An official website of the United States government
Here's how you know
Official websites use .gov
A
.gov website belongs to an official
government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you've safely
connected to the .gov website. Share sensitive
information only on official, secure websites.
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with,
the contents by NLM or the National Institutes of Health.
Learn more:
PMC Disclaimer
|
PMC Copyright Notice
. 2025 Mar 1;2025:225.
Air Pollution Exposure, Prefrontal Connectivity, and Emotional Behavior in Early Adolescence
Air Pollution Associated with Changes in Brain Connections, but Not Emotions in Preadolescent Children
This Statement, prepared by the Health Effects Institute, summarizes a research project funded by HEI and conducted by Dr. Megan M. Herting at the University of Southern California, Los Angeles, and her colleagues. Research Report 225 contains the detailed Investigators’ Report and a Commentary on the study prepared by the HEI Review Committee.
This study examined whether childhood and prenatal exposure to residential outdoor fine particulate matter (PM2.5) and nitrogen dioxide (NO2) was associated with neurodevelopment over a 1-year period in a nationally representative cohort of children transitioning to adolescence in the United States.
Neurodevelopment was assessed by brain imaging of white matter connectivity in the prefrontal cortex and measures of emotional behavior.
Higher childhood NO2 exposure was associated with less white matter connectivity, which might indicate poorer brain development. Childhood PM2.5 exposure was not associated with potentially harmful changes in white matter connectivity, and air pollution was not associated with worsening emotional behavior. Results were similar when adjusting for prenatal exposure.
Outdoor air pollution exposure in childhood might affect brain development, but additional longitudinal research is needed to determine whether these changes lead to clinical symptoms.
BACKGROUND
Early-life air pollution exposure has been associated with neurodevelopmental disorders such as autism and behavioral and mental health problems, which affect about 8.5% of children in the United States. Such impairments in thinking and mental health are lifelong and can predispose individuals to diseases such as Alzheimer disease. Recently, advances in brain imaging have enabled the detection of subtle changes in brain structure and function of children exposed to air pollution. Considerable research has focused on prenatal and early childhood, yet brain development continues through early adulthood, and less is known about how air pollution exposure might affect older children.
To evaluate the effects of early adolescent exposure to air pollution on brain development, HEI funded a study by Dr. Megan Herting of the University of Southern California titled “Air Pollution Exposure and Prefrontal Connectivity in Early Adolescence” in response to HEI’s Request for Applications 18-2: Walter A. Rosenblith New Investigator Award. Dr. Herting proposed to examine whether exposure to two major air pollutants, outdoor fine particulate matter (PM2.5) and nitrogen dioxide (NO2), are associated with changes in structural connectivity in the prefrontal cortex of the brain — an area that is involved in numerous high-level cognitive processes — and with measures of emotional health over a 1-year period in children who are transitioning to adolescence.
APPROACH
Herting and colleagues hypothesized that higher air pollution exposures would be related to decreased connectivity in the white matter of the prefrontal cortex and increased problems with emotional behavior. To evaluate this hypothesis, they leveraged data from the Adolescent Brain Cognitive Development (ABCD) Study, a nationally representative cohort study of nearly 12,000 children, ages 9–10, in the United States. The ABCD Study included 21 research institutes across 17 states and began in 2016.
The investigators estimated daily ambient PM2.5 and NO2 concentrations at a 1-km resolution across the continental United States. They used hybrid machine learning models that combined satellite-based aerosol optical depth, land use information, chemical transport models, meteorological data, and other inputs. One-year average air pollution concentrations were then linked to the geocoded home address reported at the first study visit (2016–2018), and the preceding 9-month average air pollution concentrations were linked to the geocoded home address on the date of birth.
Herting and colleagues evaluated white matter microstructure in seven tracts of the prefrontal cortex using diffusion tensor MRI (magnetic resonance imaging). White matter tracts are bundles of myelinated axons that connect neurons to facilitate communication between different brain areas. The MRI measured water diffusion through brain tissue using two metrics: mean diffusivity and fractional anisotropy. Higher fractional anisotropy and lower mean diffusivity generally indicate better white matter organization and integrity, whereas lower fractional anisotropy and higher mean diffusivity can indicate disease. Caregivers also reported emotional behavior problems using the Child Behavior Checklist, which provides scores of anxious, depressed, rule-breaking, and attention problems, as well as aggressive behaviors.
The investigators used multilevel mixed-effect models to assess associations between both air pollutants simultaneously and white matter connectivity and emotional behavior. Models were adjusted for numerous individual- and neighborhood-level factors. For the white matter tract metrics, associations were assessed by the brain hemisphere, evaluated using nonlinear concentration–response functions, and adjusted for multiple comparisons.
KEY RESULTS
Mean 1-year PM2.5 exposure was 7.7 μg/m3 and ranged from 1.7 to 15.9 μg/m3. Mean 1-year NO2 exposure was 18.6 ppb and ranged from 0.7 to 37.9 ppb. These exposures were below the level of the US National Ambient Air Quality Standards for long-term exposure.
