Minimal Association was Observed Between SES and Amygdala and Hippocampal Development in U.S. Children

• In Event: Biological embedding of early adversity: new insights on the effects on neurodevelopment and threat processing

Abstract

Introduction

Several studies have hypothesized that the amygdala and the hippocampus are disproportionately susceptible to the impact of socioeconomic status (SES) (Tottenham & Sheridan, 2009; Noble et al., 2012; Merz, Tottenham & Noble, 2018) during brain development, due to their heightened sensitivity to stress and HPA axis activities. However, the reported SES effects on the volumes of the amygdala and hippocampus so far have been inconsistent (Noble et al., 2012; Noble et al., 2015; Dufford, Bianco & Kim, 2019). With advances in analytical tools, propensity score weighting can now be applied to the ABCD Study (Heeringa & Berglund, 2020), matching its demographic and social characteristics to that of the American Community Survey. This offers a unique opportunity to estimate the magnitude of the SES effects on the development of the amygdala and hippocampus at the population level.

Methods

We applied propensity score weighting to the baseline ABCD sample (N = 8141) to assess the effects of SES on the development of the amygdala and hippocampus in U.S. children. Specifically, we estimated the association between parental education and income-to-needs ratio (INR) and the relative volume (controlled for whole brain volume), mean diffusivity (MD), and restricted isotropic diffusion (N0) of the hemispheric amygdalae and hippocampi. To probe the specificity of the SES effects to the limbic system structures, the SES effects on two control brain regions, the gray matter portion of the lateral occipital cortex (LOC) and the thalamus, were also estimated. All models included both SES measures and adjusted for age, sex, race ethnicity, scanner identification, and the random effect of family, and were Bonferroni corrected for 24 statistical comparisons.

Results

Estimating effects adjusted with propensity score weightings for the demographic and SES characteristics of 9-10 years old U.S. children, we observed no significant association between parental education and INR and the relative volume of these regions. Significant but minimal associations were found for the diffusion measures. The joint effect of parental education and INR were associated with the MD of the left amygdala (ΔR2 = 0.002) and left LOC (ΔR2 = 0.003), and the N0 of both amygdalae (L: ΔR2 = 0.003, R: ΔR2 = 0.003) and LOCs (L: ΔR2 = 0.005, R: ΔR2 = 0.004). Higher parental education, controlling for INR, was also associated with decreased MD in the left amygdala (ΔR2 = 0.002) and increased N0 of both amygdalae (L: ΔR2 = 0.003; R: ΔR2 = 0.002) and the left LOC (ΔR2 = 0.002).

Discussion

Estimating the impact of SES on brain development at the population level, we found minimal effects of parental education and INR on the macro- and microstructure of the amygdala and hippocampus. Adjusted for INR, higher parental education was associated with decreased MD and increased N0 of the amygdala, possibly reflecting a more advanced maturation profile of this brain region, its effect size was relatively small and appeared more distributed than previous hypothesized. We will subsequently present results on the distribution of these SES effects across the brain to better understand its influence on brain development.

Authors

• Weiqi Zhao, University of California - San Diego
  Presenting Author

• Marybel Robledo Gonzalez, University of California - San Diego
  Non-Presenting Author

• Terry L. Jernigan, University of California - San Diego
  Non-Presenting Author