Bispectrum analysis of EEG for motor imagery classification

Effective motor imagery (MI) classification based on electroencephalogram (EEG) signals for Brain Computer Interface (BCI) is an active area of research. Classification is largely dependent on the feature vector and the type of classifier. This paper reports a study on the use of bispectrum for classifying left and right hand MI based on surface EEG from electrode positions C3 and C4. EEG signals from publicly available dataset has been taken. Derived as well as statistical features of the bispectrum are explored. A Least Square Support Vector Machine (LS-SVM) classifier is used. Experimental results support the idea of using bispectrum for left and right hand MI classification.