Phase Lag Index in a EEG Based Brain Computer Interface Paradigm: A Comparative Analysis on Three Motor Imagery Datasets

A comparative analysis on three motor imagery datasets is performed using the feature extraction phase lag index method based on phase synchronization. Support vector machine classifier was used to classify the extracted features and Kappa coefficient, F1-score and Matthews coefficient correlation evaluated the classifier performance. Discrimination between mental tasks (right hand motor imagery, left hand motor imagery and feet motor imagery) in the frequency band of Mu rhythm led to classification rates above 90% and performance metrics values above 0.8. The results obtained suggest that phase lag index can be used in brain computer interface paradigms based on upper and lower limb for patients with motor disabilities.