Rhythmic components of spatio-temporally decorrelated EEG signals based Common Spatial Pattern

Abstract

This paper addresses the performance of Common Spatial Pattern (CSP) on rhythmic band information selected from temporally decorrelated signals by zero-phase FIR digital filter. The standard blind source separation (BSS) method is applied to the EOG corrected EEG signals to make the EOG corrected EEG signals temporally decorrelated. This work considers the standard CSP with IEEE-1057 signal reconstruction algorithm for extraction of rhythmic information from the EOG corrected EEG signals too. The selected rhythmic band information is further processed by the CSP method for the feature extraction. The performance of the proposed method has been evaluated by BCI performance evaluation parameters: classification accuracy (ACC) and Cohen's Kappa co-efficient (k).