean score for tion). The experiment was divided into four counterbalanced
the heterosexual group was 45.40, SD¼10.49 and the homo- runs of 7.25 min each, with 30 sentences presented in a
sexual group was 101.75, SD¼ 10.28. KSOG scores for the four pseudo-random manner in each run. The sentences were
groups were normally distributed (heterosexual men: W¼0.976, presented visually in a slow event-related fashion. After the
p¼0.942, homosexual men: W¼0.928, p¼ 0.392, heterosexual presentation of each sentence, participants were instructed
women: W¼ 0.909, p¼0.387, homosexual women: W¼ 0.937, to judge the emotional intensity of the event using a cursor
p¼0.463). The data of the KSOG of six subjects was lost due to that moved along a judgment bar, ranging from 5¼ highest
a technical reason and unavailable for data analysis. joy to
5¼ highest distress. Participants were instructed to
read the sentences and rate the amount of emotional inten-
sity that would be felt by the protagonists (SELF, OTHER) in
5.3. Stimuli and procedure such a situation. Event length varied from 2.5 to 7.5 s. During
the first 2.5 s, subjects had to read the sentence and indicate
In order to examine the hypothesis that individual differ- whether the event was distressful or joyful by clicking on one
ences in empathy may be related to sexual preference, we of two keys. Following this initial response, the cursor started
designed an emotional judgment task in which subjects rated to move to the left or to the right along the judgment bar. The
the emotional magnitude of events happening to another subjects then had to estimate the intensity of the emotion by
protagonist (“OTHER” condition) or to themselves (“SELF” another click, which stopped the cursor from moving at a
condition, Fig. 1). It was reasoned that the difference between point between 0 and 5 and ended the event. The sides of the
the emotional ratings for the SELF and the OTHER protago- judgment bar were switched between subjects in a counter-
nists would reflect the degree of the tendency to self-project balanced manner. Between events there were rest époques
the emotional state of others (Perry et al., 2011). Thus, an that lasted from 5 to 10 s, depending on the length of the
empathy index was created by calculating the ratio between preceding event. Each event was followed by a blank screen
the ratings for SELF and OTHER protagonists. with a fixation point.
Pre-scanning: The empathy tasks included judgments of
emotional events happening to two protagonists: “SELF” and 5.4. fMRI image acquisition and analysis
“OTHER”. Subjects were presented with short vignettes
depicting different protagonists in order to familiarize them Scanning was performed in a 3T GE scanner with an 8-
with the OTHER protagonist (for an example see Fig. 4). The channel head coil, using a gradient echo-planar imaging
vignettes matched the sexual orientation of the subject so (EPI) sequence of functional T2n-weighted images (TR/TE/flip
that heterosexual men read short stories about heterosexual angle: 2500/35/90; FOV: 20 _ 20 cm2; matrix size: 64 _ 64)
male characters, homosexual men read about homosexual divided into 40 axial slices (thickness: 3 mm; gap: 0 mm)
covering the whole cerebrum. Anatomical 3D-sequence
Vignette example - Hetreosexual Men spoiled gradient (SPGR) echo sequences were obtained with
high-resolution 1-mm slice thickness (FOV: 25 _ 18; matrix:
Asaf was born in Jerusalem in 1978. He graduated from the Academy of Arts and
256 _ 256; TR/TE: 9/3.6 ms). The fMRI data were processed
sciences in Jerusalem and now he studies cinematography in the Sam Spiegel
using the BrainVoyager QX 2 software package (http://www.
institute. Asaf rents a small apartment in central Jerusalem. He plans to finish his
brainvoyager.com).
studies by the end of 2012 and meanwhile he works during the nights in the local
Pre-processing of functional scans included head move-
café. Then, he plans a long trip to South America with Ronit, his girlfriend. His
ment assessment, high-frequency temporal filtering, and
parents got divorced ten years ago. His father remarried two years ago, and his mother
removal of low-frequency linear trends. Scans with head
lives with his 20 years old younger brother. Asaf's hobbies are: painting (he goes to an
movements 41.5 mm were rejected. In order to allow for
evening class once a week), movies and shooting pool. In the last few weeks, Asaf is
T2n equilibration effects, the first six images of each func-
helping Alon, his best friend, to organize his proposal to Michal. They make a video
tional scan were rejected. Pre-processed functional images
clip for the occasion.
were incorporated into the 3D datasets through trilinear
Asaf is an extrovert kind of guy. He has a large circle of friends around him and
interpolation. The complete dataset was transformed into
succeeds to maintain many close relationships over the years. He never forgets
Talairach space. Three-dimensional statistical parametric
birthdays and always calls to congratulate. His special Purim (Jewish holiday) parties
maps were calculated separately for each subject using a
are known all around.
