Friends, family, and community influences on adolescent development: A review of impactful findings from the ABCD study® Authors: Marybel R. Gonzalez, Kazi Priyanka Silmi, Victoria Castillo, Yailene Perez, Melissa Orama, Arlyn Cuan, Diana Alkire, Robin Aupperle, Laika Aguinaldo, William E. Pelham, Elisa M. Trucco, Michael Mason, Carlos Cardenas-Iniguez, Brynn Tepp, Bruce McCandliss, Alejandra Fernandez, Ka I. Ip, Evan White, Jana Hong, Erin L. Thompson, Rada K. Dagher, Gayathri J. Dowling, Susan F. Tapert, Raul Gonzalez Published: 2026 ABSTRACT The Adolescent Brain Cognitive DevelopmentSM (ABCD) Study was funded to investigate adolescent development across 10 + years, assessing multilevel social contexts of influence and adolescent health outcomes. As such, the ABCD study has generated a large body of literature investigating adolescent health outcomes in neurocognition, mental health, physical health and substance use. However, the extent to which family, peer, and community contexts are measured and integrated across these outcomes has not been systematically examined. We synthesized 64 articles, deemed to be of high impact using a variety of metrics, to determine the extent to which measures within the Family, Friends, and Community (FFC) domains, across multi-level contexts of identity and values, family, and peers and community were investigated in relation to adolescent health outcomes. FFC measures were most frequently examined in relation to neuroimaging and neurocognitive outcomes, as well as mental health outcomes, with substantially fewer studies incorporating FFC measures for physical health and substance use outcomes, identifying FFC measures as both risk and protective factors. The ABCD study literature to date underscores the central role of FFC measures in contextualizing neurocognitive development as well as other health outcomes during adolescence, primarily with cross-sectional data. Building on existing use of FFC measures in neuroimaging and neurocognitive studies, future research can leverage longitudinal ABCD data to better understand how proximal social contexts shape neurocognitive development across adolescence. Introduction Adolescent development is shaped by the social contexts which adolescents experience growing up. These social contexts span multilevel domains including within the individual, family, peer, school, and community levels(Bronfenbrenner, 1977). Adolescence is a sensitive developmental period, characterized by rapid neurobiological maturation, increasing autonomy from caregivers, heightened sensitivity to peer relationships, and a growing engagement with broader social contexts. Together, these changes influence adolescent development as youth are highly responsive to multiple influences across friends, family, and community (FFC) domains. The Adolescent Brain Cognitive DevelopmentSM (ABCD) Study was explicitly designed to capture this developmental complexity. In addition to comprehensive assessments of outcome domains such as neurocognitive development, mental health, physical health, and substance use, the ABCD Study includes extensive, multilevel measures of social context across family, peer, school, neighborhood, and cultural environments (R. Gonzalez et al., 2021; Volkow et al., 2018; Zucker et al., 2018). These outcome domains were selected because they represent interrelated and developmentally central aspects of adolescent functioning that are both shaped by, and reciprocally influence, youths’ social environments. Importantly, these domains do not operate in isolation. During adolescence, neurocognitive development goes hand in hand with mental health, physical health, and substance use, with social contexts influencing their co-development over time. As adolescents increasingly orient toward peers while remaining embedded within family and community systems, multilevel social contexts become particularly powerful predictors of developmental trajectories. Thus, examining FFC influences across these core outcome domains is essential for understanding pathways of risk and resilience during adolescence. FFC domains (previously named Culture and Environment) refer to the social and cultural contexts of participants across multiple levels and have evolved over time and participants' development (Gonzalez et al., 2021, Zucker et al., 2018). This protocol was designed to characterize the cultural and social ecology within which adolescent brain, behavioral, and health development unfolds. Conceptually, these measures span three related but distinct levels of influence: cultural/ethnic group membership, experiences, and values; proximal social environments, including school and neighborhood perceptions; and social-interactional influences, including family, caregiver-youth, and peer processes. Thus, FFC measures include constructs such as acculturation, ethnic identity, family cultural values, perceived unfair treatment, parental warmth, parental monitoring, family conflict, peer relationships, prosocial behavior, school engagement, and perceived neighborhood safety, among others. Although these domains overlap in lived experience, distinguishing among them is important because proximal family, peer, and school processes may operate differently from broader cultural and structural conditions, and because culturally specific developmental processes may be obscured when social context is treated only as a generic environmental exposure. At the same time, FFC measures should be distinguished from the ABCD Study’s Linked External Data Environment and Policy (LED) measures, which use geospatial, census, school, administrative, policy, and other externally linked data sources to characterize structural and environmental contexts. Many objective indicators of structural context, including residentially linked neighborhood, built-environment, natural-environment, socioeconomic, policy, and environmental-exposure measures, are represented most directly in the LED domain rather than in the FFC protocol itself. The present review focuses on publications featuring FFC measures and therefore is not intended as a synthesis of the LED-specific literature. Neurobiological susceptibility to the environment during adolescence influences brain maturation necessary for executive functioning, memory, emotional regulation, and decision-making processes, that can have important implications for academic functioning, mental health, behavioral regulation, and long-term health trajectories (Fuhrmann et al., 2015, Larsen and Luna, 2018, Zhou et al., 2025). Neurocognitive risk and protective social factors among adolescents have been primarily studied at the individual level, focusing on experiences and conditions such as material deprivation, financial difficulty, low family socioeconomic status, and neglect (Abo Hamza et al., 2024, Baysarowich et al., 2025, Johnson et al., 2016, Méndez Leal and Silvers, 2021, Qiu et al., 2025, Schibli et al., 2017). In addition, individual differences in personal identity, values, and motivations, as well as culture, have been shown to confer protection or risk to neurocognitive performance (Telzer et al., 2025). Furthermore, other studies have investigated the quality and quantity of adolescents' interpersonal interactions and their impact on neurocognition, particularly with their immediate family (i.e., parents, caregivers, and siblings) and peers in their social groups or school environments (Cardoso et al., 2025, Hyde et al., 2022). Although fewer in number, several studies have identified community-level factors related to adolescent neurocognitive functioning, including linguistic and cultural diversity, availability of resources and amenities, social cohesion, safety, and neighborhood-level deprivation and/or disadvantage (Hyde et al., 2022, Lewis et al., 2025). Mental health problems have risen among youth globally, particularly in high-income countries (Bie et al., 2024, Brody and Hughes, 2025, McGorry et al., 2025). According to a 2024 national survey, youth aged 12–19 years reported depression at a higher rate than other age groups in the US, with higher prevalence among females and those with lower family incomes (Brody and Hughes, 2025). Mental health problems that have their onset during early adolescence predict negative sequelae later in development, including impaired functioning, poorer physical health, greater risk of persistent mental health conditions, and poorer quality of life in adulthood. As such, understanding social stressors that impact mental health problems is critical to inform targets to offset downstream consequences. Key proximal social risk factors include parental mental health and socio-economic status, family environment/dynamics, peer rejection, peer victimization/bullying (including cyberbullying), school-related stress, and exposure to childhood abuse or neglect (Horne, 2024, Lin and Guo, 2024, McGorry et al., 2025). These risks are often driven and amplified by social and structural determinants of health, such as neighborhood safety, crime, and school disadvantages. Increasingly, researchers are focusing attention on resilience factors that include strong familial and peer relationships, supportive school environments, healthy lifestyle behaviors, and policies that support families to address social and economic stressors (Coley et al., 2019, McGorry et al., 2025). To address the rising rates of psychological distress and anxiety among youth, it is important to understand how the broader social environment affects the proximal experiences that youth navigate. Physical health during adolescence is shaped not only by biological development, but also by broader social and environmental conditions. Given the prevalence of obesity is 22% among US adolescents aged 12–19 and over 70% of teens report getting insufficient sleep, it is essential to identify the social risk and protective factors for physical health and influence health behaviors during this developmental period, with downstream implications for long-term health trajectories (Wu et al., 2026, Centers for Disease Control and Prevention, 2025, Centers for Disease Control and Prevention, 2024). For example, perceived unfair treatment has been consistently associated with poorer sleep quality, often mediated through increased stress and loneliness (Majeno et al., 2018, Verity et al., 2025), with longitudinal evidence showing that exposure to unfair treatment predicted decreases in protective health behaviors including sleep and physical activity (Emmer et al., 2025). Neighborhood factors, like crime and socio-economic disadvantages, have been associated with obesity and obesogenic behaviors like physical inactivity, higher screen time, and shorter sleep; whereas greenspaces and recreational facilities