Adaptive power projection method for accumulative EEG classification

For the dynamic classification of motor imagery mind states in the brain-computer interface (BCI), we propose a power projection based feature extraction method to classify the electroencephalogram (EEG) signals by combining information accumulative posterior Bayesian approach. This method improves the classification accuracy by maximizing the average projection energy difference of the two types of signals. The experimental results on two BCI competition datasets show that the classification accuracy is about 90%. The results of the classification accuracy and mutual information demonstrate the effectiveness of this method.