JCPP Advances (Dec 2022)
Differences in cortical morphology and child internalizing or externalizing problems: Accounting for the co‐occurrence
Abstract
Abstract Background Childhood internalizing and externalizing problems frequently co‐occur. Many studies report neural correlates of either internalizing or externalizing problems, but few account for their co‐occurrence. We aimed to assess specific cortical substrates of these psychiatric problems. Methods We used data from 9635 children aged 9–11 years in the baseline Adolescent Brain Cognitive Development Study. Internalizing and externalizing problem composite scales scores were derived from the Child Behavior Checklist. We standardized FreeSurfer‐derived volumes of 68 cortical regions. We examined internalizing and externalizing problems separately and jointly (covariate‐adjustment) in relation to cortical volumes, with and without adjusting for total brain volume (TBV) in multivariate linear regressions adjusted for demographics and multiple comparisons. We fit bifactor models to confirm the consistency of patterns exploring specific internalizing and specific externalizing problems. Sensitivity analyses included a vertex‐wide analysis and a replication in another large population‐based study. Results In separate TBV‐unadjusted analyses, externalizing and internalizing problems were associated with smaller cortical volumes. If adjusted for externalizing behavior, however, larger cortical volumes were associated with internalizing problems, while smaller cortical volumes remained associated with externalizing problems after adjustment for internalizing problems. The bifactor model produced similar results, which were consistently replicated in another pre‐adolescent neuroimaging sample. These associations likely represent global effects: adjusting for TBV rendered most associations non‐significant. Vertex‐wise analyses confirmed global patterns. Conclusion Our results suggest that internalizing and externalizing problems have globally opposing, and non‐specific associations with cortical morphology in childhood, which are only apparent if analyses account for their co‐occurrence.
Keywords