A mutual information based channel selection scheme for P300-based brain computer interface

This paper presents a mutual information based channel selection (MICS) for P300-based Brain-Computer Interface (BCI). The proposed method iteratively selects a new channel with maximal dependency to the target class and minimal dependency with previously selected channels. Evaluation of the proposed method on the data set II from the third BCI competition shows that the MICS can enhance the P300 detection rate compared to other state of art channel selection methods developed for P300 speller BCI.