The COI 2.0 is a national contemporary measure of neighborhood opportunity, comprising a comprehensive dataset that aggregates 29 indicators of neighborhood conditions for 72,000 census tracts in the United States. Beginning with the ABCD 4.0 data release, the ABCD Study provides scores for the COI 2.0 overall index, and the three domain indices that comprise the overall index: education , health and environment , and social and economic opportunities.Detailed documentation describing the indicators that comprise each of the domains as well as the dataset source and year for each of the 29 indicators can be found in Supplemental Table 4 and the COI 2.0 technical documentation . Given the diverse demographics of the ABCD Study participants, linking the COI 2.0 gives us objective measures of neighborhood opportunities for participants so that we can assess the influence of neighborhood quality on adolescent health and potential emerging health disparities. Crime rates are an important neighborhood characteristic that can cause distress on individuals’ mental well-being and has been linked with various children’s developmental outcomes . However, the impact of crime within the context of other neighborhood variables and how these impact neural mechanisms during children’s development is less clear. To empower researchers to investigate the impact of local crime rates in the broader context of the built environment, we obtained county-level crime statistics from Uniform Crime Reporting Data . In addition to the total crime rates, we also provided subcategories of the crime, including violent crimes, drug violations, drug sales, cannabis grow racks sales, drug possessions, and DUIs.

The LED Environment Working Group strives to include additional information about the built and natural environments of all participants in the ABCD Study. These data provide an additional perspective about differences both between study sites and individual differences among children within even a single given study site location. Integrating these external environmental factors are likely important in considering both mediating and moderating effects and allows for important questions to be asked with implications for policies that may help ensure all children can thrive. That is, given the wealth of additional data collected in the ABCD Study, the addition of understanding the built and natural environment in ABCD provides the opportunity to think more broadly about how these factors may influence neuro development of children within the established social determinants of health framework of public health . Specifically, health outcomes, including neuro development, cognition, and mental health as measured extensively by the ABCD Study, have been recognized to be influenced by complex interactions among environmental, social, and economic factors that are ultimately closely tied to one another . Dahlgren and Whitehead provided a visual representation of such complex processes as a model of the main determinants of health and well-being in public health, which has since helped shape public health policy at both national and global scales . Thus, capturing the broader physical environment makes the ABCD Study an ideal resource for researchers interested in studying how various distal and proximal factors may impact developing children and their health. While a number of development cognitive research studies have focused on individual factors, including socio-demographic factors , lifestyle , and social environments , additional natural and built environmental factors including neighborhood quality, community-level access to resources and opportunities, and exposure to harmful substances, provides an additional layer as to understanding and identifying key factors of neuro development and to promote policies that lead to better health outcomes for all children across America.

Specifically, these data can allow for researchers to examine if upstream built and natural factors might account for and/or moderate associations between physical activity and brain development, understanding the link between screen-time and mental health, determining how neighborhood conditions may impact the formation of peer groups, or exploring how recreational activities may moderate the relationship between adverse neighborhood conditions and mental health. In doing so, not only may we have a better understanding of the complex associations between the various factors contributing to neuro development across childhood and adolescence, but research findings may also point to possible public health targets for intervention and treatment. While there are clear strengths in mapping the environmental context of today’s youth in the ABCD Study, there are also several important technical limitations as well as considerations for researchers planning to use and interpret these data. A vital consideration to this type of geospatial research and the variables derived from it, is the accuracy of the assignment of the exposure assessment at any given time. Several challenges arise in trying to maximize this accuracy. Any given geospatial database has both a spatial and temporal component. How these data were derived, and the degree of resolution is important to consider. For example, census tracts can be rather large, whereas in urban areas drastic differences in the environment can sometimes be noted to vary from street to street. Furthermore, individuals who live in the same census tract should not be considered to have the same experiences or the same amount of exposure in the neighborhood as others with similar demographics. Moreover, many times, geospatial databases are compiled after data is available from other sources, such as the American Community Survey or the Environmental Protection Agency. Thus, exposure estimates can often reflect a snapshot in time that may or may not overlap directly with the time period that the child was at that residential location; requiring the researcher to consider if the exposure of interest can or cannot be assumed to be stable beyond the temporal domains of the dataset.

