Structural brain correlates of childhood trauma with replication across two large, independent community-based samples

Abstract

Introduction: Childhood trauma and adversity are common across societies and have strong associations with physical and psychiatric morbidity throughout the life-course. One possible mechanism through which childhood trauma may predispose individuals to poor psychiatric outcomes is via associations with brain structure. This study aimed to elucidate the associations between childhood trauma and brain structure across two large, independent community cohorts.

Methods: The two samples comprised (i) a subsample of Generation Scotland (n=1,024); and (ii) individuals from UK Biobank (n=27,202). This comprised n=28,226 for mega-analysis. MRI scans were processed using Free Surfer, providing cortical, subcortical, and global brain metrics. Regression models were used to determine associations between childhood trauma measures and brain metrics and psychiatric phenotypes.

Results: Childhood trauma associated with lifetime depression across cohorts (OR 1.06 GS, 1.23 UKB), and related to early onset and recurrent course within both samples. There was evidence for associations between childhood trauma and structural brain metrics. This included reduced global brain volume, and reduced cortical surface area with highest effects in the frontal (β=-0.0385, SE=0.0048, p(FDR)=5.43x10-15) and parietal lobes (β=-0.0387, SE=0.005, p(FDR)=1.56x10-14). At a regional level the ventral diencephalon (VDc) displayed significant associations with childhood trauma measures across both cohorts and at mega-analysis (β=-0.0232, SE=0.0039, p(FDR)=2.91x10-8). There were also associations with reduced hippocampus, thalamus, and nucleus accumbens volumes.

Discussion: Associations between childhood trauma and reduced global and regional brain volumes were found, across two independent UK cohorts, and at mega-analysis. This provides robust evidence for a lasting effect of childhood adversity on brain structure.

Keywords: MRI; adversity; brain; depression; trauma.

Figure 1.

Analysis structure of this study, showing the initial discovery analysis in GS due to deeper phenotyping of childhood trauma, with replication in UKB and mega-analysis for maximum sample size.

Figure 2.

Lollipop plots showing –log10 of p_(FDR) for each region and metric in the regional analysis for (A) GS, (B) UKB, and (C) the mega-analysis; and in the global and lobar analyses for (D) GS, (E) UKB, and (F) the meg-analysis. The –log10 of p_(FDR) = 0.05 is represented by the dotted gray line, any points exceeding this line achieved statistical significance after FDR correction. Regions of higher significance are labeled; the order in which other regions are presented can be found in Supplementary Material. Abbreviations: CV, cortical volume; SA, cortical surface area; ScV, subcortical volume; Th, cortical thickness.

Figure 3.

Map of showing Beta values for regions significantly associated with childhood trauma in the mega-analysis, for cortical thickness, cortical surface area, and cortical volume. The red-toned colors represent a negative association and the purple-toned represent a positive association, with the strength of that association denoted by the shade of the color.