Abstracts

Rationale: Mood disorder is a common comorbidity in people with temporal lobe epilepsy (TLE), but the functional underpinnings are unclear. Given that epilepsy is a network disorder, it is likely that changes in the network involved in emotional processing contribute to this burdensome comorbidity. This study used graph theory metrics of functional connectivity to test the hypothesis that individuals with TLE would exhibit changes in the brain network crucial for emotional processing. Methods: Functional MR images (TR 1490 ms, TE 60 ms) using a dynamic fearful faces paradigm and T1-weighted images (TR 8.4 ms, TE 3.7 ms) were obtained in 26 individuals with pharmacoresistant TLE (Left = 16; Right = 10) and 20 healthy controls. Functional images were preprocessed with SPM8 (Wellcome Department of Imaging Neuroscience, London, UK), generating maps of activation using a 1 sample t-test for controls, right TLE (RTLE), and left TLE (LTLE) (voxel-wise threshold p < 0.001, extent threshold 5). Peak values of significant clusters were used to localize brain regions involved in fearful face processing, and regions involved in any of the groups were combined into a set of masks created using the AAL atlas (WFU Pickatlas, http://fmri.wfubmc.edu/software/PickAtlas), and weighted adjacency matrices were generated from the correlation of mean time series from each region. We focused on the hypothesis of altered connectivity in the network critical for emotional processing by evaluating global network properties (efficiency, cluster coefficient, betweenness centrality) and the strength of association between the activated regions. Graph measures were calculated using Brain Connectivity Toolbox (https://sites.google.com/site/bctnet/). Significance threshold for global properties was set at p < 0.017 (= 0.05 / 3), and 0.005 uncorrected for strength of connections. Results: Connection strength between emotional salient regions were reduced in the right hemisphere for both RTLE and LTLE, compared to controls (Table). In RTLE, the reductions were predominantly between cortical and subcortical regions (Figure 1A), whereas in LTLE, the attenuation was among cortical to cortical regions (except linkage between amygdala and middle temporal gyrus; Figure 1B). No regions showed increased connection strength in TLE compared to controls. No differences in global network metrics were found between controls and individuals with TLE, although cluster coefficient approached significance (p = 0.047; LTLE >