Detecting P300 potential for speller BCI

Worldwide a number of people suffer from severe disabilities like advanced amyotrophic lateral sclerosis, brain stroke, paralysis, locked-in syndrome (LIS) wherein the motor neurons are severely affected. Such patients lose their ability to communicate via muscles including tongue which cripples their lives badly. There is a need to develop some form of communication tools for such people. P300 based BCI is a very useful tool for such patients. In this work, a simple algorithm is proposed which can help detect P300 potential in electroencephalogram (EEG) data. This P300 potential can then be used to type alphabets on a computer screen and help a disabled person to communicate with the outside world. Mean accuracy of 89.5% is obtained in the current algorithm, which suggests that it can be used in an online BCI to predict words with high accuracy.