Herting and colleagues found that associations between air pollution and white matter connectivity in the prefrontal cortex were pollutant-specific. Higher NO2 exposure was linearly associated with lower levels of fractional anisotropy in three white matter tracts (Statement Figure). These findings were consistent with the study hypothesis and suggest that NO2 exposure might lead to less organized nerve fibers. In contrast, PM2.5 exposure was associated with less mean diffusivity in six of the seven evaluated white matter tracts, particularly in the left hemisphere, indicating better connectivity. Those findings were not consistent with the study hypothesis. Many of the detected associations for both pollutants were specific to white matter tracts on either the right or left sides of the brain. The associations were also detected in the corpus callosum, which spans both sides of the brain.
Annual NO2 exposure is associated with lower white matter connectivity in the prefrontal cortex of children ages 9–10. Higher NO2 exposure was associated with lower fractional anisotropy in three of seven areas: the uncinate fasciculus (UNC), anterior thalamic radiation (ATR), and corpus callosum.
PM2.5 and NO2 exposures were not associated with emotional behavior outcomes when children were enrolled at ages 9–10. However, over the 1 year of follow-up, PM2.5 was associated with decreased anxiety-depression behaviors, withdrawn-depression behaviors, and aggression, and NO2 was associated with a decrease in total behavioral problems. These results were not consistent with the study hypothesis. Herting and colleagues noted that the magnitude of the associations was very small and might not be indicative of causal effects or correspond to clinically meaningful outcomes. The results for both white matter connectivity and emotional behavior were similar when the models were adjusted for prenatal air pollution exposure.
INTERPRETATION AND CONCLUSIONS
In its independent review of the study, the HEI Review Committee concluded that this report presents a thorough investigation into associations between exposure to air pollution and certain measures of brain development. They thought that it was novel to study the preadolescent period, which has received less focus than the prenatal and early life periods in research on environmental exposures and neurodevelopment. The Committee appreciated the high-quality design and methods, including the use of a nationally representative cohort of children, thorough statistical analyses, and the use of complementary neurodevelopmental outcomes — both brain scans and caregiver-reported mental symptoms. The Committee also appreciated the inclusion of both prenatal and concurrent air pollution exposures, whereas most prior related research assessed air pollution only once or only over shorter periods of development.
A limitation of the study was that the exposure assessment was based on residential address only, even though children in the United States spend an average of 1,000 hours at school each year. The Committee recommended that future studies in this cohort might also assess ambient air pollution exposure at school and incorporate this information into a time–activity-based exposure assessment.
Herting and colleagues detected associations between air pollution and white matter microstructure that varied by the specific pollutant, brain region, and measure of white matter connectivity. The mixed associations are consistent with prior studies, suggesting that air pollution can alter white matter microstructure. However, the specific effects and long-term implications are unclear, and such changes are not clearly linked to disease states or clinical outcomes. Indeed, the observed changes in brain microstructure did not correlate with emotional behavior problems in this study. The Committee also noted that the changes in white matter structure were subtle and that detection of changes in emotional behavior might require follow-up beyond 1 year.
In contrast to their hypothesis, Herting and colleagues observed a small but apparent protective association between air pollution and emotional behavior. Although other studies have observed sporadic protective associations, the Committee noted that it is biologically implausible that air pollution exposure would elicit a protective effect and that the observed associations might be due to chance or residual bias. An additional limitation was that caregivers only evaluated the emotional-behavioral assessment. Research suggests that preadolescent children are capable of reliably self-reporting emotional behavior; therefore, the emotional behaviors in the current study might have been underestimated. Future research could benefit from further integrating additional brain imaging techniques, self-reported symptoms, and neuropsychological assessments.
In summary, Herting and colleagues examined whether concurrent and prenatal exposures to outdoor PM2.5 and NO2 were associated with changes in white matter connectivity in the prefrontal cortex and with measures of emotional behavior over a 1-year period in children transitioning to adolescence. They observed alterations in white matter connectivity that varied by pollutant and region of the prefrontal cortex but did not find that emotional behavior was negatively affected. This study adds to the existing body of literature demonstrating that air pollution can alter neurodevelopment, even at levels below current regulatory standards for annual exposure. Future research is needed to integrate complementary outcomes to help link objective brain biomarkers with clinical disorders.
Res Rep Health Eff Inst. 2025 Mar 1;2025:225.
Air Pollution Exposure, Prefrontal Connectivity, and Emotional Behavior in Early Adolescence
1
University of Southern California, Los Angeles, California, USA
2
Harvard University, Cambridge, Massachusetts, USA
3
Columbia University, New York City, New York, USA
This Investigators’ Report is one part of Health Effects Institute Research Report 225, which also includes a Commentary by the Review Committee and an HEI Statement about the research project. Correspondence concerning the Investigators’ Report may be addressed to Dr. Megan M. Herting, University of Southern California, 1845 N. Soto St., Los Angeles, CA 90032; email: herting@usc.edu. No potential conflict of interest was reported by the authors.
