A Sliding Window Common Spatial Pattern for Enhancing Motor Imagery Classification in EEG-BCI

Accurate binary classification of electroencephalography (EEG) signals is a challenging task for the development motor imagery (MI) brain–computer interface (BCI) systems. In this study, two sliding window techniques are proposed to enhance MI. The first one calculates longest consecutive repetition (LCR) sequence prediction all windows and named SW-LCR. second mode SW-Mode. Common spatial pattern (CSP) used extracting features with linear discriminant analysis (LDA) each time window. Both SW-LCR SW-Mode applied on publicly available BCI Competition IV-2a data set healthy individuals stroke patients’ set. Compared existing state art, performed better in case left- versus right-hand MI lower standard deviation. For both sets, accuracy (CA) was approximately 80% kappa ( $\kappa $ ) 0.6. results show that window-based using robust against intertrial intersession inconsistencies activation within trial thus can lead reliable performance neurorehabilitative setting.