Preictal Time Assessment using Heart Rate Variability Features in Drug-resistant Epilepsy Patients

Area of Science:

• Clinical Neuroscience
• Biomedical Engineering
• Computational Biology

Background:

• Epileptic seizures are linked to electroencephalogram (EEG) changes and autonomic nervous system (ANS) influence on electrocardiogram (ECG) traces.
• Differences exist between interictal (normal) and preictal (seizure-preceding) periods in biosignals, but a clinically defined preictal state is lacking.
• Preictal period characteristics vary among patients and even seizures within the same patient.

Purpose of the Study:

• To investigate the existence of a seizure-specific preictal interval using heart rate variability (HRV) features from ECG data.
• To determine if HRV can differentiate between the preictal and interictal states in epilepsy patients.

Main Methods:

• Extracted time and frequency domain HRV features (linear and non-linear) from ECG data of 37 drug-resistant epilepsy patients (EPILEPSIAE database).
• Analyzed 209 temporal lobe seizures.
• Employed a linear discriminant analysis classifier to inspect the transition period before seizure onset.

Main Results:

• A combination of RRMean, NN50, and SD2 features achieved the highest accuracy (88.04%±12.30%) in discriminating between the farthest interictal and nearest preictal 50-minute intervals.
• Demonstrated the ability of specific HRV features to identify a transition period preceding seizures.

Conclusions:

• The study provides evidence for a seizure-specific preictal interval identifiable through HRV analysis.
• These findings suggest that HRV analysis holds potential for characterizing and potentially predicting the preictal state in epilepsy.