Evaluation of novel algorithm embedded in a wearable sEMG device for seizure detection

Area of Science:

• Biomedical Engineering
• Neuroscience
• Clinical Neurology

Background:

• Generalized tonic-clonic seizures (GTCS) are a major concern in epilepsy management.
• Accurate and timely detection of GTCS is crucial for patient safety and treatment efficacy.
• Existing detection methods may have limitations in portability and real-time alerting.

Purpose of the Study:

• To develop and evaluate a prototype wireless surface electromyography (sEMG) device for detecting generalized tonic-clonic seizures.
• To assess the performance of a modified, computationally efficient seizure detection algorithm integrated into the sEMG device.
• To determine the feasibility of real-time seizure alarms generated by the device in a clinical setting.

Main Methods:

• A modified algorithm for generalized tonic-clonic seizure detection was implemented in a prototype wireless sEMG device.
• The algorithm was optimized for minimal computational load and trained using prior sEMG data.
• The device was tested in a double-blind study involving five patients at an Epilepsy Monitoring Unit (EMU), with alarms annotated in the data.

Main Results:

• The prototype device successfully detected 4 out of 7 generalized tonic-clonic seizures during the study.
• The device exhibited a low false detection rate of 0.003 per hour, equivalent to one false alarm every twelve days.
• The system demonstrated the capability to provide real-time seizure alarms integrated with sEMG recordings.

Conclusions:

• The developed wireless sEMG device shows promise for the detection of generalized tonic-clonic seizures.
• The computationally efficient algorithm and low false alarm rate support its potential utility in epilepsy monitoring.
• Further validation in larger patient cohorts is warranted to confirm its clinical applicability and reliability.