Time-variant coherence between the heart rate variability and the channel-related envelopes of the EEG delta activity was used as an indicator for interactions between the autonomic nervous system and cortical processes before, during and after epileptic seizures. The tensor decomposition was applied to explore the topography-time-frequency characteristics of these correlative interactions for each patient and for two sub-groups (left and right hemispheric seizure). It can be demonstrated that tensor decomposition strongly supports the interaction analysis and its interpretation.