Discrimination of EEG related to motor imagery by combined wavelet energy feature with phase synchronization feature

For the problem of the EEG signals classification in Brain-computer interfaces (BCI), we proposed a method of feature extraction is that the combination of the wavelet energy feature and phase synchronization feature. In this paper EEG signals are transformed by means of discrete wavelet transform firstly, and then extracted 5 wavelet energy feature in different 5 frequency bands. We also obtained the instantaneous phase values based on Hilbert Transform, and extracted the phase synchronization feature through the method of Phase Locking Value (PLV). According to the distribution of the phase locking value in the time-domain, we obtained the best feature extracting time rage. Thus the final signal features are used as inputs for a Support Vector Machine (SVM) classifier. The result indicated the best feature extracting time rage is 4s~7s, and the accuracy of classification is 89.4%, providing an improvement comparing with only using wavelet energy feature.