Parallel space-time-frequency decomposition of EEG signals for brain computer interfacing

The presented paper proposes a hybrid parallel factor analysis-support vector machines (PARAFAC-SVM) method for left and right index imagery movements classification. The spatial-temporal-spectral characteristics of the single trial electroencephalogram (EEG) signal are jointly considered. Within this novel scheme, we develop a parallel EEG space-time-frequency (STF) decomposition in μ band (8-13 Hz) at the preprocessing stage of the BCI system. Using PARAFAC, we elaborate two distinct factors in μ band for each EEG trial. SVM classifier is utilised to classify the spatial distribution of the movement related factor. This factor is distinguished by its spectral, temporal, and spatial distribution.