Multivariate variational mode decomposition & phase space reconstruction based motor imagery EEG classification.

•A novel method for feature extraction of Motor Imagery EEG signals is developed.•The method is based on Multivariate Variational Mode Decomposition and Phase Space Reconstruction.•Multiple statistical and non-liner features are extracted.•The method is evaluated on BCI competition-IV(2a) dataset.•Both binary and multi-class motor imagery EEG signals are classified.•An average classification accuracy of 96.76% and 84.22% in obtained in binary and multi-class classification respectively.