Trajectories of CT changes during late childhood (from ages 9–10 to ages 11–12), and from adolescence (age 14) to young adulthood (age 19) were examined using the longitudinal data from ABCD and IMAGEN, respectively. Specifically, CT was compared between T1 and T2 using LME models (performed separately for each dataset) which included a participant’s sex, recruitment site, mode-centered PDS, intracranial volume (ICV), and sampling time (T1 or T2, with T1 as reference). The age deviation from the mean age for each participant was calculated separately for the two time points and included in the LME models to capture the age-related variance at T1 and T2. The participant’s ID was included as a random effect. The Euler number, a surface reconstruction quality index [29], was only available for IMAGEN data. It was included as a nuisance covariate in the analysis of IMAGEN data. The normality and homoscedasticity of residuals were visually checked. The Cohen’s d effect size of the sampling time was calculated according to the formula below.
