Connectivity of high-frequency bursts as SOZ localization biomarker

Area of Science:

• Neuroscience
• Biomedical Engineering
• Epilepsy Research

Background:

• Accurate localization of the seizure onset zone (SOZ) is critical for successful epilepsy surgery.
• Existing biomarkers like high-frequency oscillations (HFOs) and connectivity lack standardized estimation methods and reliability assessments.
• There is a need for robust methods to quantify inter-electrode synchronization for SOZ identification.

Purpose of the Study:

• To propose and validate channel-level connectivity dispersion (CLCD) as a novel metric for SOZ localization.
• To develop a specialized filtering technique for robust connectivity estimation in intracranial EEG (iEEG).
• To assess the association between CLCD and surgical outcomes in refractory epilepsy patients.

Main Methods:

• Developed a specialized filter to mitigate artifacts in iEEG signals.
• Created binary signals of HFOs per channel to calculate pairwise connectivity.
• Calculated CLCD by measuring the variability of combined connectivity values across electrodes.

Main Results:

• CLCD effectively identified SOZ channels, showing lower values compared to non-SOZ channels across two independent datasets (ROC-AUC 0.73-0.78).
• The proposed filtering method ensures connectivity estimates are robust to common iEEG artifacts.
• Higher CLCD interquartile range correlated with favorable surgical outcomes (Engel I), suggesting prognostic value.

Conclusions:

• CLCD is a reliable and robust metric for identifying SOZ clusters in refractory epilepsy.
• The developed method offers an improved approach to analyzing HFO connectivity for clinical application.
• CLCD shows potential as an adjunctive tool for surgical planning and predicting outcomes in epilepsy.