Parent-Child Cortical Similarity Indexes Multiple Dimensions of Sensitivity in the Association Between Family Environment and Adolescent Internalizing Symptoms

Abstract

Family environment is a robust predictor of adolescent internalizing symptoms, yet its strength varies across adolescents. We tested whether this variation is partly captured by how similar an adolescent's cortical structure is to their parent's and whether different forms of this similarity matter in different cortical systems. In 115 parent-child dyads from two cohorts, structural MRI quantified similarity across 400 cortical parcels in two forms: phenotypic similarity in regional morphometry and architectural similarity in interregional morphometric organization. Parent-child similarity was not directly associated with symptoms, whereas its interactions with family environment were jointly significant (chi-square(4) = 14.92, P = 0.005). Phenotypic similarity in the sensory-limbic system and architectural similarity in the association system each moderated the association between family environment and symptoms (both beta = 0.22). In both cases, the association was strongest among less similar dyads. The phenotypic interaction was distributed across sensory-limbic networks and was not reproduced by similarity to unrelated adults or by the child's cortical atypicality. The architectural interaction was concentrated in the dorsal attention and default mode networks and was distinguishable from unrelated-adult similarity but not from child atypicality. Together, these findings raise the possibility that markers of environmental sensitivity need not reside in the child alone. Parent-child cortical similarity may index a dyadic component of this sensitivity, one not adequately represented by a single similarity score but distributed across forms of similarity and cortical systems.