exhibited protective associations (Daniels et al., 2021, Kim et al., 2019, Saelee et al., 2020). In the United States, more than 90% of adults with substance use disorders initiated substance use during adolescence, a developmental period in which substance use is associated with poorer educational, social, health, and neurocognitive outcomes, as well as increased risk for mental health problems and justice system involvement (Gakh et al., 2020, Ignaszewski, 2021, Jordan and Andersen, 2017, Yonek et al., 2019). The influence of social factors on adolescent substance use is well established, with multi-level risk and protective factors emerging (Bronfenbrenner, 1977). At the individual level, protective factors include emotional regulation, self-efficacy, strong anti-drug beliefs and attitudes, strong cultural identity, collectivist cultural values, and spiritual engagement enhance resilience in youth, whereas impulsivity, poor mental health, low self-worth, and experiences of cultural marginalization increase vulnerability (Bozzini et al., 2021, Nawi et al., 2021, Wang et al., 2010). At the interpersonal level, familial protective factors include parental involvement, psychological support, healthy parental communication, effective monitoring, and parent-child cultural concordance, whereas risk factors include parental or household substance use, parental conflict, family poverty, abuse or neglect, and low parental supervision (Bozzini et al., 2021, Preston and Rew, 2022, Trucco, 2020). At the peer-, school- and community-level, risk factors include peer substance use, unfair treatment as a function of race, country of origin, body type and a variety of factors, low neighborhood opportunities and high neighborhood availability of substances. In contrast, school connectedness, participation in supervised programs, perceived quality of the neighborhood, and strong community norms against substance use offer meaningful protective factors (Anderson and Mayes, 2010, Arora et al., 2021, Preston and Rew, 2022, Trucco, 2020). Risk and protective factors for youth substance use operate across multiple levels, highlighting the need for an ecological lens to understand substance use during adolescence. The current literature has largely focused on individual- and interpersonal-level determinants, leaving a gap in research examining how structural contexts, both independently and in interaction with proximal factors, shape youth substance use (Silmi et al., 2025, Trucco, 2020). Thus, understanding multilevel social influences on adolescent development is critical for identifying both risk and protective factors operating across individual, family, peer, and community levels. To date, over 300 peer-reviewed studies have leveraged data from the ABCD Study to examine how FFC constructs relate to adolescent outcomes, including neurocognitive development, mental health, physical health, and substance use (Garavan et al., 2018; R. Gonzalez et al., 2021; Volkow et al., 2018; Zucker et al., 2018). The size, heterogeneity, and continued rapid growth of this literature make exhaustive coverage of all published articles impractical within a single integrative review and would necessarily limit the depth with which individual findings could be considered. Accordingly, rather than attempting to summarize the totality of published FFC research, we conducted an impact-based narrative review using transparent, prespecified criteria to identify a manageable subset of studies demonstrating substantial scholarly, public, policy, or clinical reach. This approach prioritized the most impactful published research using a variety of complementary indices. It should not be interpreted as implying that studies excluded by these criteria are scientifically unimportant. Thus, our narrative review summarizes and contextualizes a subset of high-impact ABCD studies examining FFC influences on neurocognitive development, mental health, physical health, and substance use, with the goal of advancing a more integrated understanding of adolescent development in social context. Methods Design We conducted a narrative review of publications using Adolescent Brain Cognitive Development (ABCD) Study data in which Friends, Family, and Community (FFC) constructs were substantively evaluated (i.e., as variables of interest in analyses). ABCD participants were 9–10 years of age at baseline. The publications reviewed drew on ABCD data releases up to and including release 5.0, spanning baseline through the 4-year follow-up visit. Individual studies therefore vary in the developmental window they capture, from late childhood at baseline (ages 9–10) to early adolescence at later waves (up to ages 13–14), depending on the assessment wave analyzed. Parent consent and adolescent assent were obtained by the ABCD Study (Garavan et al., 2018; R. Gonzalez et al., 2021; Volkow et al., 2018; Zucker et al., 2018). The review emphasized high-impact and high-reach empirical contributions, using a priori impact filters to prioritize papers for synthesis, described below. This narrative review was not pre-registered. Identification of candidate publications The ABCD Study has a publication-tracking infrastructure that served to identify the total pool of ABCD publications. Since 2017, ABCD-related publications have been identified for the ABCD Research Publications web page through weekly PubMed and Google Scholar searches using terms such as “ABCD Study” and “adolescent brain cognitive development,” followed by review for relevance to the ABCD Study. Beginning in 2023, this process was supplemented by an NIH iCite/iSearch Publications portfolio maintained by a co-author (DA). The iCite/iSearch portfolio is populated through two complementary mechanisms: publications linked to NIH/HHS grants proposing to use ABCD data and dynamic keyword searches for ABCD-related publications. Because iCite/iSearch draws publication data from PubMed and does not search full article text, continued Google Scholar searches are used to supplement the portfolio, and the ABCD website and iCite/iSearch publication lists are reconciled regularly. The ABCD publications database excludes preprints and includes only publications in journals indexed in MEDLINE, Web of Science, Scopus, the Directory of Open Access Journals, or the NIH Library and marked as peer reviewed. The database is further restricted to empirical articles reporting original analyses of ABCD Study data. Articles are coded into non-mutually exclusive research domains corresponding to ABCD assessment working groups, including FFC. Using this process, an ABCD publication inventory of 1403 peer-reviewed papers (as of July 28, 2025) was made available, and domain coding by our co-author (DA) identified 334 articles in which an FFC measure was a variable of interest in analyses. Impact metrics were applied after identifying papers coded as including FFC measures. High citation counts, journal visibility, or public attention alone were not sufficient for inclusion if the paper did not fall within the FFC scope of the review or if FFC constructs were not interpretable as focal variables. Thus, ABCD studies focused primarily on geocoded neighborhood disadvantage, poverty, policy environments, or other externally linked structural-context indicators (overseen by the LED working group) were not included unless they also substantively evaluated FFC measures. Sometimes, this was not possible given the creation of factors, composite score, or other data reduction methods that combined an FFC measure with other measures outside of the FFC’s scope and did not report statistics and results specific to the FFC measure. Impact-based filtering criteria From the candidate set, papers were prioritized using a three-criterion filtering algorithm designed to capture: (1) real-world relevance and impact, (2) scholarly uptake and reach, and (3) single-metric “standouts” based on various metrics. See Fig. 1 for a flowchart of article screening and selection.Fig. 1Flow chart outling the screening and selection of articles. Flow chart outling the screening and selection of articles. Criterion 1 was designed to capture papers already having “real world impact” based on whether the published study had an iCite/iSearch indicator denoting that the paper had a "Policy Mention" and/or a "Clinical Article" indicator. The former identifies publications that have been cited in policy-relevant documents, such as documents written primarily for or by policymakers and published by policy-focused sources such as governments, IGOs, NGOs, and think tanks. The latter refers to publications that are a Clinical Trial or Clinical Guidance. Clinical article status was derived from NIH iCite/iSearch metadata based on PubMed publication-type flags, generally corresponding to Clinical Trials or Clinical Guidelines. These indicators were used as markers of clinical uptake or translational relevance, not as evidence of clinical efficacy or direct practice change. This step yielded 8 with policy mentions and 31 with clinical impact (N = 37 papers, given some overlap between categories). Criterion 2 identified papers with significant scholarly uptake and reach. These filters relied on four metrics: the nonprovisional Relative Citation Ratio (RCR), the Article Citation Rate (ACR), Altmetric Attention Scores (AAS), and the Journal Citation Rate (JCR). The RCR is an NIH Office of Portfolio Analysis metric of article influence representing a field- and time-normalized citation rate, benchmarked to 1.0 for a typical (median) NIH-funded paper in the same publication year and field. The nonprovisional RCR excludes generally newer papers for which citation-based metrics are less stable. ACR is the article’s number of citations per year since publication. The AAS is an automatically calculated, weighted indicator of the amount of online attention a research output receives across tracked sources (e.g., news, blogs, social media, Wikipedia, policy documents), with weights intended to reflect the reach and visibility of different sources rather than a simple raw mention count. JCR refers to a journal-level citation rate used to compare journal impact. Based on the metrics provided for the subset of candidate publications to review, papers were included if they met both: RCR (nonprovisional) ≥ 75th percentile or ACR ≥ 75th percentile; and Altmetric Attention Score ≥ 75th percentile or Journal Citation Rate ≥ 75th percentile. This dual-criterion step contributed N = 41 (with overlap across criteria acknowledged). Finally, Criterion 3 relied on single-metric outperformers with any of the aforementioned metrics at ≥ 90th percentile threshold (RCR nonprovisional, ACR, Altmetric Attention Score, or Journal Citation Rate). Applying all criteria produced a final