Background
We used a polygenic score for externalizing behavior (extPGS) and structural MRI to examine potential pathways from genetic liability to conduct problems via the brain across the adolescent transition.
Methods
Three annual assessments of child conduct problems, attention-deficit/hyperactivity problems, and internalizing problems were conducted across across 9–13 years of age among 4,475 children of European ancestry in the Adolescent Brain Cognitive DevelopmentSM Study (ABCD Study®).
Results
The extPGS predicted conduct problems in each wave (R
2 = 2.0%–2.9%). Bifactor models revealed that the extPRS predicted variance specific to conduct problems (R
2 = 1.7%–2.1%), but also variance that conduct problems shared with other measured problems (R
2 = .8%–1.4%). Longitudinally, extPGS predicted levels of specific conduct problems (R
2 = 2.0%), but not their slope of change across age. The extPGS was associated with total gray matter volume (TGMV; R
2 = .4%) and lower TGMV predicted both specific conduct problems (R
2 = 1.7%–2.1%) and the variance common to all problems in each wave (R
2 = 1.6%–3.1%). A modest proportion of the polygenic liability specific to conduct problems in each wave was statistically mediated by TGMV.
Conclusions
Across the adolescent transition, the extPGS predicted both variance specific to conduct problems and variance shared by all measured problems. The extPGS also was associated with TGMV, which robustly predicted conduct problems. Statistical mediation analyses suggested the hypothesis that polygenic variation influences individual differences in brain development that are related to the likelihood of conduct problems during the adolescent transition, justifying new research to test this causal hypothesis.