Automated Video Detection of Epileptic Convulsion Slowing as a Precursor for Post-Seizure Neuronal Collapse

Area of Science:

• Biomedical Engineering
• Neuroscience
• Epilepsy Research

Background:

• Automated seizure monitoring enhances safety and quality of life for epilepsy patients.
• Clonic slowing at seizure end correlates with postictal generalized EEG suppression (PGES).
• Prolonged PGES is linked to increased risk of sudden unexpected death in epilepsy (SUDEP).

Purpose of the Study:

• Develop an automated, remote video sensing algorithm for real-time detection of clonic slowing.
• Utilize clonic slowing detection to provide alerts for potentially dangerous PGES periods.
• Contribute to the prevention of SUDEP through early warning systems.

Main Methods:

• Algorithm based on optical flow video processing and time-frequency wavelet spectrum analysis.
• Integral Radon-like transformation applied to detect log-linear frequency changes (clonic slowing).
• Validation against manually processed electroencephalography (EEG) data from 48 convulsive seizures.

Main Results:

• Automated detection of clonic slowing showed high correlation with manually identified interictal clonic intervals (ICIs) from EEG.
• Quantification of slowing via dominant angle in Radon transformed spectrum significantly correlated with ICI increase factors.
• Spectral ridges from Gabor-wavelet transformations closely matched manual ICI traces.

Conclusions:

• The developed algorithm reliably detects and quantifies clonic slowing, a potential precursor to PGES.
• This automated method offers a promising approach for an efficient alerting device to mitigate SUDEP risk.
• Further validation could establish this as a critical tool in epilepsy management and SUDEP prevention.