Although this report was produced with partial funding by the United States Environmental Protection Agency under Assistance Award CR–83998101 to the Health Effects Institute, it has not been subjected to the Agency’s peer and administrative review and may not necessarily reflect the views of the Agency; no official endorsement by it should be inferred. The contents of this report also have not been reviewed by private party institutions, including those that support the Health Effects Institute; therefore, it may not reflect the views or policies of these parties; no endorsement by them should be inferred.
Emerging evidence suggests that ambient air pollution may affect the developing brain and contribute to an increased risk of mental health problems. However, most studies have focused on prenatal or early postnatal periods of exposure, with less attention given to the dynamic neurodevelopment period of early adolescence. Moving forward, it is necessary to consider additional periods of exposure, such as adolescence, and the biological mechanisms that may drive potential neurotoxicological effects. This project aimed to investigate whether 1-year exposure to ambient fine particulate matter (PM2.5*) and nitrogen dioxide (NO2) at 9–10 years of age was associated with (1) concurrent prefrontal white matter connectivity at ages 9–10 years and (2) emotional health problems at ages 9–10 years as well as 1 year later. Lastly, we hypothesized that poor prefrontal white matter connectivity might be an intermediate marker (i.e., mediator) for the association between 1-year ambient exposure and mental health outcomes.
Methods
We leveraged data from the multisite, nationwide Adolescent Brain Cognitive Development Study (ABCD Study; N = 11,880), with cross-sectional data on diffusion-weighted imaging at 9–10 years (baseline visit) and longitudinal emotional health outcomes at 9–10 (baseline visit) and 10–11 years (1-year follow-up). Based on residential addresses at ages 9–10 years, novel hybrid spatiotemporal exposure models were applied to estimate 1-year average ambient exposure to PM2.5 and NO2. Diffusion tensor imaging (DTI) was used to measure white matter microstructure in tracts that innervate the prefrontal cortex. Emotional behavioral problems were measured based on caregiver reports using the Child Behavioral Checklist (CBCL). Mixed-effect two-pollutant models were fit using both PM2.5 and NO2 and adjusted for the study site, several potential sociodemographic and lifestyle characteristics, and magnetic resonance imaging (MRI) precision variables when necessary. For emotional health outcomes, longitudinal models included interaction terms for pollutant-by-time for both pollutants. Sensitivity analyses were conducted that also accounted for the number of years the child resided at the residential address, as well as adjusting for prenatal PM2.5 and NO2 exposures.
Results
The final analytic sample included 7,546 participants with DTI data and 9,334 participants with emotional behavior data. The annual exposures to PM2.5 and NO2 across 21 study sites were 7.66 μg/m3 [1.72–15.90 μg/m3] and 18.61 ppb [0.73–37.94 ppb], respectively. Annual exposure to PM2.5 was found to be significantly related to prefrontal structural connectivity, including fractional anisotropy (FA) in the right superior longitudinal fasciculus and widespread differences in mean diffusivity (MD) in the corpus callosum, bilateral uncinate fasciculus, left cingulum-hippocampal region, left anterior thalamic radiation, and left superior longitudinal fasciculus. The observed associations between PM2.5 and MD were negative and nonlinear, with greater decreases in MD seen at higher exposure levels. Annual exposure to NO2 was found to have significant, negative linear associations with FA in the right anterior thalamic radiation, left uncinate fasciculus, and corpus callosum. In terms of emotional behavior, 1-year PM2.5 annual exposure was related to slightly less internalizing, anxiety/depression, and aggression problems at the 1-year follow-up. Similarly, 1-year NO2 annual exposure was related to slightly less internalizing and total problems at the 1-year follow-up. Although some of these associations were statistically significant, small parameter estimates suggest these noted effects on emotional outcomes may not be of clinical importance. Given the later findings, the required conditions to test mediation formally were not met.
Conclusions
Our analyses indicate that white matter microstructure is uniquely associated with annual exposure to PM2.5 and NO2 at ages 9–10 years. Against our hypotheses, annual exposure was not related to more emotional problems at ages 9–10 years or after a 1-year follow-up period. These findings suggest air pollution exposure levels below US national ambient air quality standards may have important implications for child white matter development and add to the literature suggesting neurotoxicity at low exposure levels of air pollution may be critical to include in the continuing review and risk assessment for the National Ambient Air Quality Standard.