set of N = 77 unique references for narrative review. The 75th percentile thresholds used for Criterion 2 were: RCR nonprovisional ≥ 3.68, ACR ≥ 5.73, Altmetric Attention Score ≥ 18.25, and Journal Citation Rate ≥ 5.65. The 90th percentile thresholds used for Criterion 3 were: RCR nonprovisional ≥ 5.93, ACR ≥ 11.0, Altmetric Attention Score ≥ 105.6, and Journal Citation Rate ≥ 9.05. Importantly, the impact-based filtering strategy was not intended to equate scientific importance with citation count alone. Rather, it was designed to identify papers that had achieved evidence of influence through multiple pathways, including policy or clinical relevance, field- and time-normalized citation influence, annualized citation rate, public and online attention, journal-level visibility, or exceptional performance on any single metric. This approach was selected to avoid reliance on any one indicator of impact, recognizing that different types of ABCD publications may exert influence through different channels. Nevertheless, because all such indicators are imperfect and may favor work that has had more time, visibility, or disciplinary uptake, we interpret the final set as a curated sample of high-impact ABCD FFC publications rather than an exhaustive representation of all theoretically important FFC research. Screening, data extraction, and reliability Eligible records were peer-reviewed, empirical ABCD publications indexed in major bibliographic databases. As part of final screening, co-authors (AC, MO, VC, YP) were instructed to exclude review articles, methods-only papers, opinion/commentary, or other non-empirical outputs from the set slated for full narrative synthesis. Papers were excluded when an FFC measure was embedded within a composite score alongside non-FFC variables in proportions that precluded meaningful inference about the FFC measure's contribution, but retained when FFC variables constituted the majority of the composite, allowing findings to be interpreted in the context of FFC influence. Also, if an FFC measure was included but no interpretable findings were reported as a result, the paper was also excluded. Available records were split into two subsets and double-screened by independent rater pairs. Agreement was quantified with Cohen’s κ, yielding κ = 0.725 for one rater pair (n = 38), κ = 0.538 for the second (n = 39), and an overall κ = 0.655 when combined across the full set (n = 77). Discrepancies were resolved with a third rater (MG, RG). A total of 13 papers were found not to include interpretable information from FFC measures, usually resulting from reliance on LED measures or composite scores that combined FFC measures with those of other domains. This resulted in a final data set for review of 64 papers. Among the 64 papers included in the final synthesis, the median and range of impact metrics ranged considerably given that there were multiple pathways for a paper to enter the review: RCR nonprovisional, median = 3.93, range = 0.65–14.44; ACR, median = 8.0, range = [0–33.25]; Altmetric Attention Score, median = 34, range = 0–1780; and Journal Citation Rate, median = [6.49], range = [2.15–23.94]. For each included paper, we extracted bibliographic information and summarized: (1) the focal FFC measures included, (2) outcomes and domain of inquiry, (3) sample and analytic design features, and (4) the study’s substantive conclusions and implications. Findings were synthesized narratively and organized to highlight convergent patterns, divergences, and gaps across FFC constructs and measures. Findings are organized by outcome domain and, within each outcome domain, by FFC construct grouping: Identity, Values, and Culture; Family; and Peers, School, and Community. These groupings were informed by the a priori ABCD FFC framework and adapted during synthesis based on the distribution of included studies. Results Findings for a total of 64 articles with ABCD Study data were included in the review and synthesis. The number of overall studies reviewed for each FFC level by health outcome domain is reported in Table 1. The number of studies that reported FFC variables as protective and risk factors by health outcome domain are reported in Table 2. We report a summary count of the negative and protective associations between FFC measures and health outcomes in Table 3. Out of the 64 articles included in the review, 65% of studies conducted a cross-sectional analysis while only 35% presented longitudinal data. Given that FFC measures were typically analyzed as predictors, moderators, mediators, or covariates, findings below are presented based on the primary domain of outcome measures that were examined in relation to FFC measures. Subsections group by type of FFC measures and are categorized by Identity, Values, and Culture; Family; and Peers, School, Community. Papers that had analyses with multiple outcome measures that belong to different domains may appear in multiple sections below.Table 1Number of papers by FFC levels and health outcome domains.FFC LevelsNeuroimaging & Neurocognitionn (%)Mental Healthn (%)Physical Health & Substance Usen (%)Identity, Values, & Culture7 (17.0%)11 (22.9%)5 (31.3%)Family16 (39%)23 (47.9%)6 (37.5%)Peers, School, & Community18 (43.9%)14 (29.2%)4 (25.0%)Total Count41 (64%)48 (75.0%)16 (25.0%)Note: FFC (Family, Friends, and Community) measure as outcome or other outcome not above: n = 3 (4.5%).Table 2Friends, Family, and Community (FFC) measures included in reviewed ABCD studies, organized by Bronfenbrenner ecological levels. Rows list FFC measures administered in the ABCD Study, with the proportion of reviewed papers that included each measure organized by health domains. Studies for youth and caregiver reports are reported separately.FFC DomainsFFC Variables and SubscalesNeuroimaging & Neurocognition(N = 79)Mental Health(N = 69)Physical Health & Substance Use(N = 30)Youth ReportIdentity, Values, & CulturePhenX Acculturation2 (2.5%)1 (1.4%)2 (6.7%)Unfair Treatment1 (1.2%)9 (13.0%)4 (13.3%)MEIM-R2 (2.5%)00Native American Acculturation Scale000Prosocial Behavior Subscale4 (5.1%)3 (4.3%)1 (3.3%)Values Scale1 (1.2%)00Vancouver Index of Acculturation000FamilyFamily Environment Scale: Family Conflict12 (15.2%)15 (21.7%)5 (16.7%)Family Environment Scale: Cohesion000CRPBI Warmth/Acceptance16 (20.3%)11 (15.9%)4 (13.3%)Parental Monitoring14 (17.7%)15 (21.7%)8 (26.7%)Multidimensional Neglectful Behavior Scale000Pet Ownership01 (1.4%)0HOME Cognitive Stimulation Items000Peers, School, & CommunityActivity Space000Peer Behavior Profile2 (2.5%)00Peer Network Health000Resistance to Peer Influence000School Attendance & Grades000School Risk & Protective Factors15 (19.0%)9 (13.0%)4 (13.3%)Neighborhood Safety & Crime9 (11.4%)5 (7.2%)2 (6.7%)FFC DomainsFFC Variables and SubscalesNeuroimaging & Neurocognition(N = 36)Mental Health(N = 22)Physical Health & Substance Use(N = 11)Caregiver ReportIdentity, Values, & CulturePhenX Acculturation1 (2.8%)01 (9.1%)MEIM-R4 (11.1%)01 (9.1%)Native American Acculturation Scale001 (9.1%)Prosocial Behavior Subscale3 (8.3%)2 (9.1%)1 (9.1%)Values Scale2 (5.5%)01 (9.1%)Vancouver Index of Acculturation1 (2.8%)01 (9.1%)FamilyFamily Environment Scale: Family Conflict9 (25%)9 (40.9%)2 (18.2%)Family Environment Scale: Cohesion000Family Environment Scale: Expressiveness000Family Environment Scale: Intellectual/Cultural1 (2.8%)00Family Environment Scale: Active/Recreational000Family Environment Scale: Organization000Parental Knowledge3 (8.3%)2 (9.1%)1 (9.1%)Multidimensional Neglectful Behavior Scale000HOME Cognitive Stimulation Items000Peers, School, & CommunitySchool Attendance & Grades02 (9.1%)1 (9.1%)Neighborhood Safety & Crime12 (33.3%)7 (31.8%)1 (9.1%)Community Cohesion000Table 3Number of FFC measures identified as risk or protective factors in relation to each youth outcome, summarized across ABCD Study papers reviewed.Evidence for FFC Measure as Protective or Risk Factor for Outcomes by RespondentFFC DomainsFFC Variables and SubscalesNeuroimaging & NeurocognitionMental HealthPhysical Health and Substance UseProtectiveRiskProtectiveRiskProtectiveRiskYouth ReportIdentity, Values, & CulturePhenX Acculturation2--1-1Unfair Treatment-1-7-2Prosocial Behavior Subscale12-1-FamilyFamily Environment Scale: Family Conflict-5-8-3CRPBI Warmth/Acceptance5-3-1-Parental Monitoring5-814-Pet Ownership--1---Peers, School, & CommunityActivity Space------School Risk & Protective Factors: School Environment6-511School Risk & Protective Factors: School Involvement4-52-School Risk & Protective Factors: School Disengagement-2-2-3Neighborhood Safety & Crime413-1-Caregiver ReportIdentity, Values, & CultureProsocial Behavior Subscale12-1-Values Scale: Familism Referent----1-Values Scale: Familism Support----1-Values Scale: Independence/ Self-reliance-----1Vancouver Index of Acculturation----1-FamilyFamily Environment Scale: Family Conflict-2-6-1Family Environment Scale: Intellectual/ Cultural1-----Peers, School, & CommunitySchool Attendance & Grades---1--Neighborhood Safety & Crime533-1- Table 1Number of papers by FFC levels and health outcome domains.FFC LevelsNeuroimaging & Neurocognitionn (%)Mental Healthn (%)Physical Health & Substance Usen (%)Identity, Values, & Culture7 (17.0%)11 (22.9%)5 (31.3%)Family16 (39%)23 (47.9%)6 (37.5%)Peers, School, & Community18 (43.9%)14 (29.2%)4 (25.0%)Total Count41 (64%)48 (75.0%)16 (25.0%)Note: FFC (Family, Friends, and Community) measure as outcome or other outcome not above: n = 3 (4.5%). Number of papers by FFC levels and health outcome domains. Note: FFC (Family, Friends, and Community) measure as outcome or other outcome not above: n = 3 (4.5%). Table 2Friends, Family, and Community (FFC) measures included in reviewed ABCD studies, organized by Bronfenbrenner ecological levels. Rows list FFC measures administered in the ABCD Study, with the proportion of reviewed papers that included each measure organized by health domains. Studies for youth and caregiver reports are reported separately.FFC DomainsFFC Variables and SubscalesNeuroimaging & Neurocognition(N = 79)Mental Health(N = 69)Physical Health & Substance Use(N = 30)Youth ReportIdentity, Values, & CulturePhenX Acculturation2 (2.5%)1 (1.4%)2 (6.7%)Unfair Treatment1 (1.2%)9 (13.0%)4 (13.3%)MEIM-R2 (2.5%)00Native American Acculturation Scale000Prosocial Behavior Subscale4 (5.1%)3 (4.3%)1 (3.3%)Values Scale1 (1.2%)00Vancouver Index of Acculturation000FamilyFamily Environment Scale: Family Conflict12 (15.2%)15 (21.7%)5 (16.7%)Family Environment Scale: Cohesion000CRPBI Warmth/Acceptance16 (20.3%)11 (15.9%)4 (13.3%)Parental Monitoring14 (17.7%)15 (21.7%)8 (26.7%)Multidimensional Neglectful Behavior Scale000Pet Ownership01 (1.4%)0HOME Cognitive Stimulation Items000Peers, School, & CommunityActivity Space000Peer Behavior Profile2 (2.5%)00Peer