Edge-centric functional network organization reveals seizure pathways in human epilepsy

Spatiotemporal signal propagation through the human brain during epileptic seizures remains poorly understood. Using intracranial EEG from subjects with medically refractory epilepsy, we introduce an edge-centric framework that tracks how influential functional connections in the time-evolving epileptic brain network form and transition as seizures unfold. Brain regions are grouped into three functional modules – seizure onset area, nearby ipsilateral regions, and other areas – with the most important edges followed over time. A probabilistic entropy approach reveals that path-based centralities exhibit a conserved circular motif of seizure propagation across subjects. In contrast, strength-based centralities separate seizures into two distinct classes: one where cross-module edges serve as dynamic bottlenecks guiding ictal spread, and another where these edges remain stable. These findings demonstrate that an edge-focused perspective may reveal vital information about ictal communication pathways, and may have potential application for improved prediction and classification in epilepsy.