Classifying High-Frequency Oscillations by Morphologic Contrast to Background, With Surgical Outcome Correlates

Area of Science:

• Neuroscience
• Epileptology
• Biomedical Engineering

Background:

• High-frequency oscillations (HFOs) in intracranial EEG are vital for identifying seizure onset zones.
• Distinguishing pathologic from physiologic interictal HFOs remains a challenge.

Purpose of the Study:

• To develop and validate a novel method for classifying HFOs based on their morphologic contrast to background EEG.
• To improve the identification of epileptogenic zones using interictal HFOs.

Main Methods:

• Retrospective analysis of intracranial EEG data from 13 epilepsy patients.
• Automated event detection, morphologic feature extraction, and k-means clustering incorporating background contrast.
• Identification of "hotspots" of pathologic HFOs and comparison with seizure onset zones.

Main Results:

• Clustering using contrast features enhanced group separation and boundary consistency.
• A significant correlation was found between matching HFO hotspots and seizure onset zones.
• Patients with matching hotspots showed significantly better surgical outcomes (Engel I/II) compared to those with mismatches.

Conclusions:

• HFOs with higher contrast to background are likely indicative of pathologic activity.
• Accurate identification of pathologic HFO hotspots is crucial for predicting surgical success in epilepsy patients.