Abstracts

Rationale: We previously demonstrated that children with new onset epilepsy already have an altered large-scale brain structural network (Bonilha et al, HBM 2014). However, the natural history and development of further abnormalities in global brain topology in the context of ongoing active epilepsy has not been studied. Here we investigated the prospective alterations in global brain structure network over the first two years after epilepsy onset among children with idiopathic generalized epilepsy (IGE), localization related epilepsy (LRE), and healthy controls. Methods: We studied a group of 28 healthy controls (mean age 13.3 ± 3.28 years, 11 males), 21 new-onset LRE (11.6 ± 2.68 years, 12 males) and 18 new-onset IGE (15 ± 3.3 years, 7 males). Patients and controls were similar in age (p=0.92) and gender (p= 0.4). The mean age of LRE group was younger than the IGE groups (p= 0.005) but IGE and LRE groups did not significantly differ from controls (p= 0.09 and p=0.07, respectively). All subjects underwent MRI scanning utilizing the same imaging protocol yielding SPGR images (1.5 Tesla GE Signa scanner, TR = 24 ms, TE = 5 ms, Slice thickness 1.5mm). Cortical thickness data derived from FreeSurfer ‘Destrieux Atlas' were utilized for the construction of group-wise adjacency matrices (controls, LRE, IGE) where each of the 148x148 entries represented the partial correlation between the thickness from each pair of regions, controlled for age, bootstrapped 1000 times. The adjacency matrices were then utilized for the evaluation of global graphical properties (Global Efficiency, Betweenness Centrality, and Clustering Coefficient) across fixed density threshold binary matrices (ranging from 0.15 to 0.50) using of the Brain Connectivity Toolbox. Bonferroni correction was made to account for multiple comparisons (0.05/[(148x148 regions) x (8 fixed density thresholds) x (1000 bootstraps) x (3 groups) x (2 time points) = 4.75e-11. Results: All groups demonstrated prospective changes in the topology of the structural network, indicating neurodevelopmental changes over the two years. However, IGE and LRE groups showed greater clusters of increased and decreased correlations compared to controls (Figure 1). Global network metrics demonstrated that over the prospective two years, the IGE showed the most dramatic differences compared to controls with less segregation (clustering coefficient), greater centrality, reduced efficiency and small worldness (p<0.001, corrected; Figure 2). The LRE group showed increased clustering coefficient, decreased centrality, efficiency and small worldness, compared to controls (p<0.001 corrected; Figure 2). Conclusions: Children with new-onset epilepsy demonstrated an actively evolving altered large-scale neurodevelopmental trajectory over the prospective 2 years. Syndromic signatures emerged with IGE showing less segregation and LRE demonstrated greater segregation over time compared to controls, leading to abnormal neurodevelopmental structural trajectories with less network efficiency and less optimal small world configuration.