Network Health000Resistance to Peer Influence000School Attendance & Grades000School Risk & Protective Factors15 (19.0%)9 (13.0%)4 (13.3%)Neighborhood Safety & Crime9 (11.4%)5 (7.2%)2 (6.7%)FFC DomainsFFC Variables and SubscalesNeuroimaging & Neurocognition(N = 36)Mental Health(N = 22)Physical Health & Substance Use(N = 11)Caregiver ReportIdentity, Values, & CulturePhenX Acculturation1 (2.8%)01 (9.1%)MEIM-R4 (11.1%)01 (9.1%)Native American Acculturation Scale001 (9.1%)Prosocial Behavior Subscale3 (8.3%)2 (9.1%)1 (9.1%)Values Scale2 (5.5%)01 (9.1%)Vancouver Index of Acculturation1 (2.8%)01 (9.1%)FamilyFamily Environment Scale: Family Conflict9 (25%)9 (40.9%)2 (18.2%)Family Environment Scale: Cohesion000Family Environment Scale: Expressiveness000Family Environment Scale: Intellectual/Cultural1 (2.8%)00Family Environment Scale: Active/Recreational000Family Environment Scale: Organization000Parental Knowledge3 (8.3%)2 (9.1%)1 (9.1%)Multidimensional Neglectful Behavior Scale000HOME Cognitive Stimulation Items000Peers, School, & CommunitySchool Attendance & Grades02 (9.1%)1 (9.1%)Neighborhood Safety & Crime12 (33.3%)7 (31.8%)1 (9.1%)Community Cohesion000 Friends, Family, and Community (FFC) measures included in reviewed ABCD studies, organized by Bronfenbrenner ecological levels. Rows list FFC measures administered in the ABCD Study, with the proportion of reviewed papers that included each measure organized by health domains. Studies for youth and caregiver reports are reported separately. Table 3Number of FFC measures identified as risk or protective factors in relation to each youth outcome, summarized across ABCD Study papers reviewed.Evidence for FFC Measure as Protective or Risk Factor for Outcomes by RespondentFFC DomainsFFC Variables and SubscalesNeuroimaging & NeurocognitionMental HealthPhysical Health and Substance UseProtectiveRiskProtectiveRiskProtectiveRiskYouth ReportIdentity, Values, & CulturePhenX Acculturation2--1-1Unfair Treatment-1-7-2Prosocial Behavior Subscale12-1-FamilyFamily Environment Scale: Family Conflict-5-8-3CRPBI Warmth/Acceptance5-3-1-Parental Monitoring5-814-Pet Ownership--1---Peers, School, & CommunityActivity Space------School Risk & Protective Factors: School Environment6-511School Risk & Protective Factors: School Involvement4-52-School Risk & Protective Factors: School Disengagement-2-2-3Neighborhood Safety & Crime413-1-Caregiver ReportIdentity, Values, & CultureProsocial Behavior Subscale12-1-Values Scale: Familism Referent----1-Values Scale: Familism Support----1-Values Scale: Independence/ Self-reliance-----1Vancouver Index of Acculturation----1-FamilyFamily Environment Scale: Family Conflict-2-6-1Family Environment Scale: Intellectual/ Cultural1-----Peers, School, & CommunitySchool Attendance & Grades---1--Neighborhood Safety & Crime533-1- Number of FFC measures identified as risk or protective factors in relation to each youth outcome, summarized across ABCD Study papers reviewed. Neuroimaging & neurocognition The ABCD Study includes a comprehensive neurocognitive battery administered annually as well as a neuroimaging protocol administered every two years starting at baseline (see Garavan et al., 2018). Briefly, the neurocognitive protocol for ABCD includes assessments for executive function, attention, episodic memory, language abilities, inhibition, and reward processing (Luciana et al., 2018). The ABCD Study imaging protocol entailed both structural MRI and fMRI resting state and reward processing tasks (Casey et al., 2018). Identity, values, and culture Seven studies included FFC constructs at the identity-, values-, and cultural-level in association with neurocognitive development. All of these studies used cross-sectional design for analyses. In the domain of language acculturation, youth who reported bilingual use, as measured by the Youth Acculturation Survey, outperformed monolingual peers on verbal learning and exhibited distinct connectomes, with connectivity predicting working-memory performance only for multilinguals (Kwon et al., 2021). Vaughn et al. (2021) found that bilingual use among children showed widespread cortical thinning, while greater English use and vocabulary were linked to thicker cortex in fronto-temporo-parietal language regions. Together, these findings indicate that a bilingual experience and cultural-linguistic orientation shape neural systems supporting language and memory. Geckeler et al. (2022) found that fewer unfair treatment experiences were associated with greater activation in social–affective regions during implicit emotion regulation, whereas exposure to unfair treatment related to reduced middle temporal activation. Family Sixteen studies included FFC constructs at the family-level in association with neurocognitive development. Among these studies, 25% (n = 4) used longitudinal design for analyses whereas 75% (n = 12) used cross-sectional design. The findings from these studies spanned across three principal domains of family functioning: parental monitoring, parental warmth and acceptance, and family conflict, and relevant multivariate and ecological frameworks. Parental monitoring consistently emerged as a protective factor across studies. H. Zhang et al. (2020) observed that increased caregiver monitoring correlated with enhanced executive function. Gong et al. (2021) extended these findings by demonstrating that reduced monitoring predicts decreased surface area in regions integral to emotion regulation, including orbitofrontal and anterior cingulate regions. And, in cases involving subclinical obsessive-compulsive symptoms, Pagliaccio et al. (2021) indicated that lower levels of monitoring are associated with increased white matter diffusivity and disrupted functional connectivity within corticostriatal circuits. Parental warmth and acceptance, often assessed through perceived parental acceptance, are associated with adaptive connectivity patterns and help mitigate the neurobiological effects of adversity. (Brieant et al., 2021) demonstrated that higher parental acceptance at year 1 was associated with lower symptomatology at year 2, suggesting its role in the development of emotional regulation. Beck et al. (2025) indicated that emotional neglect, such as inadequate caregiver support, correlates with a younger predicted brain age, suggesting delayed maturation. These findings underscore parental warmth as a neurobiologically significant buffer during development. Family conflict has been consistently associated with neural alterations in circuits responsible for attention, emotion regulation, and cognitive control. Rakesh et al. (2021) demonstrated that neighborhood disadvantage was associated with altered resting-state connectivity across cognitive and sensorimotor networks, particularly the default mode and dorsal attention networks, and that these effects were attenuated by positive family and school environments. This suggests that positive parenting may serve as a neuroprotective factor during development (Rakesh et al., 2021). Dumornay et al. (2023) showed that racial disparities in prefrontal and limbic volumes among Black youth are largely mediated by family adversity, including conflict and material hardship. These findings position family conflict as a neurodevelopmental risk factor. Several studies incorporated family functioning within broader biopsychosocial models. Modabbernia et al. (2021) identified latent modes of brain–environment–behavior covariation, in which family-related variables, including parental characteristics and family quality of life, were consistently associated with structural and functional brain measures. Zhi et al. (2024) found that functional connectivity in control and default mode networks mediated the association between family conflict and behavior, with sleep problems acting as behavioral mediators. H. Zhang et al. (2020) showed that a composite of supportive parental warmth, prosocial behavior, reduced family conflict, and neighborhood safety predicted better cognition along with fewer impulsive behaviors. M. R. Gonzalez et al. (2020) reported that a factor of social support, including parental monitoring, family conflict, and caregiver warmth, explained additional variance in cortical surface area and cognition, even beyond the negative effect of lower household income. Sripada et al. (2022) identified parental education, a proximal family-level variable, as the strongest socioeconomic predictor of whole-brain connectivity. When considering family and environmental Interactions, contextual models emphasized that the effects of family functioning vary by broader environmental factors. Guerrero et al. (2021) found that children in low-monitoring, high-conflict families were least likely to meet movement behavior guidelines. These patterns were also associated with poorer cognitive outcomes. Rakesh et al. (2023) demonstrated that positive school environments may buffer neural effects of home adversity, particularly in networks related to self-regulation. Studies identified neural mediators linking family functioning to behavioral outcomes. Gong et al. (2021) found that cortical surface area mediated links between conflict and psychopathology. Carozza et al. (2025) reported that reduced white matter integrity mediated the relationship between early adversity, including familial conflict and economic hardship, and later cognitive deficits. Finally, Ge et al. (2023) identified three neurodevelopmental subtypes among youth with mood and anxiety symptoms, all of which showed elevated exposure to family adversity. Peers, school, and community Eighteen studies included FFC constructs at the peer-, school-, and community-level in association with neurocognitive development, as measured via neuroimaging or neuropsychological measures. Among these studies, 89% (n = 16) used longitudinal design for analyses whereas 11% (n = 2) used cross-sectional design. Results provide converging evidence that individual differences in youth neurodevelopment cannot be interpreted without considering their social ecologies. Peer experiences were highlighted in Geckeler et al. (2022) who showed that social functioning was associated with neural activation during the Emotional N-back task. Having a greater number of close friends was associated with greater activation in the fusiform gyrus, insula, temporoparietal junction, inferior parietal lobe, and superior temporal gyrus, regions implicated in socioemotional processing. School climate consistently emerged as an important correlate of neural organization. Ellwood-Lowe et al. (2021) demonstrated that the relation between frontoparietal–default network coupling and cognitive performance reverses for children in poverty and is moderated by school type and neighborhood safety, indicating that expected