Classification of senrorimotor activity in EEG signal

In this study, a Common Spatial Pattern (CSP) driven Artificial Neural Network (ANN) Classification strategy is presented to classify the mental tasks, namely, left-hand movement imagination, right-hand movement imagination, and word generation in EEG data. According to this strategy, first, electrode re-referencing and band-pass filtering are used to enhance the EEG signal. Then a multi-class extension of Common Spatial Pattern (CSP) analysis is applied to extract features from the EEG data. Finally, a feed-forward Artificial Neural Network (ANN) and Support Vector Machine (SVM) are used for classification, comparatively. The performance of the methods is evaluated using the BCI Competition III dataset and an average accuracy of 70,96% is obtained among three subjects. This result is 2,31% better than the winner of the competition.