Missing data – LGCM requires at least three timepoints, which were present in ~85% of COGA's prospective study sample overall, but only 59% of the analytic sample in the present analyses. Childhood Trauma Measures – While some traumatic events would have occurred during the study period (baseline through follow-up 3), many of the childhood traumatic events assessed relied on retrospective self-reports. Retrospective reports may be unstable over time, typically underestimating trauma exposure prevalence, potentially biasing this study's findings towards the null, especially for the individuals who were older at baseline. This may also account in part for the relatively small effect sizes observed for the effect of trauma on EEGc. Additionally, participants’ perceptions of how stressful they found the traumatic event, experiences of prolonged/repeated traumas, and childhood maltreatment (e.g. neglect) were not measured. Although our models accounted for the effect of different types of trauma, we did not present the effect of multiple types of trauma exposure (e.g. CPAT*CSAT) due to power limitations, despite evidence that children often experience multiple types of childhood abuse (Chiu et al., 2013). Future analyses
