Absence Seizure Detection Algorithm for Portable EEG Devices.

Area of Science:

• Epilepsy research
• Biomedical signal processing
• Neurology

Background:

• Absence seizures, common in childhood and juvenile absence epilepsy, manifest as generalized nonmotor seizures with abrupt onset and termination.
• Characterized by transient consciousness impairment and EEG spike-slow wave discharges (SWDs), these seizures pose significant challenges.
• Portable EEG devices offer potential for long-term, remote patient monitoring, aiding diagnosis and treatment management.

Purpose of the Study:

• To develop and validate a novel algorithm for detecting absence seizures using continuous wavelet transform (CWT) of SWDs.
• To assess the algorithm's performance in terms of sensitivity, detection rate, and overlap with actual seizures.
• To evaluate the algorithm's feasibility for real-time seizure detection on mobile devices.

Main Methods:

• A novel absence detection algorithm was developed based on the complex Morlet continuous wavelet transform of SWDs.
• The algorithm was trained and tested on EEG data from 64 patients with absence seizures and 30 healthy controls.
• Performance was evaluated using two bipolar EEG channels (Fp1-T3, Fp2-T4) with varying seizure duration thresholds.

Main Results:

• The algorithm achieved 97.6% sensitivity with a 0.7/h detection rate for seizures >2s.
• False detections were linked to non-clinical epileptiform discharges; raising the threshold to >3s reduced the false detection rate to 0.5/h.
• Automatic seizure detection showed ~96% overlap with actual seizures, and processing speed was suitable for real-time smartphone application.

Conclusions:

• The developed algorithm effectively detects absence seizures using CWT of SWDs.
• Real-time detection is feasible on smartphones, supporting remote monitoring for childhood and juvenile absence epilepsy.
• This technology can facilitate diagnosis, personalized drug titration, and pharmacotherapy duration management.