Distinct associations between multimodal brain measures and psychopathology domains predict adolescent functioning
Journal:
bioRxiv
Published Date:
Jun 4, 2026
Abstract
Adolescent psychopathology is partly rooted in measurable disruptions across key neural networks, yet the field still lacks an integrated, multimodal understanding of these brain-behavior links. Here, we examined how structural, microstructural, and functional measures across corticostriatal, corticolimbic, and executive control networks relate to psychopathology domains and explored how these associations predicted future psychosocial functioning. We used data from the Adolescent Brain Cognitive DevelopmentSM Study (n=5,408) and ran a regularized canonical correlation analysis to identify distinct modes of covariation between multiple brain measures and psychopathology domains when youth were 13-14 years old. The resulting canonical brain and psychopathology scores were used to predict school-related impairment one year later. First, higher diffusivity and decreased activation during a reward task across all three networks as well as lower corticostriatal surface area were related to higher broad psychopathology. Second, lower corticolimbic diffusivity, executive control volume and surface area, and cortical thickness across all three networks as well as higher corticostriatal and corticolimbic volumes were related to higher anxiety but lower externalizing. For the first mode, higher psychopathology scores predicted more school-related impairment one year later. For the second mode, higher brain and higher psychopathology scores predicted less school-related impairment one year later. Identifying how specific neural measures align with psychopathology domains, as well as how both forecast real-world functioning, advances the conceptualization of adolescent mental health. This approach clarifies which levels of analysis provide distinct versus shared information about youth functioning and highlights potential mechanisms that may inform future targets for change.
