Abstracts

Rationale: Cognitive and psychomotor slowing is a common neuropsychological morbidity in both children and adults with epilepsy and known to be associated with drug therapy. Traditionally, neuropsychological interest has focused on cognitive abilities such as memory, executive function and language, but there is now evidence that cognitive and psychomotor slowing is evident in children and adults with new onset epilepsies even prior to the initiation of antiseizure medications1-2. Here we inquire whether baseline psychomotor slowing is a marker of abnormal prospective brain network development as interrogated by graph theory analysis. Methods: 78 children with idiopathic epilepsies with average intelligence were dichotomized into high (HSP) (n=29) and low-speed psychomotor speed groups (LSP) (n=47) based on a median split of baseline digit symbol substitution performance (Wechsler Intelligence Scale for Children III3). Participants underwent Magnetic Resonance Imaging (MRI) and high resolution T1 images (0.78 mm x 0.78 mm) were acquired to calculate structural volumes near the time of epilepsy onset (within 12 months; baseline evaluation) and two years later. The prospective difference between volumes of 87 brain regions including cortical, subcortical, and cerebellar areas was calculated and normalized to the baseline evaluation. A graph based on the correlation coefficient of the covariance between regions at baseline and prospectively was calculated for each group controlling for IQ and syndrome, and were analyzed using graph theory metrics. Results: At baseline, LSP subcortical regions weakly correlated to the rest of the network and mostly pertaining to only one module while in HSP subcortical regions are highly correlated to different modules (Figure 1). Prospectively, HSP showed a higher level of developmental organization than LSP, which is evident in the higher modularity evident in that group (Figure 2). Specifically, LSP showed two main modules and three additional modules in the periphery of the network; therefore, showing low synchronization to the rest of the regions. HSP showed a more compartmentalized development with segregated modules that at the same time showed integration with the rest of the network. In terms of network hubs, LSP mainly showed nodes from frontal and temporal regions, while HSP showed hubs spread among frontal, temporal, parietal, and subcortical areas including the left cerebellum (bigger nodes in Figure 1). Therefore, it appears that, developmentally, LSP tend to have higher synchronization between frontal and temporal regions while HSP are presenting with a more equitable brain regional development. Conclusions: The results suggest that there are indeed widespread alterations in large scale networks between high and low psychomotor speed children with new onset idiopathic epilepsies. This was confirmed both cross-sectionally and prospectively. Evidently, irregularities in brain volumetric synchronization are present near epilepsy onset and prospectively between HSP and LSP and point to the contribution of dysregulated anatomic networks in LSP.  Funding: NIH-NINDS