INTRODUCTION
Fine particulate matter (aerodynamic diameter <2.5 μm; PM2.5) is ubiquitous in urban areas and is recognized as an important contributor to the global burden of disease (Cohen et al. 2017). Emerging evidence indicates PM2.5 may also be a neurotoxicant (Block et al. 2012, Cory-Slechta et al. 2023, Cory-Slechta and Sobolewski 2023), resulting in cognitive, psychomotor, and behavioral problems (Abid et al. 2014, Suades-González et al. 2015, Brockmeyer and D’Angiulli 2016, Clifford et al. 2016, Stingone et al. 2017, de Prado Bert et al. 2018, Bansal et al. 2021, Wang et al. 2021, Ahmed et al. 2022, Castagna et al. 2022, Yi et al. 2022). Children may be the most vulnerable to these effects because brain maturation continues well into the third decade of life (Block et al. 2012, de Prado Bert et al. 2018, Herting et al. 2019). Exposures to ambient PM2.5 and its components in early life have been inversely associated with poor neurodevelopmental outcomes (Abid et al. 2014, Perera et al. 2014a, Suades-González et al. 2015, Clifford et al. 2016, de Prado Bert et al. 2018, Loftus et al. 2020, Volk et al. 2021, Ahmed et al. 2022, Castagna et al. 2022, Yi et al. 2022, Ramaiah et al. 2023), as well as several emotion behavioral problems and mental health diagnoses (Forns et al. 2016, Harris et al. 2016, Fan et al. 2019, Zundel et al. 2022), including delinquent behavior (Younan et al. 2018), anxiety and mood disorders (Perera et al. 2012, Khan et al. 2019, Roberts et al. 2019, Yolton et al. 2019, Latham et al. 2021, Rasnick et al. 2021, Reuben et al. 2021, Yang et al. 2023), and attention-deficit hyperactivity disorder (ADHD) (Newman et al. 2013, Perera et al. 2014b, Peterson, Rauh, et al. 2015, Shih et al. 2022, Yuchi et al. 2022). Yet, these findings are incongruent with other studies, which have failed to find an association (Forns et al. 2018, Jorcano et al. 2019, Dores et al. 2021, Kim et al. 2021, Kusters et al. 2022). To further characterize the link between air pollution and mental health, it is necessary to consider the timing of both the exposure and the outcome, as well as the biological mechanisms that may underly these potential neurotoxicological effects (Myhre et al. 2018, Margolis et al. 2022, Zundel et al. 2022). For example, the existing literature has largely focused on the prenatal period and rarely assessed associations between air pollution and emotional behaviors during adolescence (Herting et al. 2019, Zundel et al. 2022, Cory-Slechta et al. 2023), defined as ages 10–19 years (WHO 2023), when mental health problems typically begin to emerge (Kessler et al. 2007, WHO 2013). Moreover, recent longitudinal studies suggest that childhood and early adolescent exposure to PM2.5 and NOx (nitrogen oxides) is related to an increased risk for greater mental health disorders, including major depression at age 18 years (Newbury et al. 2019, Roberts et al. 2019, Reuben et al. 2021). These findings suggest that it is plausible that exposure during the transition from childhood to adolescence may be associated with neurodevelopmental differences that may contribute to mental health problems later in life. Thus, studying both the adolescent period of neurodevelopment as well as finding identifiable intermediate outcomes of the potential neurotoxicological effects of ambient air pollution exposure may help clarify and strengthen the link between ambient air pollution and risk for mental health disorders.
The adolescent period is hallmarked by dynamic behavioral and neurodevelopmental changes — ultimately making adolescence a sensitive period in which the brain and emotional processing may be especially susceptible to environmental factors (Christie and Viner 2005, Paus et al. 2008, Casey 2015). As children transition into the teenage years, age-related improvements are seen in emotional regulation, or the ability to enact strategies to influence the experience and expression of emotion (Yap et al. 2007, Hare et al. 2008, Tottenham et al. 2011, Somerville et al. 2013, Heller and Casey 2016, Lantrip et al. 2016). Alongside age-related improvements in emotional regulation, the brain continues to mature well into the third decade of life (Giedd et al. 1996a, 1996b, Mills et al. 2016, Tamnes et al. 2017, Herting et al. 2018, Lebel et al. 2019, Bethlehem et al. 2022). Dendritic arborization, synaptic pruning, and myelination occur in a posterior to anterior fashion (Goldman-Rakic 1987, Benes 1989, Huttenlocher 1990, LaMantia and Rakic 1990, Gogtay et al. 2006), with brain regions important to coordinating one’s emotional reactions such as the prefrontal cortex and limbic structures (i.e., hippocampus, amygdala, anterior cingulate [Hariri et al. 2000, Cardinal et al. 2002, Ochsner et al. 2004, Likhtik et al. 2005, Phan et al. 2005, Quirk and Beer 2006, Ghashghaei et al. 2007, Wright et al. 2008]) rapidly developing during adolescence (Giedd et al. 1996a, 1996b, Hariri et al. 2000, Yurgelun-Todd and Killgore 2006, Qin et al. 2012, Gee et al. 2013, Gabard-Durnam et al. 2014, Tamnes et al. 2017, Herting et al. 2018). A combination of pruning and continued myelination of connections allows for greater specialization and efficiency of neural systems beginning in the peri-adolescent period and continuing into young adulthood (Fair et al. 2007, 2008, 2009, Dosenbach et al. 2010, Giedd and Rapoport 2010, Casey 2015). Importantly, disruptions to these maturational processes during this time can have potentially lifelong consequences (Paus et al. 2008, Fuhrmann et al. 2015). For example, brain and behavior phenotypes during the transition period of adolescence are strong predictors for subsequent mental health problems (Paus et al. 2008, Pessoa 2009, Carlson et al. 2012, Beauchaine 2015, Snyder et al. 2015, Huang-Pollock et al. 2017), including internalizing (e.g., anxiety and depression) (Yap et al. 2007, Hare et al. 2008, Tottenham et al. 2011, Somerville et al. 2013, Baune et al. 2014, Heller and Casey 2016, Lantrip et al. 2016) and externalizing problems (e.g., delinquency, aggression, and substance use) (Bø et al. 2017, Wakschlag et al. 2018). Moreover, half of all lifetime prevalence of mental illness typically begins by age 14 (Kessler et al. 2005, WHO 2013), with the poorest prognosis seen for individuals who display symptoms early on (Kessler et al. 2007, WHO 2013). Thus, the growing adolescent brain may be vulnerable to the potential neurotoxic effects of air pollution, with potential lasting impacts on emotional health.