general linear model (GLM), in which both stimuli conditions
He has close relationships with his family members and his younger brother often
(SELF, OTHER) were positive predictors, were convolved with
comes over for the weekend to hang out. Asaf is living a very intensive daily routine
a hemodynamic response function. Finally, spatial smooth-
and has always the need for new adventures which is translated to vacation around the
ing was applied (FWHM: 4 mm). False Discovery Rate (FDR)
world whenever it's possible.
correction method was used for the multiple comparisons
Fig. 4 – An example of vignette given to subjects from the problem. Beta weights extraction was performed from the
hetreosexual men group. whole cluster activated.
74 brain research 1534 (2013) 66–75

5.5. Statistical analysis Chun, J.W., Park, H.J., et al., 2012. Common and differential brain
responses in men and women to nonverbal emotional
To investigate this study's hypotheses we first computed an vocalizations by the same and opposite sex. Neurosci. Lett.
empathy index based on the ratio between emotional ratings 515 (2), 157–161.
D’Argembeau, A., Ruby, P., et al., 2007. Distinct regions of the
for SELF and OTHER protagonists. ANOVA was conducted on
medial prefrontal cortex are associated with self-referential
this index with ‘gender’ and ‘attraction’ as the between- processing and perspective taking. J. Cogn. Neurosci. 19 (6),
subject factors. Then, this index was correlated with the 935–944.
KSOG scores. To look for brain correlates, we first conducted Davis, M.H., 1994. Empathy: a Social Psychological Approach,
a qualitative analysis to look for brain regions which show Westview Press and location as Boulder, CO.
sensitivity both to the gender and sexual attraction main Deaux, K.M., Major, B., 1987. Putting gender into context: an
effects. Then we ran a second level group analysis with interactive model of gender-related behavior. Psychol. Rev. 94
(3), 369–389.
‘gender’ as the between subject factor. Following this, a
Decety, J., Chaminade, T., 2003. Neural correlates of feeling
MANOVA was conducted for the regions revealed by the
sympathy. Neuropsychologia 41 (2), 127–138.
group analysis. Afterward, a second group analysis was Decety, J., Grezes, J., 2006. The power of simulation: imagining
conducted, this time with ‘sexual attraction’ as the between one's own and other's behavior. Brain Res. 1079 (1), 4–14.
group factor. Here, again followed a MANOVA for the regions Decety, J., Jackson, P.L., 2004. The functional architecture
revealed by the second group analysis. Finally, brain and of human empathy. Behav. Cogn. Neurosci. Rev. 3 (2),
behavior correlations were explored using the empathy 71–100.
index, KSOG score and TPJ beta weights. Decety, J., Lamm, C., 2007. The role of the right temporoparietal
junction in social interaction: how low-level computational
processes contribute to meta-cognition. Neuroscientist 13 (6),
580–593.
Author contributions Decety, J., Svetlova, M., 2012. Putting together phylogenetic and
ontogenetic perspectives on empathy. Dev. Cogn. Neurosci. 2
(1), 1–24.
The authors declare no conflict of interest.
Denny, B.T., Kober, H., et al., 2012. A meta-analysis of functional
neuroimaging studies of self- and other judgments reveals a
references spatial gradient for mentalizing in medial prefrontal cortex.
J. Cogn. Neurosci. 24 (8), 1742–1752.
Derntl, B., Finkelmeyer, A., et al., 2010. Multidimensional
Assaf, M., Kahn, I., et al., 2009. Brain activity dissociates assessment of empathic abilities: neural correlates and
mentalization from motivation during an interpersonal gender differences. Psychoneuroendocrinology 35 (1), 67–82.
competitive game. Brain Imaging Behav. 3 (1), 24–37. Drummond, K.D., Bradley, S.J., et al., 2008. A follow-up study of
Bandstra, N.F., Chambers, C.T., et al., 2011. The behavioural girls with gender identity disorder. Dev. Psychol. 44 (1), 34–45.
expression of empathy to others' pain versus others' sadness Ellis, L., Ames, M.A., 1987. Neurohormonal functioning and sexual
in young children. Pain 152 (5), 1074–1082. orientation: a theory of homosexuality–heterosexuality.
Baron-Cohen, S., Wheelwright, S., 2004. The empathy quotient: Psychol. Bull. 101 (2), 233–258.
an investigation of adults with Asperger syndrome or high Fan, Y., Duncan, N.W., et al., 2011. Is there a core neural network
functioning autism, and normal sex differences. J. Autism in empathy? An fMRI based quantitative meta-analysis.
