Interictal Epileptogenic Zone Localization using Neural Fragility in Simulated Electroencephalogram Data.

Area of Science:

• Computational neuroscience
• Epilepsy research
• Biomarker discovery

Background:

• Epilepsy affects 50 million globally, with one-third having drug-resistant epilepsy (DRE).
• Accurate localization of the epileptogenic zone (EZ) is crucial for successful epilepsy surgery.
• Current EZ localization methods require invasive intracranial EEG (iEEG) monitoring and prolonged hospitalization.

Purpose of the Study:

• To evaluate neural fragility as a computational biomarker for identifying epileptogenic nodes in DRE.
• To assess the predictive accuracy of neural fragility in localizing seizure-generating regions using in-silico data.
• To explore advanced computational methods for improving epilepsy surgical planning.

Main Methods:

• Utilized in-silico data from a phenomenological network model with pre-defined EZs.
• Assessed neural fragility, a dynamical network-based metric, on resting-state iEEG data.
• Employed threshold-based classification to identify epileptogenic nodes based on fragility scores.

Main Results:

• Neural fragility demonstrated a bimodal distribution of scores.
• Threshold-based classification accurately identified epileptogenic nodes in 45% and 54% of simulations across two datasets.
• Variability in predictions indicates a need to investigate factors influencing accuracy.

Conclusions:

• Neural fragility shows potential as a computational biomarker for EZ localization in DRE.
• Further research is needed to optimize network models and validate findings with clinical iEEG data and surgical outcomes.
• This work contributes to advancing computational methods for more precise epilepsy surgical planning and improved patient outcomes.