State-level macro-economic factors moderate the association of low income with brain structure and mental health in U.S. children

Macrostructural characteristics, such as cost of living and state-level anti-poverty programs relate to the magnitude of socioeconomic disparities in brain development and mental health. In this study we leveraged data from the Adolescent Brain and Cognitive Development (ABCD) study from 10,633 9-11 year old youth (5115 female) across 17 states. Lower income was associated with smaller hippocampal volume and higher internalizing psychopathology. These associations were stronger in states with higher cost of living. However, in high cost of living states that provide more generous cash benefits for low-income families, socioeconomic disparities in hippocampal volume were reduced by 34%, such that the association of family income with hippocampal volume resembled that in the lowest cost of living states. We observed similar patterns for internalizing psychopathology. State-level anti-poverty programs and cost of living may be confounded with other factors related to neurodevelopment and mental health. However, the patterns were robust to controls for numerous state-level social, economic, and political characteristics. These findings suggest that state-level macrostructural characteristics, including the generosity of anti-poverty policies, are potentially relevant for addressing the relationship of low income with brain development and mental health.

population density tend to have both higher costs of living and more generous antipoverty policies. However, the direct mechanisms through which population density may mitigate or augment the impacts of low SES are not currently known.
Unemployment is the average percentage of the labor force in the state that was unemployed in 2017 as reported by the U.S. Bureau of Labor Statistics. 4 Unemployment rates reflect economic conditions and would be expected to relate to family incomes and potentially to the generosity of anti-poverty policies.
Incarceration rate is the number of individuals incarcerated in the state out of every 100,000 in population as reported by the U.S. Bureau of Justice Statistics. 5 Because people who are incarcerated tend to come from lower SES backgrounds than those who are never incarcerated, 6 it is plausible that the impacts of low SES may be augmented in states that are more punitive and incarcerate a greater proportion of their population.
State Preschool Enrollment is the proportion of 4-year old children in each state enrolled in state-funded prekindergarten during the 2016-2017 school year as reported by the National Institute for Early Education Research. 7 It is plausible that providing preschool to a greater proportion of the population may mitigate the impacts of low SES on neurodevelopmental outcomes.

th Grade
Reading proficiency is the average reading score on the National Assessment of Educational Progress (NAEP) among students receiving free or reduced price lunch as reported by the National Center for Education Statistics. 8 It is plausible that these educational outcomes reflect the quality of the schools attended by low SES children in the state, potentially mitigating the impacts of low SES on neurodevelopmental outcomes.
Political preferences were estimated based on the results of the 2020 presidential election. These vales reflect the proportion of votes in each state that were cast for Joe Biden during the 2020 presidential election as reported by CNN and officially confirmed by the Federal Election Comission. 9,10 Regions that vote more Democratic tend to have both higher costs of living and more generous antipoverty policies. However, outside of a greater tendency to implement more generous antipoverty policies, the direct mechanisms through which political preferences may mitigate or augment the impacts of low SES on developmental outcomes are not currently known.
Women's political participation (Women's PP) is a composite score calculated by the Institute for Women's Policy Research 11 based on four indicators of women's political status: voter registration, voter turnout, representation in elected office, and women's institutional resources. Regions where women have higher political participation tend to have both higher costs of living and more generous antipoverty policies. However, the direct mechanisms through which women's political participation may mitigate or augment the impacts of low SES on developmental outcomes are not currently known.
Reproductive rights is a composite score calculated by the Institute for Women's Policy Research 11 based on nine indicators of women's reproductive rights: mandatory parental consent or notification laws for minors receiving abortions, waiting periods for abortions, restrictions on public funding for abortions, the percent of women living in counties with at least one abortion provider, pro-choice governors or legislatures, Medicaid expansion or state Medicaid family planning eligibility expansions, coverage of infertility treatments, same-sex marriage or secondparent adoption for individuals in a same-sex relationship, and mandatory sex education. Regions with greater reproductive rights tend to have both higher costs of living and more generous antipoverty policies. Further, greater access to abortion, contraception, adoption, and sex education may directly impact family SES by increasing family size and needs. However, the potential role of reproductive rights in mitigating the impacts of SES on developmental outcomes is not currently known.
Tightness-looseness is a measure of the strength of punishment (tightness, e.g., the severity of punishment for violating laws) and the degree of permissiveness (looseness, e.g., access to alcohol) in a state. 12 Higher scores indicate greater tightness and lower looseness.
Tightness-looseness has been found to be predict other state characteristics, such as rates of inequality, discrimination, social stability, and homelessness. However, the direct mechanisms through which tightness-looseness may augment or mitigate the impacts of low SES on developmental outcomes are not currently known. .268 Note: Analyses were conducted using linear mixed-effects models with the nlme package in R using two-tailed tests. Age, sex, and the proportion of participants at each site that were White, Black, and Latinx were also included as covariates in all analyses. .033 Note: Analyses were conducted using linear mixed-effects models with the nlme package in R using two-tailed tests. Age, sex, and the proportion of participants at each site that were White, Black, and Latinx were also included as covariates in all analyses. Minimum wage is the states' minimum wage in 2017 as reported by the U.S. Department of Labor. 3 The federal minimum wage was used for states with no established minimum wage or with a minimum wage lower than the federal minimum wage. Note: Analyses were conducted using linear mixed-effects models with the nlme package in R using two-tailed tests. Age, sex, and the proportion of participants at each site that were White, Black, and Latinx were also included as covariates in all analyses. Note: Analyses were conducted using linear mixed-effects models with the nlme package in R using two-tailed tests. Age, sex, and the proportion of participants at each site that were White, Black, and Latinx were also included as covariates in all analyses. Note: Analyses were conducted using linear mixed-effects models with the nlme package in R using two-tailed tests. Age, sex, and the proportion of participants at each site that were White, Black, and Latinx were also included as covariates in all analyses.  .009 Note: Analyses were conducted using linear mixed-effects models with the nlme package in R using two-tailed tests. Age, sex, and the proportion of participants at each site that were White, Black, and Latinx were also included as covariates in all analyses. .009 Note: Analyses were conducted using linear mixed-effects models with the nlme package in R using two-tailed tests. Age, sex, and the proportion of participants at each site that were White, Black, and Latinx were also included as covariates in all analyses.