Detecting seizure onset in the ambulatory setting: demonstrating feasibility

Area of Science:

• Biomedical Engineering
• Neuroscience
• Digital Signal Processing

Background:

• Ambulatory electroencephalogram (EEG) recorders are widely available for patient monitoring.
• Current ambulatory EEG devices primarily capture and store data, lacking real-time seizure detection capabilities.
• Existing real-time detectors offer no guarantee on detection timing, limiting their use for immediate therapeutic interventions.

Purpose of the Study:

• To integrate a patient-specific seizure onset detector into a commercially available ambulatory EEG recorder.
• To enable real-time seizure onset detection in the ambulatory setting.
• To assess the feasibility of running seizure detection algorithms on resource-limited digital signal processors within ambulatory recorders.

Main Methods:

• Development and integration of a patient-specific seizure onset detection algorithm.
• Implementation of the algorithm onto the digital signal processor of a commercial ambulatory EEG recorder.
• Testing and validation of the integrated system for ambulatory seizure onset detection.

Main Results:

• Successful integration of a seizure onset detector into an ambulatory EEG recorder.
• Demonstration of the system's capability to detect seizure onset in the ambulatory setting.
• Proof of concept for running advanced detection algorithms on embedded hardware.

Conclusions:

• Integrating seizure onset detectors into ambulatory EEG recorders is feasible.
• This integration can enable real-time seizure detection, paving the way for alarms and therapies.
• The developed system holds potential for acute seizure management in ambulatory patients.