network architecture is environment-dependent. Rakesh et al. (2023) found that more positive school climate (availability of extra-curricular activities, perceived safety, teacher support) was associated with greater within-network connectivity of the cingulo-opercular network (CON) and reduced CON–default mode network (DMN) connectivity, patterns that in turn related to better mental health. Additional effects included increased connectivity between the auditory and ventral attention networks, and reduced frontoparietal–retrosplenial temporal network connectivity, though these did not relate directly to mental health. Structural brain measures did not show significant associations with school climate. Hackman et al. (2022) extended these findings by showing that school climate was in fact associated with cortical thickness and subcortical volume, but specifically for youth experiencing socioeconomic disadvantage. These results indicate that school environments may interact with contextual stress to shape brain development. Research conducted thus far with ABCD data confirms that community-level and socioeconomic factors have a strong impact on neurodevelopment and neurocognition. The findings of H. Zhang et al. (2020) highlight the importance of community-level factors on cognitive function. They found that higher socioeconomic status (parental education, income, financial security) related to greater general cognition and executive function (but less behavioral inhibition) while greater quality of youth proximal social environments (e.g., neighborhood safety, school environment) related to less impulsive behavior and behavioral inhibition (but greater behavioral activation). ABCD research further suggests that the effects of community-level factors on neurodevelopment may serve as a mechanism underlying impacts on cognitive function and mental health. These studies have utilized functional measures of brain activity and connectivity, as well as structural measures of gray and white matter. Rakesh et al. (2021) found that neighborhood disadvantage was associated with altered resting-state connectivity across DMN, dorsal attention, and sensorimotor networks, reflecting widespread differences that predicted lower cognition and worse mental health. Positive family and school environments again buffered these effects. Sripada et al. (2022) similarly reported that socioeconomic resources, especially parental education, related to brain-wide differences in intrinsic functional architecture, with effects moderated by home enrichment, youth academic performance, and cognitive ability. Using graph-theory, Jeong et al. (2024) found that greater interpersonal support, a latent subfactor that consisted of school connectedness, caregiver warmth, and parental monitoring, was associated with lower whole-brain modularity during the stop-signal task and differential efficiency in wiring in the visual 2 network during the stop-signal task, which the study notes could reflect adaptive neurodevelopmental responses to specific environmental contexts. Zhi et al. (2024) showed that community exposures were stronger predictors of mental health and cognition than connectivity alone, examining not only cross-sectional associations at baseline but also at one-year follow-up. Connectivity in the DMN and cognitive control network (CCN) best predicted cognitive ability, while subcortical network connectivity (e.g., thalamus–temporal and thalamus–postcentral connections) best predicted mental health. Connectivity partially mediated associations between environmental exposures and outcomes. Similarly, (Ellwood-Lowe et al., 2021) showed that environmental conditions moderated the association between brain network coupling (between lateral frontoparietal network and default mode network) and cognitive function. These results highlight how the association between brain and behavior may depend on the socioeconomic context of the youth. In regard to structural neuroimaging, (M. R. Gonzalez et al., 2020) found that among 9–10 year old youth, socioeconomic disadvantage predicted reduced total cortical surface area and lower cognition, while latent factors for resource-to-adversity (including, FFC measures for caregiver warmth, parental monitoring, neighborhood safety, school disengagement, and non-FFC measures for duo-parent households, parental education, financial hardship, sleep) and social support (i.e., FFC measures for caregiver warmth, parental monitoring, school disengagement, school environment, and neighborhood safety) predicted greater total cortical surface area and better cognitive performance. Dumornay et al. (2023) demonstrated that race-related differences in amygdala, hippocampal, and prefrontal volumes were largely eliminated after accounting for adversity exposure. Utilizing canonical correlation analysis, (Alnæs et al., 2020) found that a component correlating lower neighborhood safety with lower cognitive performance, among many other factors, mapped onto reduced cortical thickness and surface area, consistent with the theory that accelerated maturation under chronic stress. While there has been less research focused specifically on white matter structure, (Carozza et al., 2025) demonstrated associations between adversity exposures (social vulnerability, trauma, financial strain, family conflict) and widespread reductions in fractional anisotropy (FA). Lower FA mediated links between adversity and later cognitive performance. (Alnæs et al., 2020) also found urbanicity factors associated with microstructural properties in the parahippocampal cingulum, uncinate fasciculus, corpus callosum, and forceps minor. Numerous researchers have taken advantage of ABCD’s multimodal neuroimaging to support a more holistic understanding of the impact of community environmental factors on neurodevelopment. (Beck et al., 2025) showed differential effects of adversity dimensions on brain-age gap estimates (calculated utilizing T1-weighted structural, diffusion tensor and resting-state functional neuroimaging). Specifically, emotional neglect predicted younger brain age, whereas cumulative adversity (family conflict, trauma, SES disadvantage) predicted older brain age. Modabbernia et al. (2021) used canonical correlation analyses to identify multivariate patterns linking environmental exposures to diffuse alterations in neurodevelopment (cortical thickness, surface area, subcortical volume, gray/white matter contrast, white matter integrity, and resting state connectivity), with the most robust predictors including socioeconomic circumstances, quality of parental engagement, perinatal events, cognitive ability, and psychopathology. Ge et al. (2023) used Heterogeneity Through Discriminative Analysis (HYDRA) multimodal clustering (incorporating regional morphometry, neurite density, intracortical myelination) to subtype youth with mood/anxiety symptoms as compared to typically developing youth. All subtypes reported higher family conflict, adverse events, and bullying. The subtype with globally lower cortical thickness and surface area and higher myelination/neurite density had the highest adversity burden and showed distinct profiles in school environment, parental engagement, and peer support (including reduced positive school climate, lower parental engagement, more grade decline, and fewer regular friends). Overall, these findings highlight a converging theme: peer, school, and community factors are deeply intertwined with youth neurodevelopment, cognition, and mental health. Positive environments buffer adversity, while environmental stress—including neighborhood disadvantage, family disengagement, and weak school climate—is consistently associated with less optimal structural and functional brain patterns and elevated risk for behavioral and emotional difficulties. Further, findings suggest that how neurodevelopment relates to positive or negative mental health or cognitive outcomes may depend on the socioeconomic context of the youth. These results strongly support ongoing efforts within ABCD to prioritize social ecological variables as key developmental determinants and moderators of brain-behavior relationships. Results are limited by the primarily cross-sectional nature of analyses. Future work is needed that leverages that longitudinal nature of the ABCD Study to identify how different socioenvironmental factors influence neurodevelopment and cognitive functioning over time. Mental Health The ABCD Study protocol included assessments of youth mental health through youth-reported and parent-reported annual questionnaires (see Barch et al., 2018). Constructs assessed include psychopathology (e.g., externalizing and internalizing behaviors), emotional functioning, exposure to trauma, family history of mental health and screening for other psychiatric disorders based on the DSM criteria. Identity, values, and culture Eleven studies included FFC constructs at the identity-, values-, and cultural-level in association with mental health outcomes. Among these studies, 27% (n = 3) used longitudinal design for analyses where as 73% (n = 8) used cross-sectional design. Identity and cultural-context-linked exposures were found to consistently predict youth mental health across the studies. Exposure to unfair treatment was prominently a risk factor in adolescent mental health. Specifically, disparities in psychotic-like experiences (PLEs) during middle childhood between racial/ethnic groups were attenuated when accounting for youths’ unfair treatment experiences, identifying exposure to unfair treatment as a salient social correlate of PLEs (Karcher et al., 2022). In eating disorders, early adolescents reporting unfair treatment due to identifying with a race or ethnicity had approximately threefold higher odds of binge eating disorder, with student-perpetrated unfair treatment most clearly linked to both diagnosis and binge eating behaviors (Raney et al., 2023). Moreover, 10–11-year-olds experiencing unfair treatment in multiple domains (e.g., weight plus other types of unfair treatment) showed incrementally greater odds of disordered eating, underscoring intersectional risk (Boswell et al., 2024). Work centered on sexual minority (SM) youth indicates that interpersonal SM unfair treatment and state-level structural stigma contribute to elevated psychopathology and suicidality, beyond non-identity specific stressors (Gordon et al., 2024). Consistent with interpersonally targeted stressors, cyberbullying victimization was associated with suicidal ideation and attempts after accounting for offline peer aggression (Argabright, Tran, et al., 2022). Additionally, stress related to exposure to unfair treatment due to racial/ethnic identity was independently associated with suicidality in preadolescents, and