Abstracts

Rationale: In the past decade, the notion that epilepsy is a network disorder was prevailing. Understanding the spatiotemporal propagation profiles is the inherent requirement for a deep understanding of the seizure mechanism, which could also facilitate the localization of the epileptogenic zone. Much of what we know about the epileptogenic network has been gleaned from intracranial studies that relied on clinical observation from epilepsy phenotypes (temporal epilepsy, frontal epilepsy, insula epilepsy, etc.) and anatomical labeling. However, the spatiotemporal dynamics of seizure activity within (intra) and across (inter) the resting-state intrinsic networks remain almost entirely unknown. To address this topic, we collected information from a large number of recording sites in a group of human subjects implanted with SEEG.

Methods: We have collected 78 seizures from 41 subjects with SEEG implantation. The ictal data were analyzed as follows, in each electrode, we calculated the anatomic localization (Branvisa), the intrinsic network labeling (SPM, Matlab),  the epileptogenicity index (EI), the propagation time, and the energy ratio (ER) of high-band to low-band frequency (https://github.com/babaoriley6767/Chao_SEEG). The data were pooled together in the R program for the group-level analysis.

Results: We recorded 2476 sites in the YEO7 networks and found that the EI value was higher within the network than across the network (p < 0.001). When comparing intra-network versus inter-network, seizures propagated faster (p < 0.001) and the ER was stronger (p < 0.001). When considering Euclidean distance, we found the significant differences in EI and propagation time between intra-network and inter-network were limited to 60 mm (p < 0.001, respectively). Intra-network ER was higher than inter-network ER within a 30 mm distance (p < 0.001).