EEG filtering based on BSS algorithm and its modification for BCI

Biomedical signals have a significant temporal structure. It is obvious to use the blind source separation (BSS) algorithm based on the time lagged covariance matrix for the independent source separation from the multichannel EEG signals. In this paper work, a novel signal preprocessing technique for the enhancement of the rhythmic information related to movement imagination has been presented. The technique is based on the BSS algorithm called as AMUSE algorithm. AMUSE algorithm estimates the separating matrix by an eigenvalue decomposition of a single time lagged covariance matrix of the EEG signals. Human brain has an asymmetrical structure. The asymmetry over the primary motor cortex areas has been incorporated for the modification of the estimated separating matrix. Preprocessing by the modified separating matrix achieve almost 100% classification accuracy between the left and right hand movement imagination. Consequently there is also a significant improvement in Cohen's kappa coefficient.