data demonstrate Black youth were disproportionately exposed (Argabright, Visoki, et al., 2022). In addition, we found evidence to support that social and familial contexts also shape risk and resilience. Using ABCD baseline data, parental psychopathology, socioeconomic status, and the proximal social environment were the strongest correlates of children’s psychopathology and cognition (H. Zhang et al., 2020). Investigations employing a dimensional approach demonstrated the general “p factor”, defined as a latent factor capturing shared variance between internalizing and externalizing disorders, exhibited diffuse associations spanning temperament, cognition, family history, and social adversity, while internalizing and externalizing factors retained more domain specific signatures when p was modeled (Brislin et al., 2022). During the early COVID 19 period, pre-pandemic adverse childhood experiences (ACEs) forecasted poorer mood and mental health; race/ethnicity did not strengthen ACE-symptom links overall, but minoritized youth reported more pandemic-related unfair treatment and worry, and girls reported more negative symptoms (Stinson et al., 2021). Taken together, these ABCD reports indicate that identities (racial/ethnic, sexual minority, body weight) and culturally structured contexts (structural stigma, unfair treatment, socioeconomic position, and online/offline peer ecologies) are tightly coupled to transdiagnostic burden and specific symptom domains, with convergent evidence from social, dimensional, and exposomic frameworks. Family Twenty-three papers included in the review linked a family-level construct in ABCD to teen mental health. Among these studies, 48% (n = 11) were cross-sectional and 52% (n = 12) were longitudinal. Measures for parental warmth, parental monitoring, family conflict, and pet ownership were included, with 10 of the 23 studies used all three of the warmth, monitoring, and conflict measures. There were 7 papers that treated the family measure as a predictor of interest. (Brislin et al., 2022) studied cross-sectional predictors of the ‘p factor’ at ages 9–10 and found that p was positively correlated with family conflict and negatively correlated with parental monitoring. (Zhi et al., 2024) studied brain-by-environment interactions at ages 9–10 in predicting teens’ mental health 1 year later, finding that family conflict was one of the top 5 environmental predictors of internalizing and externalizing problems. (Hill et al., 2025) used data spanning ages 9–14 and developed prognostic models to forecast changes in mental health over a 1-year timeframe. Like (Zhi et al., 2024), they found that family conflict was the 4th most important predictor of a participant being classified as high-risk for mental health problems. DeVille et al., 2020 and Janiri et al. (2020) studied predictors of teen suicidality, both using cross-sectional data from ages 9–10. Janiri et al. (2020) showed that less parental warmth, less parental monitoring, and higher family conflict were all associated with higher suicidality risk among youth, and (DeVille et al., 2020) showed that monitoring and conflict were also associated with non-suicidal self-injury. Guerrero et al. (2021) performed a latent profile analysis of warmth, monitoring, and conflict measures at ages 9–10 and related them cross-sectionally to the number of sleep, physical activity, and screentime levels recommended for children by the Canadian 24-h Movement Guidelines. The latent profile of families who were high on warmth and monitoring and low in conflict were most likely to meet recommendations. (Mueller et al., 2022) studied whether youth with pets had better mental health during the COVID−19 pandemic. They did not find a buffering impact of pets, although those with dogs were more likely to have a walking routine and spend time outdoors. There were 3 papers that treated the family measure as a mediator/moderator of interest. (Argabright, Tran, et al., 2022) tested whether parental monitoring or family conflict mediated the association between household wage loss during the COVID−19 pandemic and youth depressive symptoms, finding significant mediation for increased conflict but not for monitoring. Brieant et al. (2021) tested whether parental warmth moderated the association between adversity and cortico-limbic rsFC between ages 9–12, finding no such moderation. Finally, (Hamatani et al., 2022) studied parental monitoring as a moderator of the impact of the COVID−19 pandemic on youth mental health, finding those with lower monitoring were more likely to show worsening mental health. Other work employing a neural network approach found that family context had greater predictive probability of all mental health risk (i.e., p-factor scores) compared to either the school or neighborhood contexts (Hill et al., 2025). In study on depression and brain structure, it was reported that discrepancies between youth and caregiver reports of depressive symptoms were significantly related to negative socio-environmental factors such as family conflict, highlighting the interpretive complexity of multi-informant data in developmental psychopathology research (Shen et al., 2021). The remaining 12 of the 23 papers used the family measures in such a way that they did not yield empirical findings specific to the family variables. One paper included monitoring and conflict only as nuisance covariates, with parental monitoring being associated with lower odds of suicidality and conversely higher family conflict being associated with higher suicidality (Arnon et al., 2022). Another paper included warmth, conflict, and monitoring as inputs to a PCA model constructing predictors of neurodevelopmental outcomes, but none of the resulting components ended up corresponding to a measure of family (H. Zhang et al., 2020). Finally, the remaining 10 papers included family measures as one of many inputs to a composite exposure variable. The composites indexed adversity (Baldwin et al., 2022, Chu et al., 2022, Petrican and Fornito, 2023, Raney et al., 2024, Stinson et al., 2021, Trompeter et al., 2024), maltreatment (Warrier et al., 2021), stressful environments (L. Zhang et al., 2023), or the “exposome” (Choi et al., 2022, Hoffman et al., 2024), and most papers related them to a youth mental health outcome. Often, these use-cases involved reverse-scoring or relabeling items from the family scales to reflect different forms of “neglect.”. Peers, school, and community There were 14 ABCD studies reviewed that included peer-, school-, or community-level measures in relation to youth mental health. Among these studies, 36% (n = 5) used cross-sectional data and 64% (n = 9) used longitudinal data. These studies take a fine-grained approach to understanding the impact of specific socialization contexts on youth mental health. For example, adverse school environments (i.e., low involvement, greater disengagement) was linked to mental health problems, both at baseline (Brislin et al., 2022) and longitudinally (Brislin et al., 2022, Zhi et al., 2024). Conversely, a positive school climate (i.e., higher engagement) was related to lower levels of internalizing and externalizing symptoms (Hackman et al., 2022). Similarly, school engagement was protective against youth suicidality (Janiri et al., 2020). Other work indicates that a decline in school performance (i.e., decrease in school grades) was associated with the initial emergence of a “high risk” prodromal-like state of psychosis (Smucny et al., 2024). Several studies examined the association of neighborhood risk factors on PLEs. This work supported the association between PLEs and neighborhood risk (i.e., urbanicity, deprivation, and to a lesser extent years at current residence) and that global structural brain metrics (especially volume) may mediate this link (Karcher et al., 2021). Subsequently, Karcher and colleagues (2022) demonstrated that environmental adversity (e.g., neighborhood safety, years at current residence, neighborhood deprivation) collectively, as well as fewer years at current residence, specifically, were associated with persistent distressing PLEs. Lastly, one study included school environment as a covariate in investigating the associations of cyberbullying and suicidality, reporting that more protective school environments were correlated with lowers suicidality (Arnon et al., 2022). Given the multitude of environmental factors that contribute to adolescent mental health, some researchers have determined that the exposome (i.e., the totality of environmental exposures including family and neighborhood level factors) is useful for comprehensively characterizing risk and protective social factors. Two studies demonstrated that exposomic burden, which encompassed the family and neighborhood environment, was associated with mental health problems via allostatic load (i.e., physiological wear and tear on the body due to chronic stress) (Hoffman et al., 2024) and inhibition-relevant functional brain alterations (Petrican and Fornito, 2023) while accounting for polygenic risk. Zhang and colleagues (2023) also demonstrated that sleep problems mediate the link between proximal contextual risk factors (e.g., school and neighborhood) on impulsivity. Other studies indicated that proximal contextual risk factors (e.g., peer, school) were not only associated with impulsive behavioral problems (H. Zhang et al., 2020), but they had the largest effect of externalizing behavior problems compared to both broader contextual factors (e.g., neighborhood) and life events (e.g., injury, illness; (Choi et al., 2022). Physical health and substance use The ABCD Study assessed youth physical health, including anthropomorphic (e.g., weight and height) and sleep measures (Barch et al., 2018). The ABCD Study also annually assessed lifetime history and past year use of substances among youth via youth self-report (see Barch et al., 2018). Identity, values, and culture There were 5 articles for which identity and values-based constructs were investigated as predictors of physical health and substance use. Specifically, these were: perceived unfair treatment (n = 4), acculturation (n = 2), cultural values (n = 1), ethnic identity (n = 1), and prosocial behavior (n = 1). All, except one of these studies, used longitudinal data. Findings on perceived unfair treatment and physical health were mixed. Perceived racial unfair treatment was positively associated with adiposity, including body mass index and waist circumference (Cuevas et al., 2023), and drastically increasing screen time use (Shao et al., 2023) but not significant for total sleep time (Giddens et al., 2022). Regarding substance use, youth-reported acculturation (bilingual status) was associated with an increased likelihood