A machine-learning algorithm for detecting seizure termination in scalp EEG.

Area of Science:

• Neurology
• Biomedical Engineering
• Machine Learning

Background:

• Automated detection of seizure cessation in electroencephalograms (EEGs) is crucial for clinical applications.
• Current methods for determining seizure termination can be time-consuming and subjective.
• Accurate seizure duration estimation is vital for treatment and understanding seizure dynamics.

Purpose of the Study:

• To present and evaluate a novel machine learning-based algorithm for detecting the termination of electrographic seizure activity.
• To assess the algorithm's performance against expert electroencephalographer annotations.
• To explore the utility of the seizure end detector in estimating seizure duration when combined with a seizure onset detector.

Main Methods:

• Development of a machine learning algorithm to identify the electrographic end of seizures from scalp EEG data.
• Testing the algorithm on 133 seizures from a publicly available database.
• Pairing the seizure end detector with a pre-existing seizure onset detector to estimate seizure duration.

Main Results:

• The novel algorithm successfully detected the end of 132 out of 133 seizures.
• Detection occurred within an average of 10.3 ± 5.5 seconds of the electroencephalographer-determined end time.
• When combined with a seizure onset detector, the system estimated the duration of 85% of seizures within a 15-second margin of error.

Conclusions:

• The developed machine learning method provides a robust and accurate approach for detecting electrographic seizure termination.
• This automated tool has the potential to significantly improve clinical applications, including seizure duration estimation and status epilepticus detection.
• The findings support the advancement of automated seizure detection technologies for improved patient care and research.