Phase locking analysis of motor imagery in brain-computer interface

Currently existing brain-computer interfaces (BCIs) extract feature vectors derived from amplitude information. However, they were not use the rich phase dynamics in the EEG. Phase synchronization was opposed to use for classification of motor imagery. In suitable time window, the electrodes of C3, C4 and central regions to match were selected and then Hilbert transform signal processing method was used for extracting the degree of phase synchronization between two electroencephalogram (EEG) signals by calculating the so-called phase locking value (PLV). The support vector machine (SVM) was used for classification of the motor imagery by a feature selection algorithm. It shows that the satisfactory results are obtained with single-trial accuracies of 92.5% and that synchronization differences between motor imagery depends upon frequency selection, length of data and electrode selection.