Dev. Disord. 34 (2), 163–175. Neurosci. Biobehav. Rev. 35 (3), 903–911.
Bem, D.J., 1996. Exotic becomes erotic: a developmental theory Frith, U., Frith, C.D., 2003. Development and neurophysiology of
of sexual orientation. Psychol. Rev. 103 (2), 320–335. mentalizing. Philos. Trans. R. Soc. London B Biol. Sci. 358
Blakemore, S.J., Frith, C., 2003. Self-awareness and action. Curr. (1431), 459–473.
Opin. Neurobiol. 13 (2), 219–224. Gallagher, H.L., Frith, C.D., 2003. Functional imaging of ‘theory of
Brody, L.R., 1985. Gender differences in emotional development: a mind’. Trends Cogn. Sci. 7 (2), 77–83.
review of theories and research. J. Pers. 53 (2), 102–149. Geangu, E., Benga, O., et al., 2010. Contagious crying beyond the
Canli, T., Desmond, J.E., et al., 2002. Sex differences in the neural
first days of life. Infant Behav. Dev. 33 (3), 279–288.
basis of emotional memories. Proc. Natl. Acad. Sci. USA 99
Gettelman, T.E., Thompson, J.K., 1993. Actual differences and
(16), 10789–10794.
stereotypical perceptions in body image and eating
Chakrabarti, B., Baron-Cohen, S., 2006. Empathizing:
disturbance: a comparison of male and female heterosexual
neurocognitive developmental mechanisms and individual
and homosexual samples. Sex Roles 29 (7), 545–562.
differences. In: Anders, S., Ende, G., Junghofer, M., Kissler, J.,
Hall, J.A., Matsumoto, D., 2004. Gender differences in judgments
Wildgruber, D. (Ed.), Progress in Brain Research, vol. 156.
of multiple emotions from facial expressions. Emotion 4 (2),
Elsevier, Amsterdam, Netherlands, pp. 403–417.
201–206.
Chapman, E., Baron-Cohen, S., et al., 2006. Fetal testosterone and
empathy: evidence from the empathy quotient (EQ) and the Haslam, N., 1997. Evidence that male sexual orientation is a
“reading the mind in the eyes” test. Soc. Neurosci. 1 (2), matter of degree. J. Pers. Soc. Psychol. 73 (4), 862–870.
135–148. Johnson, S.C., Baxter, L.C., et al., 2002. Neural correlates of self-
Cheon, B.K., Im, D.M., et al., 2011. Cultural influences on neural reflection. Brain 125 (Pt 8), 1808–1814.
basis of intergroup empathy. Neuroimage 57 (2), 642–650. Klein, F., 1993. In: The Bisexual Option. The Haworth Press,
Chong, T.T., Cunnington, R., et al., 2008. fMRI adaptation reveals Binghamton, NY.
mirror neurons in human inferior parietal cortex. Curr. Biol. 18 Klein, K.J.K., Hodges, S.D., 2001. Gender differences, motivation
(20), 1576–1580. and empathic accuracy – when it pays to understand. Pers.
Chou, K.H., Cheng, Y., et al., 2011. Sex-linked white matter Soc. Psychol. Bull. 27, 720–730.
microstructure of the social and analytic brain. Neuroimage 54 Kranz, F., Ishai, A., 2006. Face perception is modulated by sexual
(1), 725–733. preference. Curr. Biol. 16 (1), 63–68.
 brain research 1534 (2013) 66–75 75

Lamm, C., Decety, J., et al., 2011. Meta-analytic evidence for Rogers, C.R., 1957. The necessary and sufficient conditions of
common and distinct neural networks associated with therapeutic personality change. J. Consult. Psychol. 21 (2),
directly experienced pain and empathy for pain. Neuroimage 95–103.
54 (3), 2492–2502. Roth-Hanania, R., Davidov, M., et al., 2011. Empathy development
Le Vay, S., 1991. A difference in hypothalamic structure between from 8 to 16 months: early signs of concern for others. Infant
heterosexual and homosexual men. Science 253, 1034–1037. Behav. Dev. 34 (3), 447–458.
Lippa, R.A., 2000. Gender-related traits in gay men, lesbian Salais, D., Fischer, R.B., 1995. Sexual preference and altruism.
women, and heterosexual men and women: the virtual J. Homosex. 28 (1–2), 185–196.
identity of homosexual–heterosexual diagnosticity and Samson, D., Apperly, I.A., et al., 2004. Left temporoparietal
gender diagnosticity. J. Pers. 68 (5), 899–926. junction is necessary for representing someone else's belief.
