Regularized CCA identifies sex-specific brain-behavior associations in adolescent psychopathology

Adolescence is a critical period of neural development and a sensitive window for the emergence of psychiatric symptoms. Resting-state functional MRI (rs-fMRI) provides a unique opportunity to investigate brain-behavior associations. However, the role of sex-specific differences in these associations remains underexplored, despite their potential to reveal heterogeneous neurobiological mechanisms and guide personalized interventions. In this study, we analyzed data from the Adolescent Brain Cognitive Development (ABCD) Study, comprising 7,892 adolescents (9-10 years old, 3,896 females). Using Canonical Correlation Analysis (CCA) and a rigorous cross-validation framework, we identified associations between cortical-to-cortical (Cor-Cor) and cortical-to-subcortical (Cor-Sub) functional connectivity and eight symptom domains from the Child Behavior Checklist (CBCL). Unlike previous approaches, we directly examined sex differences within the brain-behavior mappings by applying separate CCA models in boys and girls to uncover differential connectivity-behavior relationships. Our analysis uncovered two reproducible components for both Cor-Cor and Cor-Sub mappings on the whole cohort (r₁ = 0.130, p < 0.001, r₂ = 0.122, p < 0.01 for Cor-Cor; r₁ = 0.157, p < 0.001, r₂ = 0.115, p < 0.01 for Cor-Sub). Importantly, sex-stratified analyses revealed distinct patterns of brain-behavior associations. Among females, high loadings on attention and thought problems were linked to high loadings on default mode network, whereas in males, attention and thought problems were linked to sensorimotor networks. Compared to females, males also had higher loadings on internalizing symptoms, such as anxious/depressed and withdrawn/depressed symptoms, coupled with lower loadings on putamen and hippocampus functional connectivity. These findings suggest there may be fundamentally different brain-behavior mappings across the sexes in adolescence, in addition to previously reported sex differences in functional connectivity and behavioral profiles. By revealing sex-specific neural correlates of psychiatric symptoms in early adolescence, this study paves the way for sex-informed strategies in clinical risk assessment and personalized treatment design.