A multi-modal BCI system based on motor imagery

Brain-computer interface based on motor imagery is currently considered as the most promising brain-computer interface (BCI). By off-line analytic comparison, this paper used wavelet packet decomposition (WPD) and short-time Fourier transform (STFT) to feature extraction, which provided the basis for real-time online BCI system. With combination of advantages of two electroencephalograms (EEG) as Alpha wave and motor imagery and design of control strategy, multimodal brain-computer interface was built on LabVIEW to implement functions of mouse click and web browser. The experiment results suggested that the accuracy of four motor imagery movements were above 80% and the classification accuracy of research was at a ideal level. It proved that this system is feasible and has a high application value.