Publication

Abstract : Autonomic dysfunction is often associated with seizures in patients with epilepsy. Autonomic dysfunction during seizures can cause serious events like cardiac arrest and sudden unexplained death. Early detection of autonomic dysfunction is crucial to avoid such medical emergencies. In our study, we analyzed the heart rate variability (HRV) in physiological recording during the preictal, ictal and postictal periods. A total of 142 seizures with various semiologies recorded from 49 epilepsy patients were included. Time-domain measurements including mean hear rate (HR), RR interval (RRI), root-mean-square of successive R-R interval differences (RMSSD), The standard deviation of N-N intervals (SDNN) and percentage of successive RR intervals that differ by more than 50 ms(pNN50) were analyzed at different time points, 20 minutes before the seizure, 1 minute before the seizure, during the seizure and 20 minutes after the seizure. We observed there was a significant difference in HRV parameters during and just before the seizure onset (RMSSD with p-value<0.001 and SDNN with p-value<0.001). We also observed that there was a suppression of parasympathetic activity and activation of sympathetic stimulation in the ictal and peri-ictal period, which was characterized by an elevation in the mean heart rate (HR), whereas the mean RRI showed an inversely proportional trend. There was a reduction in mean RMSSD, SDNN and pNN50 prior to seizure onset and was more pronounced in generalized seizures as compared to focal seizures with preserved or impaired awareness. There was a greater sympathetic activation and parasympathetic inhibition during an epileptic seizure in females as compared to male patients. These findings suggest that autonomic dysregulation was increased during a generalized seizure when compared to a focal seizure. HRV analysis can be used as a valid method for quantifying central influences on the autonomic nervous system and its cardiac control. A major advantage of electrocardiogram (ECG)-based seizure detection is that the ECG is an essentially easier signal to obtain, with a higher signal-to-noise ratio than EEG (electroencephalogram). Based on the inferences from this study future studies could be done to build machine learning models which could form the base for the development of wearable seizure prediction devices.