Motor Imagery Classification using local region CSP features with high-gamma band

In this paper, we present enhanced MI classification method in electroencephalogram(EEG)-based Brain-Computer Interfaces(BCI) with high-gamma band and local region CSP features. In order to improve the performance of MI classification, the use of the local region CSP feature with high eigenvalue disparity score and the high-gamma band related to advanced cognitive information processing such as reasoning and judgment can provide improved performance compared to existing CSP based methods. As a result of the experiment through BCI competition III-IVa dataset, it is shown that MI classification performance is improved through the proposed method.