Abstracts

Rationale: Average intelligence children with focal epilepsy (FE) have psychiatric diagnoses[1] that In children without epilepsy are associated with abnormal cortical thickness (CT) in frontal, precentral, postcentral, parietal, temporal and occipital regions[2-5]. We examined the relationship of psychopathology severity with CT and sulcal depth (SD) in these regions using categorical (multiple (M-PsyDg), single (S-PsyDg), and no psychiatric diagnoses (n-PsyDg)) and continuous Child Behavior Checklist (CBCL) scores (externalizing, internalizing, social problem, thought problem, attention problems, social competence). Methods: High-resolution 3D MR images were obtained on 43 FE(6.1-15.3 years) in a 1.5 Tesla scanner. After removal of non-brain tissue from MR images, image voxels were classified into different tissue types. Cortical surface representations at the geometric center of the 3D gray matter (GM) tissue were extracted. CT at each point in the cortical GM mantle was defined as the sum of the distances from this point to the GM/white matter and GM/cerebrospinal-fluid tissue boundaries. A surface-based spatial normalization technique was used to match anatomically homologous cortical features across subjects before performing cross-subject comparison. Structured psychiatric interviews provided DSM-IV-TR psychiatric diagnoses. Parents completed CBCLs and provided seizure-related information. Multivariate analyses of covariance (MANCOVA), with regional CT or SD (left and right averaged) as dependent variables, were estimated separately for psychiatric groups (8M-PsyDg, 10S-PsyDg, 25N-PsyDg) or CBCL scores as independent variables with age, seizure control and age of seizure onset as covariates. Post-hoc univariate analyses examined which regions contributed to the significant MANCOVA findings. Results: MANCOVAs with CT not SD yielded a significant psychiatric diagnosis and externalizing, social problems, and thought problems scores effects (Table 1). Univariate analyses indicated significantly reduced CT in the M-PsyDg inferior frontal, lateral fronto-orbital, and inferior temporal regions compared to the S-PsyDg (p<.01-.004) and N-PsyDg groups (p<.02-.01). They also demonstrated that higher externalizing scores were significantly related to thinning in the frontal, cingulate, insular, precentral, post-central, parietal, middle occipital, subcallosal, parahippocampus, and fusiform regions. Increased social problems scores were associated with frontal, parietal, pre-central, post-central, and cingulate thinning. Higher thought problem scores were related to temporal and parietal thinning. Conclusions: In pediatric FE abnormal fronto-temporal pruning is associated with multiple psychiatric diagnoses and widespread cortical thinning with externalizing behavior problems. Involvement of fronto-parietal and cingulate regions in social skills [6], tempero-parietal cortex in language, thought, and communication [7], and supramarginal cortex in attention [4] support the more localized association of thinning with these children's social, thought, and attention problems.