Multiscale detrended cross-correlation of EEG and RR intervals during focal epilepsy

Abstract

To evaluate the interaction between epilepsy-related brain activities and heart rate dynamics, we analyze electroencephalogram (EEG) and heart rate variability (HRV) using detrended moving-average cross-correlation analysis (DMCA) for pre- and postictal periods in subjects with focal epilepsy. The DMCA was applied to the 5 min. long periods of heart beat-to-beat intervals and power spectral density time series, located 5 and 10 min. before and after seizures. Statistically significant differences were found in the cross-correlation in δ (0.5–4 Hz) band for the periods before and after seizure, which shows the correlative coupling between RR intervals that δ band activity is changing while approaching to and after the epileptic seizure, that suggests the presence of nonlinear mechanisms of interactions between low-band EEG and RR intervals in observed periods of brain and heart activity in epilepsy. In contrast, no statistically significant changes could be observed while comparing brain-heart coupling in preictal periods. The wide scatter of distributions of all frequency bands in periods before epileptic seizure suggests that the multiscale cross-correlation coefficient is highly subject dependent and needs further subject-specific analysis for this particular period with longer time series.