Extracting Common Spatial Patterns from EEG time segments for classifying Motor Imagery classes in a Brain Computer Interface (BCI)

Brain Computer Interface (BCI) is a system which straightly converts the acquired brain signals such as Electroencephalogram (EEG) to commands for controlling external devices. One of the most successful methods in BCI applications based on Motor Imagery is Common Spatial Pattern (CSP). In the existing CSP methods, common spatial filters are applied on whole EEG signal as one time segment for feature extraction. The fact that ERD/ERS events are not steady over time motivated us to break down EEG signal into a number of sub-segments in this study. Since the importance of EEG channels for classification varies for different time segments. Accordingly we extract a feature vector from each time window of EEG channels using the analysis of CSP. In order to classify a Motor Imagery EEG signals, we apply a LDA classifier based on OVR (One-Versus-the Rest) scheme on the extracted CSP features. The considered Motor Imagery four classes are: left hand, right hand, foot and tongue. We used dataset 2a of BCI competition IV to evaluate our method. The result of experiment shows that this method outperforms both CSP and the best competitor of the BCI competition IV.