Fig. 4.
Cortical volume mediates the association between gestational age and cognitive assessments in 11,203 adolescents at baseline. The standardized mediations by cortical volume were significant from the gestational age on total intelligence (A β = 0.01, P = 1.51 × 10−4), crystallized intelligence (B β = 0.01, P = 1.50 × 10−4), and fluid intelligence (C β = 0.003, P = 1.37 × 10−4). Path A: the effect of the independent variable (gestational age) on the mediator (cortical volume); Path B: the effect of the mediator (cortical volume) on the dependent variable (cognitive intelligence); Path C: the regression coefficient (β value) representing the total effect of the independent variable (the gestational age) on the dependent variable (cognitive intelligence) when the mediator (cortical volume) was not taken into account; Path C’: the direct effect of the gestational age on the cognitive intelligence when controlling for the cortical volume. The regression coefficient in Path C’ shows some reduction in contrast to that in Path C. Path AB: the indirect effect of gestational age on cognitive intelligence through cortical volume can then be quantified as the product of Path A multiply by Path B. CI: confidence interval. The cortical volume used as mediator was the averaged value across regions showing significant group difference of gestational age at baseline