For example, many databases may create variables using 5-year averages that have then been linked to the baseline residential addresses which were collected in 2016–2018. Another technical challenge is that retrospective address collection is hindered by recall bias, or the differences in the accuracy or completeness of caregivers in the ABCD Study to recall address details over the 9–10 years prior to study enrollment. In addition, exposure assessment based on residential geospatial location also fails to capture individual data on percentage of time in which children in the current study spend time at their primary address versus other daily activities and/or various locations, such as in school. Of course, it is important to note that misclassification of exposure may be lower for children in that they may spend more of their time around the home, as compared to other populations such as adults who may spend more time commuting, time at work, or so forth. Although children do spend a substantial period of time at school, which may or may not be in a similar geographical location to that of their primary residence. Lastly, there is not a direct correlation between external environmental exposures to chemicals and internal exposure doses. For some environmental toxins, internal biomarkers exist to determine internal dose , whereas others, like air pollution, do not. Nonetheless, these geospatial factors can lead to misclassification, or information bias, which can severely affect observed associations between the exposure and the outcome. Therefore, given these limitations, it is important to note that while the current LED Environment measures may help provide a snapshot as to the built and natural environment surrounding ABCD participants’ residential homes, the current data fall short of fully characterizing participant exposomes. Thus, while continued efforts by the LED Environment Working Group aim to mitigate these challenges, findings should be interpreted considering these potential pitfalls, and misclassification should be acknowledged and discussed when necessary. Another potential challenge for researchers using these data is conceptual and/or statistical collinearity and potential confounders. Environmental variables included from various databases can greatly overlap in terms of theoretical construct. For example, various factors may represent broad constructs of economic advantage, and many variables from the same databases may be highly collinear. It is also important to note that although some estimates may draw from similar linked databases , they may implement any number of transformations or operations when computing measures. In addition to considering the exposure of interest from these data, a number of spatial contextual variables may also be important to consider as source of confounding. For example, ecological variables,cannabis grow system such as air pollution, may be an important spatial confounder in examining associations between neighborhood socioeconomic factors and child health outcomes in ABCD. Some models of exposures may also include other important geospatial or socioeconomic factors in establishing estimates of exposure, such as temperature and humidity in estimating ambient air pollution, or age of housing in compiling a metric for lead risk. Therefore, it is vital in the early stages of planning analyses with these data to consider the choice of which variables to use for a given construct, identifying potential ecological or spatial confounders, and understanding the raw datasets that were utilized in calculating various environmental and societal variables included in the ABCD Study. Additional sensitivity analyses should always be considered to evaluate the impact of potential confounds and the specificity of the tested environments.

Lastly, researchers should note that the environmental estimates do not represent the ‘lived’ or subjective experience of these exposures, with careful consideration given to the potential interpretation of any effects seen between these variables and brain and cognitive outcomes of interest. For example, these data are derived from outside databases that may capture an objective perspective of a given geospatial location, as they do not rely on the subjective report of the participants. However, these objective constructs do not necessarily reflect any individual’s subjective experience in a given state, census tract, or even residential neighborhood. It is likely that subjective experiences may moderate or mediate associations of external estimates of exposures. Further, neighborhood socioeconomic factors, environmental exposures, and potential health and behavioral outcomes should also be considered in light of local, state, and federal policies of racism, segregation, and inequality that has resulted in persistent inequalities in social, economic, and educational opportunities . For these reasons, socioeconomic and other family-level factors are likely to also be highly correlated to various built and natural exposure variables. Thus, thoughtful consideration is vital in reporting on potential exposure and outcome associations but also the nexus of neighborhoods, communities, and environmental justice and equity. The LED Environmental Working Group has primarily focused on baseline residential addresses to provide additional contextual information about the places where ABCD Study participants are growing up. In this process, we continually aim to implement ways to reduce exposure misclassification. Current efforts include historical reconstruction of each child’s residential history, which offers the opportunity to create a better understanding about each child’s physical environmental exposures across their lifespan. In doing so, quality assurance of retrospective residential addresses using commercial credit-reporting data is underway to help reduce recall bias . Further, efforts are under way to improve syncing the temporal domains of linked database estimates with temporal changes in residential information for retrospective and prospective addresses. The ABCD Study’s Physical Health Working Group is also collecting biomarkers to measure exposure to some chemical toxins. Beyond improving exposure assessment, both the working group and its discussions with the greater larger scientific community has identified additional important linkage databases with other information regarding environmental toxins, urban settings, and neighborhood factors, such as green space and food deserts. The ABCD LED Environment Working Group envisions an ever-increasing resource for researchers who are keen to understand environmental impacts on the human brain. Given surges in electronic cigarette use among young people in the U.S. and concerns about vaping-related lung injury, federal policy raised the minimum retail tobacco product sales age to 21 in 2019 and in early 2020 prioritized enforcement against the sale of unapproved flavored, cartridge-based e-cigarette products . Individual states and local jurisdictions also have passed retail policies to restrict the sale of e-cigarettes and/or flavored tobacco, including flavored e-cigarettes . For example, in 2019, San Francisco banned direct and online sales of e-cigarettes without Food and Drug Administration Marketing Orders . To date, no e-cigarette company has received FDA Marketing Orders, effectively resulting in a ban on e-cigarette sales for now in San Francisco. Other jurisdictions  prohibit the sale of flavored nicotine vaping products. National youth e-cigarette prevalence in 2020 shows a decline from 2019 to 2018 levels, which were deemed epidemic . An estimated 19.6% of high school students reported past-month e-cigarette use in early 2020, prior to the COVID-19 pandemic shelter-in-place policies .