of substance use initiation by age 12 (Green et al., 2024). In a baseline subsample, several factors differentiated adolescents who were classified as at higher or lower risk for cannabis use (Zucker et al., 2018). These protective factors included parent-reported adherence to heritage and mainstream culture, as well as prosocial behavior that were youth-reported and parent-reported. Other variables like acculturation, cultural values, and multigroup ethnic identity did not exhibit discriminability between these groups. Family Six studies included FFC constructs at the family-level in association with physical health and substance use. Among these studies, 33% (n = 2) used cross-sectional data and 66% (n = 4) used longitudinal data. Across the reviewed studies, family relationships and parenting practices emerged as important factors associated with adolescent substance use and were included as predictors of substance use and physical health. Family-focused measures at baseline distinguished youth identified at higher risk for later cannabis use: youth identified at lower-risk reported higher parental acceptance, monitoring, and familism, whereas youth identified as being at higher risk showed more family conflict (Zucker et al., 2018). When examining direct associations with substance use, parenting behaviors, including parental rules around alcohol, nicotine, and cannabis, substance use availability, and parental monitoring, were among the most robust predictors of early substance initiation, performing only second to sociodemographic characteristics and outperforming hormonal, neurocognitive, and neuroimaging indicators (Green et al., 2024). Although primarily a methodological paper, (McCabe et al., 2022) modeled parental monitoring in an applied example, demonstrating that its protective effect on alcohol sipping was amplified under higher school disengagement and varied by sex and race/ethnicity. Another ABCD investigation found that caregiver monitoring reduced early adolescent substance use, primarily by preventing use opportunities rather than through punishment, indicating a deterrence rather than enforcement pathway (Pelham et al., 2024). Family processes were also associated with some aspects of physical health. Greater sleep disturbance among sexual minority youth was partially explained by higher family conflict and lower parental monitoring (Nagata et al., 2023). In assessing risk for concussion and head or neck injury, (Saadi et al., 2024) used an index of ACEs that included several FFC family measures and found higher ACE exposure increased injury risk, although the FFC measures were not uniquely associated with these outcomes. Overall, the findings underscore the important role of family environments in early substance use and suggest possible links with some aspects of physical health. However, most studies were cross-sectional, excluding Green et al. (2024) and Pelham et al. (2024), highlighting the need for longitudinal work to clarify directionality and developmental timing. Peers, school, and community There were 4 studies that included peer-, school-, and community-level constructs as predictors and consequences of substance use initiation during adolescence, focusing on the interplay of sociodemographic, neurocognitive, environmental, and social factors. Among these studies, 50% (n = 2)used cross-sectional data and 50% (n = 2)used longitudinal data. Green et al. (2024) identify parental religiosity, race, income, and prenatal substance exposure as significant predictors of early substance use initiation. Modifiable factors such as sensation seeking, peer influence, and parenting behaviors are highlighted as key intervention targets. Ma et al. (2025) show that parental monitoring, peer influence, as well as impulsivity as major contributors to substance use initiation. Their findings underscore the need for early prevention strategies tailored to high-risk groups to mitigate substance use risks. McCabe et al. (2022) demonstrated that school disengagement reduced the protective effects of parental monitoring on alcohol sipping and varied by sex and race/ethnicity. Paulich et al. (2021) integrate neuroimaging data to explore structural brain differences associated with risk-taking behaviors, highlighting the importance of addressing both social factors and neurodevelopmental vulnerabilities. Zucker et al. (2018) focus on FFC influences on adolescent development, with a focus on family dynamics, socioeconomic status, and neighborhood safety as critical determinants of risk and protective factors for substance use initiation. Together, these studies show the importance of understanding environmental and biological vulnerabilities, while considering complex cultural contexts and stressors. These studies provide a foundation for developing targeted policies aimed at reducing adolescent substance use risks and promoting mental health resilience across diverse populations. FFC as outcomes FFC variables have been investigated in relation to other adolescent social and academic outcomes, including disciplinary action. Fadus et al. (2021) examined youth- and parent-reported family conflict as predictors of receiving school disciplinary actions. Family conflict was not significantly associated with higher odds of disciplinary actions; instead, other factors, such as race and absence of a secondary caregiver, were stronger contributors. We found two studies that investigated FFC variables as outcomes by examining their patterns and predictors. Nagata et al. (2021) reported on the prevalence of child-reported racial and ethnic perceived unfair treatment, that is the perception of being treated unfairly or feeling unaccepted in society due to their racial or ethnic background. Approximately 5% of children reported experiencing perceived unfair treatment, with Black children reporting the highest prevalence (10%), and most race/ethnic groups reporting higher odds of perceived racism compared to their White peers (except for the “Other” race/ethnic category). Using latent profile analysis to identify unique neurocognitive profiles, (Menu et al., 2025) found that early-term or preterm children had worse parent-reported academic grades as they scored at or below the average on neurocognitive tasks. From a socio-ecological perspective, these factors likely play an important role at the interpersonal and organizational level and may serve as meaningful contributors to child development. Discussion Our review highlights some of the most impactful research conducted with the ABCD Study dataset linking contextual measures of family, friends, and community to adolescent health outcomes. We synthesized findings from 64 studies examining FFC measures across three contextual levels: (1) identity, values, and culture, (2) family, and (3) peer, school, and community, in relation to adolescent health outcomes. The overall goal of this review was to describe how FFC processes operate across levels to shape adolescent neurobiological, mental health, and physical health outcomes, as well as the development of FFC constructs themselves. Across outcome domains, FFC factors at multiple levels were consistently associated with adolescent functioning, though the strength and nature of these associations varied by domain. Family-level processes emerged as the most robust and transdiagnostic predictors across domains, followed by peer-, school-, and community-level factors and then identity and cultural values. Together, these findings underscore the importance of considering adolescent health as embedded within interconnected social contexts rather than isolated domains of risk or resilience. Neuroimaging and neurocognition Studies examining neuroimaging and neurocognitive outcomes most frequently identified associations with family-level factors, including family conflict, caregiver warmth, and parental monitoring, which were linked to brain structure, functional connectivity, and cognitive performance (Beck et al., 2025, Carozza et al., 2025, Ge et al., 2023, Modabbernia et al., 2021). Parental monitoring and warmth promoted adaptive neurodevelopment, while family conflict disrupted connectivity and structure. Many of these effects remain significant after adjusting for socioeconomic status, suggesting that proximal family dynamics contribute uniquely to developmental outcomes. Identity and values-based measures, such as prosocial values and self-perceived unfair treatment, were also associated with neural systems supporting emotion regulation, reward processing, and executive functioning, suggesting that self-concept and cultural context may shape neurodevelopmental pathways (Geckeler et al., 2022). Peer, school, and community contexts, such as peer relationships, school connectedness, and neighborhood safety, were found to positively influence cognition and brain structure, highlighting the multilevel nature of the impact of social contexts on youth neurocognitive development (Alnæs et al., 2020, Geckeler et al., 2022, Hackman et al., 2022, Rakesh et al., 2023; Zhang et al., 2020). These findings support integrating family-level interventions into developmental neurocognitive models to promote resilience and reduce risk. Mental health Across studies focusing on mental health outcomes, FFC factors at all three levels were implicated, with most studies including constructs at the family-level and at the peer-, school-, and community-level. At the identity and values level, higher reported unfair treatment was a risk factor while higher youth prosocial values were protective for internalizing symptoms, well-being, and stress-related outcomes. Family-level variables, including family conflict, caregiver warmth, and parenting monitoring, demonstrated robust associations with both internalizing and externalizing symptoms (Brislin et al., 2022; H. Zhang et al., 2020). Higher family conflict was consistently reported as a risk factor for youth mental health, while higher caregiver warmth and parental monitoring were protective (Brislin et al., 2022, Guerrero et al., 2021, Hill et al., 2025, Zhi et al., 2024). Peer, school, and community contexts, such as peer relationships, school connectedness, and neighborhood safety, were found to positively influence mental health outcomes (Arnon et al., 2022, Brislin et al., 2022, Hackman et al., 2022, Janiri et al., 2020, Karcher et al., 2021, Karcher et al., 2022, Smucny et al., 2024; H. Zhang et al., 2020; Zhi et al., 2024). Physical health and substance use Studies examining physical health and substance use outcomes placed greater emphasis on peer, school, and community contexts, particularly peer norms, school environments, and neighborhood opportunity structures. Family-level factors continued to play an important role, especially