Automatic ictal HFO detection for determination of initial seizure spread

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• High-frequency oscillations (HFOs) in electrocorticography (ECoG) are crucial biomarkers for identifying the seizure onset zone (SOZ) in epilepsy.
• Current methods for HFO detection often rely on visual analysis, which can be time-consuming and subjective.

Purpose of the Study:

• To introduce a novel, automated method for detecting ictal HFOs in the ripple band (80-250 Hz) using CFAR matched sub-space filtering.
• To enable the tracking of early HFO propagation and identify initial and follow-up epileptic activity.

Main Methods:

• Development and application of a novel CFAR matched sub-space filtering technique for automatic ictal HFO detection.
• Analysis of ECoG recordings from two seizures in a patient with focal epilepsy.
• Comparison of automated detection results with clinical visual HFO analysis and conventional v-activity for SOZ correlation.

Main Results:

• The proposed method successfully detected ictal HFOs in the ripple band.
• The identified electrodes showed strong agreement with clinician-based visual HFO analysis.
• Electrodes exhibiting initial HFO activity were well-correlated with the seizure onset zone (SOZ) identified by conventional v-activity.

Conclusions:

• The novel CFAR matched sub-space filtering method provides an accurate and automated approach for detecting ictal HFOs.
• This technique facilitates the precise localization of the SOZ by tracking early epileptic activity.
• The findings support the utility of automated HFO detection for epilepsy diagnosis and surgical planning.