Implementation method of brain-computer interface system based on Fourier transform

Brain-computer interface(BCI) systems require real-time processing speed and higher and more accurate identification rate.An implementation method of on-line BCI systems for identification of mantel tasks was presented on the basis of analysing the amplitude-frequency and phase synchronization of brain's electrical signals.Taking the extracted spectral peak and phase synchronization coherency index as the characteristic parameters for expressing the imaginary movements in the brain,a time-variable linear classifier was designed on the basis of informations accumulation for identifying the imaginary mental tasks of left and right hands,so that a satisfactory result was obtained with the maxim accuracy of 90.72%.The investigation results showed that the characteristics of spectral peak was a sensitive parameter for quantifying the measure of event-related desynchronization/synchronization,and the sptetra peak combined with phase coherency index could provide more information reflecting the state of brain's mental tasks.The presented algorithm with fast Fourier transform and linear discriminant analysis for feature extraction and classification was simple and quick-operational,providing a new idea and means to realize the on-line BCI system.