in relation to substance use initiation and escalation (Green et al., 2024, McCabe et al., 2022, Pelham et al., 2024, Zucker et al., 2018). Identity and values-based factors were less frequently examined in this domain and were often indirectly associated with outcomes through family or peer processes, suggesting potential mediation pathways that warrant further investigation. Methodological and measurement considerations The literature exhibited several limitations, including uneven representation of FFC levels, limited longitudinal and multilevel modeling, and inconsistent measurement of social context. Even fewer studies examined FFC measures as outcomes, with most treating FFC primarily as an exposure or contextual covariate. Neuroimaging studies, in particular, underutilized peer, school, and community-level measures, while studies of physical health and substance use often lacked integration with neurobiological outcomes. These patterns should be interpreted considering our publication filtering approach. Although more than 300 publications include FFC-related measures, the present review necessarily focused on a smaller, curated subset based on multiple impact and relevance metrics. Although the use of multiple indicators was intended to reduce bias by avoiding overreliance on any single metric, no impact-based strategy is exhaustive or neutral. Indeed, several important FFC measures were represented only sparsely in the final group of included articles. Such filters may underrepresent newer publications, emerging areas of inquiry, or studies focused on topics that have not yet accrued substantial citation, public-attention, policy, or clinical-impact metrics. This limitation may be especially relevant for constructs such as family values, acculturation, ethnic identity, perceived unfair treatment, and social determinants of health. Importantly, the filtering strategy did not categorically exclude these areas; several included studies examined perceived unfair treatment, identity-related stressors, and related social-context measures. Similarly, family cultural values, assessed in ABCD appeared in relatively few included studies. However, the Values Scale, which includes modified subscales form the Mexican-American Values Scale yielded few results in our subset. Similar patterns were observed for several other FFC measures, including ethnic identity, Native American acculturation, community cohesion, peer network health, and other constructs that may be theoretically important but have not yet generated a large high-impact publication base. Thus, the limited representation of these measures should not be interpreted as evidence of limited developmental relevance. Rather, it reflects the current distribution of ABCD publications captured by our review criteria and highlights opportunities for future research using the broader FFC protocol. The present review focused on studies in which FFC measures were substantively evaluated. As a result, it does not provide a comprehensive synthesis of studies focused primarily on LED-derived indicators, such as geocoded measures of neighborhood deprivation, residential segregation, built and natural environments, environmental exposures, school characteristics, or state-level policy environments. Indeed, the scope of this review should be distinguished from the broader ABCD literature on structural context. Many of these studies rely primarily on externally linked or geocoded indicators curated through ABCD’s LED working group rather than FFC. These studies are important for understanding adolescent development in structural context, but they were not the primary target of the present review unless FFC measures were substantively evaluated and interpretable. Furthermore, several studies incorporated FFC measures as part of composite scores or latent factors, which limited our ability to identify the unique contribution of specific FFC measures to associations with adolescent health outcomes. This boundary likely contributed to the limited representation of some structural-context studies and highlights the need for future integrative work. Nevertheless, the review points to the need for more sustained ABCD research that treats structural and culturally patterned contexts as developmental mechanisms rather than background covariates. Implications for research A central implication of this review is that future ABCD research should move beyond single-level models of social context. The reviewed literature shows that family-level variables have been used frequently and productively, but comparatively fewer studies have focused on culturally specific values or their interactions in linking social position, proximal experience, and adolescent health. Future analyses should therefore test models that explicitly distinguish proximal rearing processes from broader neighborhood, cultural, and structural contexts; examine whether family, peer, and school processes mediate or moderate the effects of structural factors; evaluate culturally specific protective processes such as familism, ethnic identity, bicultural orientation, and community cohesion; and model complex exposures rather than treating race, ethnicity, income, neighborhood disadvantage, and unfair treatment as interchangeable covariates (Cardenas-Iniguez and Gonzalez, 2024, Gonzalez et al., 2025). Future studies hoping to model more complex exposures should consider applying intersectionality frameworks (Crenshaw, 1991) to investigate how multiple intersecting identities and exposures, across social positions (e.g., race and ethnicity, sex, and socioeconomic status), determine the relationships between family, friend, and community factors and youth outcomes. Applying intersectionality frameworks in ABCD research can help researchers contextualize and assess the generalizability of findings by revealing how intersecting social positions and exposures structure developmental outcomes, producing more precise knowledge about which intervention components could work for whom and when. The longitudinal structure of ABCD makes answering these questions possible because investigators can examine whether structural and cultural contexts shape developmental trajectories directly, indirectly through proximal social experiences, or conditionally across subgroups and developmental periods. Future investigations of adolescent health with the ABCD Study dataset will be able to leverage longitudinal data spanning ages 9–18 years of age. Only a third of the studies in our narrative investigated longitudinal outcomes, and only three of the longitudinal investigations included neurocognitive outcomes. Longitudinal investigations will be necessary to establish causal inference and uncover mechanistic pathways between multi-level contextual factors and adolescent health outcomes (Kim et al., 2019). In addition, as ABCD participants continue to mature, future research should examine longitudinal trajectories, critical periods, and mechanisms through which interconnected contexts, including individual, family, friends and community shape adolescent brain development. The ABCD dataset presents an unparalleled opportunity to investigate the socio-ecologically connected contexts in predicting adolescent development. Adolescent development occurs within a multi-level context; thus, future analyses should consider how these FFC constructs interact to contribute to either risk or resilience for adolescent health outcomes. For example, cumulative contexts such as exposure to multiple chronic stressors like parental unemployment, low family income, and high neighborhood crime magnify the risk for poor outcomes, including substance involvement and academic failure in adolescence and young adulthood (Mason et al., 2019). Similarly, the buffering effects of protective factors like parental monitoring and engagement is amplified when combined with strong neighborhood collective efficacy and can also increase resilience against the influences of the risk factors like ACEs (Davis et al., 2019, Sharma et al., 2019). Conclusion Taken together, this body of work demonstrates that adolescent health outcomes across neurocognitive, mental health, and physical health domains are influenced by identity, family, peer, school, and community contexts. Advancing adolescent health science will require moving toward integrated and developmentally informed models that center multi-level context as a core driver of adolescent development. The ABCD Study dataset will continue to provide opportunities for which scientists could consider such multi-level models to elucidate modifiable social factors, including family, friends, and community contexts, that are promotive for neurocognitive and behavioral health outcomes among youth. CRediT authorship contribution statement Rada K. Dagher: Writing – review & editing, Writing – original draft. William E. Pelham: Writing – review & editing, Writing – original draft. Gayathri J. Dowling: Writing – review & editing, Project administration, Conceptualization. Elisa M.Trucco: Writing – review & editing, Writing – original draft, Conceptualization. Jana R. Hong: Writing – review & editing. Robin Aupperle: Writing – review & editing, Writing – original draft, Conceptualization. Erin L. Thompson: Writing – review & editing, Writing – original draft, Funding acquisition. Laika Aguinaldo: Writing – review & editing, Writing – original draft. Ka I. Ip: Writing – review & editing, Writing – original draft. Arlyn Cuan: Writing – review & editing, Investigation, Formal analysis, Data curation. Evan White: Writing – review & editing, Writing – original draft. Diana Alkire: Writing – review & editing, Writing – original draft. Yailene Perez: Writing – review & editing, Formal analysis, Data curation. Melissa Orama: Writing – review & editing, Investigation, Formal analysis, Data curation. Alejandra Fernandez: Writing – review & editing, Writing – original draft. Brynn Tepp: Writing – review & editing, Writing – original draft. Bruce McCandliss: Writing – review & editing. Susan F. Tapert: Writing – review & editing, Funding acquisition, Conceptualization. Michael Mason: Writing – review & editing, Writing – original draft. Raul Gonzalez: Writing – review & editing, Visualization, Supervision, Methodology, Formal analysis, Data curation, Conceptualization. Carlos Cardenas-Iñiguez: Writing – review & editing, Writing – original draft. Kazi Priyanka Silmi: Writing – review & editing. Victoria Castillo: Writing – review & editing, Formal analysis, Data curation. Marybel Robledo Gonzalez: Writing – review & editing, Writing – original draft, Visualization, Supervision, Methodology, Formal analysis, Data curation, Conceptualization. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.