Development of P300 Detection Algorithm for Brain Computer Interface in Single Trial

This paper proposed a combinational method to estimate and discriminate the P300 component. The Kalman filter and the wavelet transformation were used to estimate the P300 component and PCA and three discrimination methods; linear discrimination analysis, Mahalanobis distance, and Bayes´ theorem, were used to discriminate the estimated P300 from single trial. The results show the effectiveness of the combinational technique in discrimination. The combination of wavelet transformation and a linear discrimination analysis can perform discrimination in the highest accuracy and the fastest processing speed.