Liu, Y., Jiang, Y., et al., 2011. Molecular regulation of sexual Nat. Neurosci. 7 (5), 499–500.
preference revealed by genetic studies of 5-HT in the brains of Savic, I., Lindstrom, P., 2008. PET and MRI show differences in
male mice. Nature 472 (7341), 95–99. cerebral asymmetry and functional connectivity between
Lombardo, M.V., Chakrabarti, B., et al., 2011. Specialization of homo- and heterosexual subjects. Proc. Natl. Acad. Sci. USA
right temporo-parietal junction for mentalizing and its 105 (27), 9403–9408.
relation to social impairments in autism. Neuroimage 56 (3), Saxe, R., 2006. Uniquely human social cognition. Curr. Opin.
1832–1838. Neurobiol. 16 (2), 235–239.
Lombardo, M.V., Chakrabarti, B., et al., 2010. Shared neural Saxe, R., Wexler, A., 2005. Making sense of another mind: the role
of the right temporo-parietal junction. Neuropsychologia 43
circuits for mentalizing about the self and others. J. Cogn.
(10), 1391–1399.
Neurosci. 22 (7), 1623–1635.
Schulte-Ruther, M., Markowitsch, H.J., et al., 2008. Gender
Mercadillo, R.E., Diaz, J.L., et al., 2011. Perception of suffering and
differences in brain networks supporting empathy.
compassion experience: brain gender disparities. Brain Cogn.
Neuroimage 42 (1), 393–403.
76 (1), 5–14.
Shamay-Tsoory, S.G., 2011. The neural bases for empathy.
Meyer, M.L., Masten, C.L., et al., 2012. Empathy for the social
Neuroscientist 17 (1), 18–24.
suffering of friends and strangers recruits distinct patterns of
Singer, T., Seymour, B., et al., 2006. Empathic neural responses are
brain activation. Soc. Cogn. Affect. Neurosci. 8 (4), 446–454.
modulated by the perceived fairness of others. Nature 439
Pavlova, M.A., Wecker, M., et al., 2010. Perception of intentions
(7075), 466–469.
and actions: gender stereotype susceptibility. Brain Res. 1311,
Van Honk, J., Schutter, D.J., Bos, P.A., Kruijt, A.W., Lentjes, E.G.,
81–85.
Baron-Cohen, S., 2011. Testosterone administration impairs
Perry, D., Hendler, T., et al., 2011. Projecting memories: the role of
cognitive empathy in women depending on second-to-fourth
the hippocampus in emotional mentalizing. Neuroimage 54 digit ratio. Proc. Natl. Acad. Sci. USA 108 (8), 3448–3452.
(2), 1669–1676. Van Overwalle, F., 2009. Social cognition and the brain: a meta-
Ponseti, J., Siebner, H.R., et al., 2007. Homosexual women have analysis. Hum. Brain Mapp. 30 (3), 829–858.
less grey matter in perirhinal cortex than heterosexual Vogeley, K., Fink, G.R., 2003. Neural correlates of the first-person-
women. PLoS One 2 (1), e762. perspective. Trends Cogn. Sci. 7 (1), 38–42.
Rahman, Q., 2005. The neurodevelopment of human sexual Winston, J.S., O’Doherty, J., et al., 2003. Common and distinct
orientation. Neurosci. Biobehav. Rev. 29 (7), 1057–1066. neural responses during direct and incidental processing of
Rametti, G., Carrillo, B., et al., 2011. White matter microstructure multiple facial emotions. Neuroimage 20 (1), 84–97.
in female to male transsexuals before cross-sex hormonal Yang, C.Y., Decety, J., et al., 2009. Gender differences in the mu
treatment. A diffusion tensor imaging study. J. Psychiatr. Res. rhythm during empathy for pain: an electroencephalographic
45 (2), 199–204. study. Brain Res. 1251, 176–184.
Rametti, G., Carrillo, B., et al., 2011. The microstructure of white Young, L., Dodell-Feder, D., et al., 2010. What gets the attention of
matter in male to female transsexuals before cross-sex the temporo-parietal junction? An fMRI investigation of
hormonal treatment. A DTI study. J. Psychiatr. Res. 45 (7), attention and theory of mind. Neuropsychologia 48 (9),
949–954. 2658–2664.
Rieger, G., Linsenmeier, J., et al., 2010. Dissecting “gaydar”: Zysset, S., Huber, O., et al., 2002. The anterior frontomedian
accuracy and the role of masculinity–femininity. Arch. Sex. cortex and evaluative judgment: an fMRI study. Neuroimage
Behav. 39 (1), 124–140. 15 (4), 983–991.