Ambulatory Epileptic Monitoring Device

Rachel Kolb, Kevin Pineda, Mark Curran, Timothy Skinner, A. Adegbege, B.F. BuSha The College of New Jersey (TCNJ) Ewing, NJ 08628
Abstract Epilepsy is a brain disorder that affects millions of adults and children in America. It is characterized by abnormal neuronal signaling and can cause strange sensations, emotions, behavior, loss of consciousness, muscle spasms, and convulsions. These episodes are very difficult to predict with the current technologies available, and almost impossible to accurately record outside clinical settings. This project aims to create an ambulatory device that can continuously measure long-term electroencephalography (EEG) outside of a hospital. The Epileptic Monitoring Device (EMD) is a portable device with two separate units: a head-mounted EEG recording system and a data viewing application on an Android platform. The EEG is recorded using five electrodes in referential montage attached to a headset and passes the signal through an isolation amplifier and then into a microprocessor for analog-to-digital conversion. The digital signal is sent via Bluetooth transceiver to the mobile device where it is processed and viewed using the Android application. An alarm will be triggered when the application detects seizure-like activity. The EMD will provide accurate monitoring for epileptic patients within the comfort of day-to-day activities. Caretakers and medical professionals can use this data to determine types of seizures, number of seizures, and the efficacy of medication. KeywordsAmbulatory Device, Computer Interface, Non-Invasive. Epilepsy, EEG, Brain-

Figure 2: Schematic of EMD design overview and signal flow II. HEADSET

I.

INTRODUCTION

Epilepsy is a neurological disorder, characterized by sudden, recurrent episodes of sensory disturbance caused by abnormal electrical activity in the brain. A key diagnostic tool for epilepsy is EEG, a test that measures the voltage fluctuations along the scalp from ionic current due to neuronal activity. Epileptic seizures are extremely difficult to predict and can be tricky to accurately record. It is also difficult for caretakers to provide continuous monitoring. To resolve this issue, an ambulatory Epileptic Monitoring Device (EMD) will alert the caregiver at seizure onset, record, and store electroencephalograhy (EEG) recordings. See Figure 1 for the SolidWorks design of the EMD Headset, and Figure 2 for the schematic of the EMD design overview.

A. Electrodes The number of electrodes used for EEG is related to the desired spatial resolution of cortical potential distribution, or electrical impulses in the brain1. For epilepsy the clinically relevant EEG frequency band is 0.1 to 100 Hz and amplitudes measured along the scalp tend to range between 10 to 100 V2. The electrodes will be placed at T7, C3, Cz, C4, and T8 using the 10-20 international system. These locations span the width of the head and are spaced equidistant. The electrodes will be arranged in a referential montage. The reference electrode is Cz and is centrally located, thus not subject to asymmetric electrical impulse propagation from the brain . The EMD will use reusable electrodes so the user wont need to continuously replace them. The EMD will use wet electrodes because they are more commercially available. The EMD will use gold plated silver electrodes similar to the Emotiv EPOC headset. The EMD will use an electrode system like that of the Emotiv, with a felt piece attached to the electrode and allowing for the conductive saline solution to be absorbed and leaves less residue. B. Headset Design The headset is designed to support and house the electrodes to record EEG data, incorporate and store the circuitry such as the wires, batteries, and breadboard. It will fit the head comfortably for an extended period of time, be elastic enough to wrap around the head with ease, and have a simple assembly. The points of contact for the headset include the back and side of the head where the headpiece wraps around the head.

Figure 1: EMD Headset design created in SolidWorks

Solidworks was used to model the headset with the electrodes attached to the human head. The Objet30 Pro 3D printer was used to create the headset because of its ability to print out models with a resolution of 600 x 600 x 900 dpi and an accuracy of 0.1 mm. The headset is designed to fit an average sized head comfortably, and the fabrication process is precise and versatile. VeroBlue was selected as the 3D printing material. III. HARDWARE

