Mental Workload Classification Method Based on EEG Cross-Session Subspace Alignment

Electroencephalogram (EEG) signals are sensitive to the level of Mental Workload (MW). However, random non-stationarity EEG will lead low accuracy and a poor generalization ability for cross-session MW classification. To solve this problem different marginal distribution in time periods, an classification method based on Cross-Session Subspace Alignment (CSSA) is presented identify induced visual manipulation tasks. The Independent Component Analysis (ICA) used obtain Components (ICs) labeled unlabeled signals. energy features ICs extracted as source domains target domains, respectively. distributions subspace base vectors aligned with linear mapping. Kullback–Leibler (KL) divergences between two calculated select approximately similar transformed subspace. all selected trained build new classifier using Support Vector Machine (SVM). Then it can realize signals, has good accuracy.