Beyond the important consideration of examining exposure during adolescence, the existing literature suggests the importance of identifying neural biomarkers of PM2.5 exposure–related risk before clinical onset of emotional disorders. In this regard, an increasing number of animal exposure studies and in vivo human MRI studies suggest that, in addition to the prenatal period (Boda et al. 2020), postnatal PM2.5 exposure impacts both brain structure and function. For example, long-term exposure to airborne particles, including PM2.5 and nano-sized PM during early life (including childhood and adolescence), leads to changes to neuronal and glial cell function (Solaimani et al. 2017, Gomez-Budia, et al. 2020, Morris et al. 2021), including decreases in dendritic spine density and branching in the hippocampus (Fonken et al. 2011), impaired neurogenesis (Woodward et al. 2018), as well as more anxiety (Nephew et al. 2020) and depressive-like symptoms in adult rodents (Fonken et al. 2011, Woodward et al. 2018). Exposure to nano-sized particles has also been found to increase microglial cell count and lead to changes in morphology in white matter in adult mice (Babadjouni et al. 2018). Similarly, in vivo brain imaging studies also suggest that postnatal exposure to PM2.5 and its components impact brain structure and function in childhood and adolescence, including differences in gray matter macro- and cytoarchitecture (Pujol et al. 2016a, Guxens et al. 2018, Mortamais et al. 2019, Cserbik et al. 2020, Lubczyńska et al. 2021, Guxens et al. 2022, Sukumaran et al. 2023), white matter microstructure (Lubczyńska et al. 2020, Burnor et al. 2021, Binter et al. 2022, Guxens et al. 2022, Peterson et al. 2022), as well as functional connectivity patterns (Pujol et al. 2016b, Guxens et al. 2022, Pérez-Crespo et al. 2022, Cotter et al. 2023), including the prefrontal cortex as well as frontal and limbic white matter tracts in both children and adolescents. Interestingly, these neural targets of ambient air pollution converge with neural phenotypes involved in both emotional regulation and linked to psychiatric disorders. Specifically, structural white matter connections between the prefrontal cortex and limbic regions allow for improved emotional regulation across adolescence (Giedd et al. 1996a, 1996b, Hariri et al. 2000, Yurgelun-Todd and Killgore 2006, Qin et al. 2012, Gee et al. 2013, Gabard-Durnam et al. 2014, Tamnes et al. 2017, Herting et al. 2018). Moreover, alterations in prefrontal white matter structural connectivity have been linked with various mental health disorders suggesting it may be a potential neurobiological marker of mental health risk (Jenkins et al. 2016, Hinton et al. 2019, Vanes and Dolan 2021). For example, altered prefrontal white matter connectivity is seen in youth diagnosed with internalizing disorders (i.e., anxiety and depression [Cullen et al. 2010, LeWinn et al. 2014, Liao et al. 2014]), externalizing disorders (i.e., ADHD and conduct disorder [Nagel et al. 2011, Puzzo et al. 2018]), and risk for psychosis (Roalf et al. 2020). Thus, prefrontal white matter structural connectivity may be an especially useful biomarker of PM2.5 neurotoxicity as well as possibly an intermediate marker (i.e., mediator) of air pollution–induced mental health effects.
