טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentSerby Hilit
SubjectDevelopment of a Brain-Computer Interface based on Event-
Related Potentials
DepartmentDepartment of Electrical Engineering
Supervisor Professor Emeritus Gideon Inbar (Deceased)


Abstract

A brain-computer interface (BCI) is a system for direct communication between brain and computer. A BCI can serve as a substitute for the normal human communication channel, especially for individuals that suffer from severe motor disabilities, as it relys solely on the brain’s activity and do not require motor ability. The system exploits event-related brain potentials (ERP), natural responses of the brain to external visual stimuli, as a medium for communication. In this work independent component analysis (ICA) was used in order to separate the ERP source from the background noise. A matched filter was used together with averaging techniques for detecting the existence of ERPs. The BCI developed in this work allows a subject to communicate one of 36 symbols. A further enhancement of the BCI allows it to adapt the communication rate to the users’ ability. The processing method was evaluated offline on data recorded from six healthy subjects. The method achieved a communication rate of 5.45 symbols/min with an accuracy of 92.1%, a communication rate equal to 23.75 bits/min (the communication rate takes error rates into account). The on-line interface was tested with the same six subjects. The average communication rate achieved was 4.5 symbols/min with an accuracy of 79.5%, a communication rate equal to 15.43 bits/min. This research presents a working, simple, and easy-to-use brain-computer interface. Data acquisition is performed by noninvasive means with only a few scalp electrodes. The interface works on-line, adapting the communication rate to the users’ performance to achieve good results. The presented BCI achieves good performance compared to other existing BCIs, and allows a reasonable communication rate, while maintaining a low error rate.