Classification of EEG signals during Working- Memory Maintenance based on Phase Space Reconstruction of Empirical Mode Decomposition

Electroencephalogram (EEG) signal is a conventional method to detect the cognitive function of the brain. In this paper, we have proposed a feature selection approach based on empirical mode decomposition (EMD) and phase space reconstruction (PSR), which is used to carry out the classification of different material during working memory processing process. We calculated the 2D-PSR of intrinsic mode functions (IMFs) of EEG signals from five channels in the visual cortex based on the data set of six people. New features based on the central tendency measure (CTM) of 2D-PSR have been proposed for classification. By systematically comparing the effects of three classifiers, including Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Random Forest (RF). The results show that the classification accuracy obtained on the task of memorizing picture materials is 66.3%±7.3% and the one of characters materials is 63.2%±9.5%. This work is the first experimental provenance to show that the CTM features of PSR after EMD are the effective features of memory processing maintenance tasks based on working memory materials among single subject.