Despite a growing body of literature supporting the notion that ambient air pollution may increase the risk of mental health problems and impact brain development, there remain substantial knowledge gaps. Existing questions remain as to the potential neurotoxic effects of air pollution exposure during the sensitive window of adolescent brain development (Clifford et al. 2016, Herting et al. 2019, Zundel et al. 2022, Cory-Slechta et al. 2023), which is also an important time when many mental health symptoms begin to emerge (Kessler et al. 2005, 2007, WHO 2013). Moreover, few studies have been population-based and the majority have been limited by small sample sizes, cross-sectional outcomes, and/or narrow geographical representation — thus limiting the potential generalizability of findings (Paus et al. 2008, Fordyce et al. 2018, Herting et al. 2019, Fan et al. 2020). Additionally, many studies have not addressed potential confounding factors, such as socioeconomic factors or noise, which may contribute to inconsistencies in the literature (de Prado Bert et al. 2018, Herting et al. 2019, Fan et al. 2020, Volk et al. 2021). Lastly, although the levels of ambient air pollution have declined significantly over the last decades in the United States, an increasing number of epidemiological studies have reported associations with adverse health effects at even lower levels of exposure (Dominici et al. 2019). Thus, more air pollution research is needed to better understand the potential health effects of low exposure concentrations experienced by today’s youth (Landrigan et al. 2018). In this project, we aimed to address these knowledge gaps to investigate the associations between ambient air pollution, prefrontal white matter connectivity, and emotional behavior during adolescence.
SPECIFIC AIMS
The overarching aim of this project was to expand the current knowledge as to the effects of early adolescent exposure to ambient PM2.5 on prefrontal cortex (PFC) white matter neurodevelopment and mental health problems. Based on the extant literature, our primary exposure of interest is PM2.5, however, our analyses also examined ambient NO2 to ensure the robustness of our results. To accomplish this, we investigated the following specific aims:
Aim 1: To determine the association of 1-year average air pollution exposure at ages 9–10 and effects on concurrent PFC white matter structural connectivity at 9–10 years old.
We hypothesized that exposure to higher levels of PM2.5 exposure during adolescence is associated with impaired white matter connectivity between the PFC and regions involved in emotional processing, including limbic regions of the amygdala, hippocampus, and anterior cingulate gyrus. We also expected these effects to persist even after adjusting for NO2 co-exposure during adolescence and any potential prenatal exposure effects.
Aim 2: To assess the relationship between 1-year average air pollution exposure at ages 9–10 on emotional problems during adolescence, specifically at both ages 9–10 years old as well as 1 year later at ages 10–11 years.
We hypothesized that higher levels of PM2.5 exposure during adolescence are associated with more emotional problems at ages 9–10 as well as at the 1-year follow-up. Again, we expected these effects to persist even after adjusting for NO2 co-exposure during adolescence and any potential prenatal exposure effects. We also hypothesized that poorer white matter prefrontal connectivity might be an intermediate marker (i.e., mediator) of air pollution–induced mental health effects.
The current aims were modified slightly from the initial specific aims. Within Aim 2 of the study, the initial goal was to examine both continuous measures of emotional behavior as well as diagnostic criteria for mental health disorders using a computerized version of the Kiddie Schedule for Affective Disorders and Schizophrenia. However, errors were found in the processing of the final assignment of diagnostic criteria by the ABCD Study. These errors were beyond the control of the investigators to correct. Thus, the final specific aims of the current study focus on dimensional emotional behavior symptomology. The initial study aims also included exploring mediation of prefrontal white matter connectivity in the association between exposure and emotional outcomes. However, given for Aim 2 that the associations between PM2.5 and NO2 on emotional outcomes were very small to negligible effect sizes and thus may not be clinically meaningful, it was jointly decided by the Research Committee and the investigators that the required conditions to test mediation formally were not met.
STUDY DESIGN AND METHODS
The current study leverages (1) novel spatiotemporal estimates of annual average PM2.5 and NO2 at the child’s home address; (2) one measurement of diffusion tensor imaging of white matter microstructure of the child; (3) two measurements of the child’s emotional behavior; and (4) additional demographic, lifestyle, and spatial environmental confounders, from the ongoing ABCD Study. An overview of the current study can be seen in Appendix A, Figure 1 (Appendix A is available on the HEI website). The main analyses for the current study focused on data collected in person from children and their caregivers at the first two annual ABCD study visits that occurred when the children were 9–10 years and 10–11 years of age. We conducted both single- and two-pollutant analyses. Sensitivity analyses were then conducted to account for (1) an alternative set of confounders; (2) the duration the child lived at the primary address provided; and (3) prenatal exposure based on address that overlapped with the child’s birth date, based on retrospective recall from the child’s caregiver. Details as to the location of these corresponding details for these analyses are provided in the Research Roadmap.
RESEARCH ROADMAP.
