Node degree volatility for seizure onset zone localization

Drug-resistant epilepsy often requires invasive treatment, yet biomarkers for localizing the seizure onset zone remain limited. Here, we introduce node degree volatility, a dynamic network biomarker defined as the rate of change in a node’s functional connectivity over time. We analyzed 82 seizures from 80 patients by reconstructing Granger-causal networks from intracranial electroencephalography, where nodes represent recording contacts and edges represent directed predictive relationships. Nodes with the highest degree volatility at seizure onset identified the clinically annotated seizure onset zone in 75% of successful surgical cases, using a stringent top-quantile threshold. Under a uniform ranking framework, this biomarker showed higher seizure onset zone concordance than benchmark implementations of neural fragility and common graph-theoretic metrics, demonstrating the value of capturing rapid local dynamics in functional brain networks. This metric is computationally efficient, interpretable, and readily deployable in clinical workflows. More broadly, node degree volatility provides a generalizable framework for identifying dynamically critical nodes in functional neurophysiological networks and may extend to other neurological disorders and recording modalities.