Feature extraction of Motor Imagery Task Based on wavelet transform

Feature extraction of motor imagery task for EEG is an important issue in brain-computer interface. Considering that the electroencephalogram (EEG) signal is nonstationarity, a method of wavelet analysis is proposed to extract EEG based left and right hand motor imagery. First of all, the instantaneous average power of EEG was calculated. Then, a wavelet decomposing for five layers proceeded to EEG's average power by using db5 wavelet. Finally, the feature of motor imagery EEG was extracted by reconstructing the approximation coefficients and the detail coefficients. The results have shown that this method can effectively extract the signals of event-related synchronization and desynchronization (ERS / ERD) based on left and right hand motor imagery task. Meanwhile, we also obtain the time when the features change. Therefore, this method can extract the EEG's time-frequency information which can not be obtained by the conventional EEG analysis, and make a foundation for the classification of motor imagery tasks.