Introduction: Childhood adversity is associated with perturbations in brain development. Sources of childhood adversity range from proximal, occurring in a child’s home, to distal, occurring in the child’s neighborhood. This study takes a novel statistical approach, multilayer network analysis, to explore relationships among environmental and resting-state functional MRI measures in children (n = 131, 4-10 years old).

Methods: A bilayer network was constructed using a graphical Lasso approach, which estimates sparse partial correlation networks. We used 6 exposome measures: total crime; Quality of Life index reflecting access to education, food, and healthcare; GINI index of income inequality; percentage of adults with a Bachelor’s degree; unemployment; and rate of high child blood lead levels. We used 2 proximal SES measures: parent education and family income. Participation coefficients (PC) were extracted from 400 parcels in 7 networks (Yeo et al., 2011). PC is a graph measure of integration of the network's nodes (parcels): a lower PC indicates more intra-network connectivity and less variation in the inter-network connections. PC typically declines with age as cortex specializes.

Results: The Quality of Life index was positively associated with the PC of the salience network. The GINI index was positively correlated with dorsal attention PC. Blood lead levels were associated with the PC of the default mode network. No associations were specific to proximal SES.
Conclusions: These results suggest that children who live in higher opportunity neighborhoods have less mature and less specialized external attention networks, consistent with the theory that safety and cognitive enrichment prolong brain development. The link between neighborhood lead and the default mode network is consistent with evidence that toxins are deposited in the brain along the midline due to circulatory system structure, but more work is needed to explore this correlation.