also favorable because it has an over current and reverse polarity protector. C. Android Application This design will use an application on the Android based operating system to display and store the EEG signals. Android is more popular globally than iOS. Also, the open source nature of Android and the availability Java coding support are added benefits4. Application development will be on Google's 2013 Nexus 7 running Android 4.4 (KitKat) operating system. A tablet will be used based on the cost and available operating system. Android 4.4 (KitKat) uses level 19 API and is designed to optimize performance on devices with lower hardware specifications. The Nexus 7 has Bluetooth 4.0, which is optimized to be low power and remain paired or connected without requiring a continual stream of data to be transferred between devices. IV. CONCLUSION We have designed the ambulatory epileptic monitoring device with efficiency, cost, and ease of use in mind. The comfort of the user is a top priority as well as a long-term battery to support continuous usage. It will characterize a signal as non-epileptic, epileptic, or indeterminant. The indeterminant classification requires doctor consultation for additional interpretation. The device EEG capabilities will be compared to those of the emotive EPOC device, a currently approved device available commerically. The signal processing is being simulated, tested, and verified using labVIEW software. The EEG is integrated, displayed, and stored digitally using a novel Android application. ACKNOWLEDGMENT We would like to thank TCNJ School of Engineering faculty and staff for their continued 3D printing and parts support and Dean Steven Schreiner for funding and support. REFERENCES
[1] Vaisanen, Outi. "Multichannel EEG Methods to Improve the Spatial Resolution of Cortical Potential Distribution and the Signal Quality of Deep Brain Sources." Tampere University of Technology, 13 June 2008. Web. Bohorquez, Davies P.J. "Design of a Portable Wireless EEG System Using a Fully Integrated Analog Front End." IEEE Xplore. Southern Biomedical Engineering Conference, 3 May 2013. Web. 26 Nov. 2013. Smith, S. J M. "EEG in the Diagnosis, Classification, and Management of Patients with Epilepsy." Journal of Neurology, Neurosurgery, and Psychiatry. BMJ Group Journals, 2010. Web. 25 Nov. 2013. !"#$%"&'()* +%$$,-* ./0123* /4((4(5)* 678* 1$* 9(:$14:;* 9<<* =0'* >?@A'$3)* 9<<*4(*7('*B<%"'*C*D6!E-"1@-.*!"#$%&'&()*"+-*D4@'*D'"0)*FG*9H$-*IJFK-* /'A-*IL*>1M-*IJFK-.

A. Signal Processing The signal-processing portion of the EMD needs to amplify EEG signals from microvolts to millivolts, cut off any high, unwarranted, frequencies, send the signals via Bluetooth, and characterize the behavior as seizure like, non-determinant, or non-epileptic. A power-source isolated bioinstrumentation amplifier was sourced to connect to the EEG electrodes. The circuit uses a non-inverting Op-Amp level shifter, ensuring only positive voltages are passed to the Arduino Pro Mini during analog to digital conversion (ADC). A filter was designed to cutoff any frequencies greater than 25Hz. The sampling rate was set to 125Hz. The Bluetooth Transceiver Module attached to the microcontroller transmits data between the Arduino board and Android Tablet. The characterization of the EEG signals will be done by the android application. Several defining EEG characteristics can be used to determine seizure-like activity, such as an increase in amplitude or frequency, or mismatch between the left-sided and right-sided signal. However, articles published discuss how a frequency spike lasting longer than 5-10 seconds in the 5 Hz range is an indication of seizure like activity3. Using the Fast Fourier Transform (FFT) to analyze normal and epileptic signals, the device will be able to characterize seizure-like signals based on the episodes lasting longer than 5 seconds. B. Microcontroller The Arduino Pro Mini, using Atmel 328P processor with 32 kB of flash memory and a max operating frequency of 20Mhz was selected. The integrated ADC has a 10-bit resolution at 15ksps. Operating voltage for the Pro Mini is 5V with an input voltage between 5-12V. Polymer lithium ion batteries are used to meet the design specification. Each battery produces 3.7 volts, two in series yields an appropriate 7.4 V. DC current for each I/O pin is 40mA The Arduino platform was selected for this project as the most cost-effect microcontroller that met the design specifications. Arduino is also cross-platform compatible and open sourced. The Pro Mini was specifically selected because of its small dimensions, 18x33mm. With an Atmel 328P processor, the Pro Mini is a low power microcontroller that has an ADC that fulfills the needs of the project. The Pro Mini is

[2]

[3]

[4]