Abstracts

Rationale: Temporal lobe epilepsy (TLE) is the most common focal epilepsy. Structural changes in the mesial temporal lobe (MTL) such as mesial temporal sclerosis predispose to TLE. Previous studies demonstrate altered resting state functional MRI connectivity in TLE patients compared to controls. However, changes in connectivity structure within the TLE epileptic network versus the rest of the brain remain unclear. Here we investigated topographical changes in distances-specific connectivity and graph theoretical degree of direct connectivity in patients with TLE compared to healthy volunteers. Methods: 35 patients with known TLE and 31 age- and gender-matched healthy controls (age range 20- 57) were recruited in the context of the Epilepsy Connectome Project (ECP). Epilepsy patients had no structural MRI abnormalities other than mesiotemporal sclerosis. High resolution anatomical and resting state functional MRI data were obtained using 3T MRI Human Connectome Project (HCP) sequences and were preprocessed using the HCP pipeline to obtain surface-based time series. For each subject, vertex-wise matrices estimating the average strength of distance-specific correlation (DSC) between each vertex and its neighbours were constructed for a series of distance ranges (0-5, 5-10, 10-15, 15-20, 20-30, 30-50, 50-100, 100- 150, and 150-200 mm) using Matlab. Graph theoretical degree of direct connectivity (DDC) was also estimated at each vertex by counting edges with a correlation value above the one best separating a distribution of vertices with direct anatomical connectivity (short-distance connectivity (<7mm) within the primary visual cortex) from a distribution of vertices with indirect anatomical connectivity (between prefrontal and contralateral visual cortices) using signal detection theory for each subject. For each subject, spatial z scoring was then performed to assess topographical distributions of DSC and DDC across the cortex. Group means of Z-scored DSC and DDC were compared between TLE patients and controls using non parametric statistics with Threshold Free Cluster Enhancement (TFCE) as implemented in HCP Workbench. All analyses were thresholded at whole-brain corrected TFCE corrected p<0.05. Results: Compared to healthy controls, patients with TLE showed increased DSC in bilateral MTL and limbic networks, which was especially pronounced at short cortical distances (<10mm) (p<0.05, TFCE corrected, Figure 1). Patients with TLE also displayed increased DSC in bilateral motor and somatosensory cortices, which was especially pronounced at longer distances (150- 200 mm) (p<0.05, TFCE corrected). By contrast, DSC was decreased at all cortical distances in occipital, default network and lateral frontal cortices (p<0.05, TFCE corrected). DDC analysis confirmed increased bilateral limbic and sensorimotor networks connectivity and reduced occipital, default network and lateral frontal connectivity in TLE patients compared to controls (Figure 2). Conclusions: Compared to controls, patients with TLE demonstrated increased short range (<10 mm) connectivity in bilateral MTL and limbic networks and increased long range (150-200 mm) connectivity in bilateral motor and somatosensory cortices. These findings provide functional MRI neuroimaging support for the involvement of local, nearest-neighbor networks in epileptogenesis, and suggest a localizing diagnostic value of functional MRI short range connectivity for TLE. Future studies will correlate DSC and DDC with individual seizure focus location and cognitive function. Funding: NIH U01NS093650 (Epilepsy Connectome Project)