The research of mental task classification based on wavelet packet and phase synchrony

For BCI system based on motor imagery, most algorithms are only based on power changes of mu and beta rhythms. A method of classification based on feature combination of wavelet packet and phase synchrony was proposed to improve the correct classification rates of mental task EEG signal. First, With the help of wavelet package, the energy related to frequency bands and time were computed, then the energy and phase synchrony as a vector and the vector was used by BP Neural network to recognize the pattern of the EEG. The classification of imaginary left and right hand movements Results show that the methods are effective for recognition of mental tasks.