Bispectrum analysis of EEG for motor imagery based BCI

Summary form only given. Electroencephalography (EEG) is the recording of electrical activity of the brain from multiple electrodes placed on the scalp. Recent years have seen a surge of research in the use of EEG for control of an output device such as a computer application or a neuro-prosthesis. Motor imagery (MI), the mental process by which an individual rehearses or simulates an action induces changes in EEG signals. EEG based classification of MI holds promise for development of pervasive Brain Computer Interface (BCI). This paper discusses work undertaken at Tezpur University in pursuit of this goal.Most of the current BCI systems are developed based on traditional signal processing techniques under the assumption that the signal is Gaussian and has linear characteristics. However, MI related EEG is highly non-Gaussian, non-stationary and non-linear. We have undertaken experiments to explore the use of bispectrum of EEG in estimation of MI. Bispectrum analysis is presented to analyze EEG signals recorded during imagination and observation of hand movements. EEG signals are recorded from primary hand areas i.e. from electrode position C3 and C4. Our preliminary investigation, have shown the ability of bispectrum in detecting non-linear phase coupling between alpha and beta rhythms during observation and imagination of hand movement. Bispectrum analysis of EEG provides a way to efficiently discriminate MI.