Automated algorithm for generalized tonic-clonic epileptic seizure onset detection based on sEMG zero-crossing rate

Area of Science:

• Biomedical Engineering
• Neurology
• Signal Processing

Background:

• Generalized tonic-clonic seizures (GTCS) render patients unable to call for assistance.
• Current methods for seizure detection may lack robustness or portability.

Purpose of the Study:

• To develop a generic, robust algorithm for automatic GTCS detection using surface electromyography (sEMG).
• To ensure the algorithm is suitable for implementation in a portable device for real-time patient monitoring.

Main Methods:

• Analysis of 22 GTCS events from 11 patients using sEMG data.
• Implementation of a high-pass filter (150 Hz cutoff) and zero-crossing count with hysteresis (±50 μV).
• Utilized a single sEMG electrode placed on the deltoid muscle.

Main Results:

• Achieved 100% sensitivity in detecting tonic-clonic seizures.
• Mean detection latency was 13.7 seconds.
• Maintained a low false detection rate of 1 false alarm per 24 hours.

Conclusions:

• The developed generic algorithm demonstrates high performance for automatic GTCS detection.
• The algorithm's robustness and low false alarm rate make it suitable for clinical implementation on portable devices.
• This technology has the potential to significantly improve patient safety and care during seizures.