336 Connectivity Between Resting-State Networks and Mesial Temporal Regions May Distinguish Between Right and Left Mesial Temporal Lobe Epilepsy

Abstract

INTRODUCTION: Mesial temporal lobe epilepsy (mTLE) is the most common surgical epilepsy. Confidently lateralizing mTLE preoperatively may be challenging with only noninvasive evaluations. Resting-state fMRI has been used to identify connectivity abnormalities in mTLE, involving mesial temporal structures and other networks such as default mode network (DMN). METHODS: We acquired 20 minutes of resting-state fMRI in 52 mTLE patients (28 females, 39.5 ± 12.2 years) and 52 age/gender matched controls. In mTLE patients versus controls, we evaluated fMRI functional connectivity of bilateral hippocampi/amygdalae and resting-state networks (default mode, central executive, salience, visual) using non-directed Pearson correlation (Z-scored), network-based statistic (NBS), and directed dynamic causal modeling (DCM). We separately acquired 2 minutes of resting-state SEEG in 17 unilateral mTLE patients (10 females, 34.6 ± 12.5 years). Using SEEG, we measured hippocampal versus extrahippocampal inward partial directed coherence (PDC) strength. RESULTS: Right mTLE patients demonstrated reduced undirected functional connectivity (Z) between right hippocampus and DMN compared to left mTLE patients and controls (p < 0.001, corrected, ANOVA). These were not seen between the hippocampus and other resting-state networks. Focusing on mesial temporal and DMN nodes, NBS analysis demonstrated the network with most decreased connectivity in right versus left mTLE included right hippocampus/amygdala, bilateral inferior parietal, precuneus, and medial orbitofrontal, and right lateral orbitofrontal regions (9 of 14 nodes, p = 0.006, corrected). With model exceedance probabilities, DCM analysis revealed cross-hemispheric connectivity between hippocampi/amygdalae was mostly inward to epileptogenic-side. FMRI finding resembled SEEG findings, where inward PDC strength was higher in epileptogenic hippocampi compared to other regions in right/left mTLE (p < 0.05, t-tests, corrected). CONCLUSION: If improved, this may aid surgical planning and reduce need for intracranial monitoring. Inward patterns using fMRI and SEEG appear altered at epileptogenic hippocampi, warranting further comparative study of directional connectivity between fMRI and SEEG.