Aims and Research Conducted
Methods Description
Aim 1: Annual exposure at ages 9–10 years and prefrontal white matter connectivity
Main report and Appendix
Two-pollutant models
Main report
Single-pollutant models
Appendix
Sensitivity analyses of two-pollutant models
Alternative minimally sufficient adjustment set
Appendix
Adjustment without geospatial confounders
Appendix
Exclusion of participants in rural areas
Appendix
Residential duration
Main report and Appendix
Prenatal exposure
Main report and Appendix
Aim 2: Annual exposure at ages 9–10 years and emotional outcomes at ages 9–10 and 10–11 years
Centralized institutional review board (IRB) approval was obtained from the University of California, San Diego IRB. Study sites obtained approval from their local IRBs. Written informed consent was provided by each parent; each child provided written assent. All ethical regulations were complied with during data collection and analysis. Participants were compensated for their participation in the current study.
STUDY DESIGN AND POPULATION CHARACTERISTICS
We used data from the geographically and sociodemographically diverse ABCD Study cohort (Auchter et al. 2018, Volkow et al. 2018). The ABCD Study is the largest long-term study of adolescent brain development from 21 communities throughout the United States (Figure 1), with 11,880 children, ages 9–10 years, enrolled between years 2016 and 2018 and followed annually for up to 10 years (Volkow et al. 2018). The ABCD cohort study design, eligibility criteria, recruitment, and data collection procedures have been described elsewhere in detail (Barch et al. 2018, Casey et al. 2018, Feldstein Ewing et al. 2018, Garavan et al. 2018, Luciana et al. 2018, Uban et al. 2018). Primary inclusion criteria for participant enrollment into the ABCD Study were age (9.0 to 10.99 years) and fluency in English. Exclusionary criteria for participation were MRI contraindications; history of significant traumatic brain injury or major neurological disorder (i.e., seizure disorder, cerebral palsy, or other conditions requiring neurological or medical care); uncorrectable sensory and motor impairments that preclude the youth’s ability to participate in study procedures; current or persistent major Axis I psychiatric disorder (i.e., psychosis and bipolar disorder); current medication of antipsychotics or mood stabilizers; current or history of persistent severe learning disorder, mental retardation, or pervasive developmental disorder or substance-use disorder diagnosis; parent report of intellectual disability; prematurity, very low birth weight or perinatal complications; and knowledge at baseline of impending move to an area not within reasonable traveling distance to an ABCD study site.
Geographic distribution of Adolescent Brain Cognitive Development (ABCD) study sites.
Study site abbreviations: CHLA = Children’s Hospital of Los Angeles, CUB = University of Colorado Boulder, FIU = Florida International University, LIBR = Laureate Institute for Brain Research, MUSC = Medical University of South Carolina, OHSU = Oregon Health and Science University, ROC = University of Rochester, SRI = SRI International, UCLA = University of California, Los Angeles, UCSD = University of California San Diego, UFL = University of Florida, UMB = University of Maryland Baltimore, UMICH = University of Michigan, UMN = University of Minnesota, UPMC = University of Pittsburgh Medical Center, UTAH = University of Utah, UVM = University of Vermont, UWM = University of Wisconsin-Milwaukee, VCU = Virginia Commonwealth University, WUSTL = Washington University in St. Louis, YALE = Yale University.
STUDY SITE CATCHMENT AREAS
The recruitment approach for the ABCD Study was primarily through public, charter, and private schools (Garavan et al. 2018). Each ABCD study site created a catchment area, defined of all schools within 50 miles of the research institute, and used a stratified sampling of schools within each site’s catchment area. The subset of schools was then randomly selected from this list of potential schools. Beyond this primary recruitment strategy, an embedded twin design (n = 4 sites) used a birth record approach, and approximately 10% were recruited via other methods (i.e., community events, nontargeted schools) (Garavan et al. 2018, Karcher and Barch 2021). Thus, it is important to note that while participants completed their study visits at a given ABCD study site, the participants’ residential location, and thus exposure in the current study, at any given site may vary given this study design. Given that residential address (i.e., geocoded coordinates) data are considered protected health information, they cannot be shared outside of the ABCD consortium’s Data Analytics Information and Resource Center (DAIRC) (Fan et al. 2021). Therefore, study area maps for each site are currently not available for ABCD study participants.
ESTIMATION OF RESIDENTIAL PM2.5 AND NO2 EXPOSURE
Geocoded information about participants’ residential addresses was used to define the locations where annual average PM2.5 and NO2 exposures were estimated (Fan et al. 2021). Primary residential addresses when the child participants were 9–10 years of age were collected in person from the participant’s caregiver during the study visit between October 2016 to October 2018. We used birth addresses, as reported retrospectively by the participant’s caregiver, for the residential location for estimating exposure during the prenatal period. All residential addresses were then geocoded by the ABCD consortium’s DAIRC (Fan et al. 2021). Importantly, only complete addresses with valid geocode matches were used to map exposure estimates. At each geocoded primary residential address provided for the child at ages 9–10 years (and for the 9-month pregnancy period based on birth address), PM2.5 and NO2 exposure was estimated using hybrid spatiotemporal modeling (Di et al. 2016, 2019, 2020).
