A new coupled PARAFAC2 decomposition for joint processing of somatosensory evoked magnetic fields and somatosensory evoked electrical potentials

The application of tensor-based signal processing in EEG (Electroencephalogram) and MEG (Magnetoencephalogram) data analysis has received growing research interest. In particular, coupled tensor decompositions are regarded as a promising tool for the joint processing of simultaneously recorded EEG and MEG signals. Therefore, we develop a coupled version of the PARAFAC2 decomposition that has been employed for multi-way component analysis of EEG or MEG data separately. Then we discuss the application of this new coupled PARAFAC2 decomposition in the combined analysis of somatosensory evoked magnetic fields (SEFs) and somatosensory evoked electrical potentials (SEPs) to extract the time, frequency, and spatial features of these signals.