Abstracts

Rationale: Functional connectivity changes in and around the epileptic zone (EZ) in patients with focal epilepsy are still unknown. Although many researches in focal epilepsy have demonstrated increased pattern of connectivity surrounding the EZ (Morgan VL,2015, Negishi M, 2011), some have included contradictory observations (Doucet G, 2013). We hypothesized increased regional connectivity in EZ as well as some propagation areas, which could be identified by “hub analysis”, a novel method to analyze functional connectivity using resting-state fMRI (rsfMRI), with comparison to large normal cohort datasets. Methods: Eighteen patients with medically uncontrollable focal epilepsy were included in this study. EZ was identified with presurgical routine checkup, and resection surgery was performed in all patients. Out of 18 patients, 17 cases demonstrated MRI positive lesions including tumors, cortical dysplasia, vascular malformations, and hippocampal sclerosis. Preoperative rsfMRI was taken with a 3T MRI scanner (TR = 2.5s, 152 volumes). After preprocessing, the voxel-based degree, defined as the number of voxels with time courses that are strongly correlated with that of the given voxel (r>0.75), was computed. Compared to datasets of normal healthy cohorts (n=83), the cluster of voxels with degree z-score greater than 3 was defined as “hub” (hub analysis). Accordingly, hub demonstrated the region which had strong connectivity to non-specific areas. The location, T-score, and size of each cluster were evaluated and compared to presumed EZs, surgical resected areas, and surgical outcomes. Results: In 16/18 cases (88.9%), obvious hubs were seen in and around EZ (EZ-hubs; mean T-score = 9.33). Postoperatively, epilepsy disappeared in 14 cases and remained in 4 cases. EZ-hub was included in resection area in 11/18 cases, demonstrating a trend to correlate to good surgical outcome, but the correlation was not statistically significant (P=0.198). Mirror hubs were also observed in the contralateral side in 15/18 cases and were stronger than EZ-hubs. Interestingly, laterality indices (LI; EZ hub – mirror hub / EZ hub + mirror hub) were significantly high in good outcome group (P < 0.01). LI cut-off values were -0.25 for T-score and -0.13 for cluster size. Other hubs were also seen in symptomatic zone and presumed propagated areas (propagation hubs). Propagation hubs in ipsilateral frontal and temporal lobes were significantly bigger and stronger than those in the contralateral side. Conclusions: Hubs may represent regions with pathologically strong connections related to epilepsy, and these results supported the hypothesis for increased regional connectivity in EZ as well as some propagation areas. Relatively strong and big EZ-hub may predict good surgical outcome. EZ-hubs may not always appear inside the EZ, but in adjacent areas, therefore resection of those may not be required. Funding: This research was supported by JSPS KAKENHI grant number JP17K10890, and the Japanese epilepsy research foundation.