Measurement matters: An individual differences examination of family socioeconomic factors, latent dimensions of children’s experiences, and resting state functional brain connectivity in the ABCD sample

Abstract

The variation in experiences between high and low-socioeconomic status contexts are posited to play a crucial role in shaping the developing brain and may explain differences in child outcomes. Yet, examinations of SES and brain development have largely been limited to distal proxies of these experiences (e.g., income comparisons). The current study sought to disentangle the effects of multiple socioeconomic indices and dimensions of more proximal experiences on resting-state functional connectivity (rsFC) in a sample of 7834 youth (aged 9–10 years) from the Adolescent Brain Cognitive Development (ABCD) study. We applied moderated nonlinear factor analysis (MNLFA) to establish measurement invariance among three latent environmental dimensions of experience (material/economic deprivation, caregiver social support, and psychosocial threat). Results revealed measurement biases as a function of child age, sex, racial group, family income, and parental education, which were statistically adjusted in the final MNLFA scores. Mixed-effects models demonstrated that socioeconomic indices and psychosocial threat differentially predicted variation in frontolimbic networks, and threat statistically moderated the association between income and connectivity between the dorsal and ventral attention networks. Findings illuminate the importance of reducing measurement biases to gain a more socioculturally-valid understanding of the complex and nuanced links between socioeconomic context, children’s experiences, and neurodevelopment.

Assessment of Parent Income and Education, Neighborhood Disadvantage, and Child Brain Structure

“…Third, to ensure findings are driven by variation in the construct of interest rather than measurement artifacts, future work should use models that adjust for measurement biases. 79 Fourth, the effect sizes in this study were small, but small effects can accumulate over time and be meaningful at the population level. 80 However, the practical implications of these effects and how they evolve (ie, magnify or diminish) over time should be tested in future work.…”

Socioeconomic resources are associated with distributed alterations of the brain’s intrinsic functional architecture in youth

“…Several recent studies have also examined associations between SER and resting state functional connectivity in the same ABCD study dataset and reached somewhat different conclusions ( Rakesh et al, 2021a , Rakesh et al, 2021b , DeJoseph et al, 2022 ). These studies generally found much weaker brain-behavior associations than what is reported here.…”

Socioeconomic Resources are Associated with Distributed Alterations of the Brain’s Intrinsic Functional Architecture in Youth

“…As shown in Figure S5 and Table S2, we found that between 68 to 80% of the multivariate signal associated with SER-related variables is lost when using the cell mean summary data rather than connection-resolution data. It is possible that this sizable reduction in signal explains the weaker effect sizes observed in these recent studies (45–47) as well as the different pattern of effects observed. Of note, there were in addition other differences in the current study that could contribute to differing results.…”

“…All analyses reported in the Main Manuscript were performed on whole-brain connectomes, each with 87,153 connections. Recent studies (45–47) of SER and SER-related variables in ABCD use summary statistics in which each subject’s connectome is reduced to 78 numbers representing mean connectivity between each pair among 12 large-scale networks, and these summary statistics are available through the ABCD Data Exploration and Analysis Portal; https://deap.nimhda.org. We repeated our multivariate predictive modeling approach with summary statistics data to assess whether the underlying signal associated with SER-related variables is preserved in the summary data.…”

“…Several recent studies have also examined associations between SER and resting state functional connectivity in the same ABCD study dataset and reached somewhat different conclusions (45–47). These studies generally found much weaker brain-behavior associations than what is reported here.…”