Person:

Rivkin, Michael

Abstract Objective: Little is currently known about the impact of congenital heart disease (CHD) on the organization of large‐scale brain networks in relation to neurobehavioral outcome. We investigated whether CHD might impact ADHD symptoms via changes in brain structural network topology in a cohort of adolescents with d‐transposition of the great arteries (d‐TGA) repaired with the arterial switch operation in early infancy and referent subjects. We also explored whether these effects might be modified by apolipoprotein E (APOE) genotype, as the APOE ε2 allele has been associated with worse neurodevelopmental outcomes after repair of d‐TGA in infancy. Methods: We applied graph analysis techniques to diffusion tensor imaging (DTI) data obtained from 47 d‐TGA adolescents and 29 healthy referents to construct measures of structural topology at the global and regional levels. We developed statistical mediation models revealing the respective contributions of d‐TGA, APOE genotype, and structural network topology on ADHD outcome as measured by the Connors ADHD/DSM‐IV Scales (CADS). Results: Changes in overall network connectivity, integration, and segregation mediated worse ADHD outcomes in d‐TGA patients compared to healthy referents; these changes were predominantly in the left and right intrahemispheric regional subnetworks. Exploratory analysis revealed that network topology also mediated detrimental effects of the APOE ε4 allele but improved neurobehavioral outcomes for the APOE ε2 allele. Conclusion: Our results suggest that disruption of organization of large‐scale networks may contribute to neurobehavioral dysfunction in adolescents with CHD and that this effect may interact with APOE genotype.

Abstract Objective: Adolescents with d‐transposition of the great arteries (d‐TGA) who had the arterial switch operation in infancy have been found to have structural brain differences compared to healthy controls. We used cortical thickness measurements obtained from structural brain MRI to determine group differences in global brain organization using a graph theoretical approach. Methods: Ninety‐two d‐TGA subjects and 49 controls were scanned using one of two identical 1.5‐Tesla MRI systems. Mean cortical thickness was obtained from 34 regions per hemisphere using Freesurfer. A linear model was used for each brain region to adjust for subject age, sex, and scanning location. Structural connectivity for each group was inferred based on the presence of high inter‐regional correlations of the linear model residuals, and binary connectivity matrices were created by thresholding over a range of correlation values for each group. Graph theory analysis was performed using packages in R. Permutation tests were performed to determine significance of between‐group differences in global network measures. Results: Within‐group connectivity patterns were qualitatively different between groups. At lower network densities, controls had significantly more long‐range connections. The location and number of hub regions differed between groups: controls had a greater number of hubs at most network densities. The control network had a significant rightward asymmetry compared to the d‐TGA group at all network densities. Conclusions: Using graph theory analysis of cortical thickness correlations, we found differences in brain structural network organization among d‐TGA adolescents compared to controls. These may be related to the white matter and gray matter differences previously found in this cohort, and in turn may be related to the cognitive deficits this cohort presents.