Phase-Sensitive Common Spatial Pattern for EEG Classification
Biswadeep Chakraborty, Saptak Ghosal, Lidia Ghosh, Amit Konar, Atulya K. Nagar · 2019
This paper addresses an interesting problem to model common spatial pattern (CSP) using an objective function employed to segregate EEG signals for a given cognitive task into two classes. The novelty of the present research is to include phase information with amplitude of the EEG signals to differentiate class boundaries. A new formulation of CSP is introduced and solved using Lagrange's multiplier method taking phase information of EEG into account. In addition, the proposed CSP is also realized with Schur Decomposition technique instead of using the conventional eigenvalue decomposition to overcome the disadvantage of using the latter one. Experiments undertaken confirm that the proposed phase-sensitive CSP and the CSP with Schur decomposition yield better performance than their non-phase sensitive counterpart by a large margin with respect to classification accuracy with the best results obtained by former CSP algorithm.