Classification of two-class motor imagery EEG signals using empirical mode decomposition and hilbert–huang transformation

Abstract

This work proposes classification of two-class motor imagery EEG signals using instantaneous frequencies as the feature, which is extracted from intrinsic mode functions (IMFs) using Hilbert–Huang transformation (HHT). IMF is computed by using Empirical Mode Decomposition (EMD). The classification of two-class motor imagery is prerequisite for brain–computer interface (BCI). In this work, the EMD method is used for identification of the right-hand and feet imagery movements. The proposed method has been applied on the subjects 05 and 06, which are publicly available on BNCI-horizon-2020. Upon investigation of the performance of the proposed feature extraction method, the following results are obtained; we found that the BFTree turned out to be the best classifier with 59.375% accuracy with subject 05 and DecisionStump classifier with 58.75% accuracy with subject 06. From the above results, it is evident that the proposed method is reliable for classification of two-class motor imagery EEG signals.