Phase Locked Feature Based BCI Speller for P300 Analysis

摘要

The main purpose of Brain computer interface (BCI) research is to develop assistive devices for people with motor impairment, thus resuming their interaction with the outside world. The P300 based BCI speller is an important choice for such patients in lieu of providing the degree of freedom to its users to write anything intended on the screen. Windowing is performed for better visualization of data and improving the signal to noise ratio (SNR). Four features viz., mean, range, instantaneous phase and instantaneous amplitude are extracted. The K-fold cross validation technique is used to find the generalised classification performance using a simple linear discriminant analysis (LDA) classifier. Classification performance is measured using accuracy, sensitivity, specificity and precision. The performance of P300 based BCI speller, while considering the effect of individual feature was evaluated. It is observed that the instantaneous phase among the tested features obtained the highest classification performance with an accuracy of 99.5%.