Figure 3. Association of brain metrics with childhood psychopathology.

A) Standardized regression β effect sizes and 95% CIs from mixed-effect regressions assessing the association of significant (pfdr<0.05) brain metrics with psychopathology. B) Standardized mediation effect and 95% quasi-Bayesian CIs from longitudinal mediation analyses testing whether significant brain metrics mediate the association of prenatal cannabis exposure (PCE) with psychopathology. Analyses were run first collapsing across PCE groups (i.e., any exposure) and post-hoc analyses computed effects for each group separately. C) Mediation path diagrams for significant mediations (p<0.05 fdr) with regions which survived correction for all comparisons in primary analyses (Figure 2; restricted total diffusion of the forceps minor and transverse diffusivity of the right pars triangularis grey matter). Values reflect standardized effects and 95% CIs. c = total effect, c’ = direct effect.