An automatic patient-specific seizure onset detection method using intracranial electroencephalography

Area of Science:

• Neurology
• Biomedical Engineering
• Signal Processing

Background:

• Epilepsy diagnosis relies heavily on electroencephalography (EEG) analysis.
• Accurate and timely seizure detection is crucial for patient management and treatment planning.
• Current automated seizure detection methods face challenges with accuracy and robustness.

Purpose of the Study:

• To develop and evaluate a novel multichannel, patient-specific seizure detection system.
• To leverage empirical mode decomposition (EMD) for feature extraction from intracranial EEG.
• To utilize a support vector machine (SVM) classifier for discriminating seizure and non-seizure epochs.

Main Methods:

• Intracranial EEG data from 17 patients (463 hours, 51 seizures) were analyzed.
• EMD was employed to extract relevant features from EEG signals.
• A postprocessing algorithm was implemented to enhance robustness and reject artifacts.

Main Results:

• The method achieved an average sensitivity of 92%.
• A low false detection rate (FDR) of 0.17/hour was recorded.
• A time delay (TD) of 12 seconds was observed, with potential for FDR reduction via TD extension.

Conclusions:

• The patient-specific seizure detection method demonstrates high performance, aiding clinical staff in automated seizure marking.
• The system offers potential for online seizure onset detection with high accuracy.
• This tool can facilitate early and precise seizure detection, supporting epilepsy intervention planning.