We obtained daily concentrations of ambient PM2.5 and NO2 at 1-km2 spatial resolution across the United States from two ensemble prediction models that combine multiple machine-learning algorithms (Di et al. 2019, 2020). These hybrid models use a generalized additive model that accounts for geographic differences to combine exposure estimates from three machine-learning algorithms, including neural network, random forest, and gradient boosting. The machine-learning algorithms use multiple predictor variables, including satellite-based aerosol optical depth models, land-use variables (i.e., land use coverage types, road density, restaurant density, elevation, and normalized difference vegetation index), chemical transport models, as well as meteorological and other ancillary variable inputs (i.e., air temperature, accumulated total precipitation, downward shortwave radiation flux, accumulated total evaporation, planetary boundary layer height, low cloud area fraction, precipitable water for the entire atmosphere, pressure, specific humidity at 2 meters, visibility, wind speed, medium cloud area fraction, high cloud area fraction, and albedo, etc.). For missing data, separate prediction models for each predictor variable with missing values were created and used the predicted values to fill in the missing records. These models are then trained for the continental United States from monitoring data collected from 2000 to 2016 and validated using a tenfold cross-validation. It is important to note, however, that monitoring sites for these data are not equally distributed across the United States, with more sites in the Eastern United States, West Coast, and urban areas, and fewer sites in mountainous regions and rural areas (Appendix Figure 2). For PM2.5, monitoring data were obtained from the Environmental Protection Agency (EPA), the Interagency Monitoring of Protected Visual Environments (IMPROVE), Clean Air Status and Trends Network (CASTNET), and other regional or local monitoring data with 2156 monitoring sites. For PM2.5, the tenfold cross-validation R2 was 0.86 (spatial R2 was 0.89 and temporal R2 was 0.894) for annual averages, and the average root mean square error was 2.786 μg/m3 (1.26 μg/m3 spatially and 2.53 μg/m3 temporally) (Di et al. 2019). For NO2, monitoring data were obtained from the EPA, including 912 monitoring sites. For NO2, the tenfold cross-validation R2 was 0.788 for annual averages (spatial R2 was 0.844 and temporal R2 was 0.729), and the average root mean square error was 7.146 ppb (4.51 ppb spatially and 25.57 ppb temporally) (Di et al. 2020).
Using the primary residential address provided at the baseline study visit, annual PM2.5 and NO2 exposures were calculated by averaging the daily estimates from these ensemble models over the 2016 calendar year (Figure 2). For the prenatal period estimates, exposure was assigned by averaging across the daily exposure estimates from these ensemble models for 9 months of pregnancy based on the child’s birthdate (birth years 2005–2009) and weighting the pollution exposure accordingly for participants who reported multiple addresses that overlapped with the birthdate for the child (Appendix A, see section Prenatal Residential Addresses and Exposure Estimates). Quality-controlled prospective residential addresses are not currently available from the ABCD consortium. Therefore, for the 1-year follow-up, we assumed the spatial contrast remained similar over this 1-year follow-up period, which has been demonstrated using this ensemble-based model from 2000 to 2016 (Di et al. 2019, 2020).
Annual averages of daily PM2.5 and NO2 across the continental United States for the 2016 calendar year. Overlay of 21 ABCD study site locations included for visualization purposes. See Figure 1 for study site abbreviations.
STUDY OUTCOME DATA
The current study included white matter microstructure imaging collected from brain MRI from the children when they were 9–10 years of age (for Aim 1) and emotional behavior assessments collected by questionnaire from their caregiver when the child was 9–10 years old and at a 1-year follow-up timepoint when the child was 10–11 years old (for Aim 2). As is important for multisite studies, all ABCD methods and assessments have been optimized and harmonized across the 21 sites.
Brain MRI White Matter Imaging Outcome Variables
Harmonized MRI imaging methods and image acquisition were utilized across all 21 sites using three Tesla scanners (Siemens Prisma, General Electric 750, and Philips) (Casey et al. 2018). Diffusion MRI (dMRI) sequence details per scanner are provided in the Appendix Table 1. The dMRI acquisition was multiband echo planar imaging with a slice acceleration of 3, with 96 diffusion directions, 7 b = 0 images, and four different b values (6 directions at b = 500, 15 directions at b = 1,000, 15 directions at b = 2,000, and 60 directions at b = 3,000), with a spatial resolution of 1.7 mm isotropic (Hagler et al. 2019). All collected MRI images then underwent a harmonized quality control (QC) protocol as well as image preprocessing at the DAIRC as previously detailed (Hagler et al. 2019) (see also Appendix Figure 3). QC steps prior to image preprocessing include (1) protocol compliance (checking parameters and completeness), (2) automated QC calculations of signal-to-noise and head motion, and (3) visual inspections by trained technicians checking for image quality and imaging artifacts. Postprocessing QC steps of the diffusion-weighted images include the manual assessment by trained technicians of the presence and severity of artifacts or processing problems including B0 warping, motion-related and magnetic susceptibility artifacts, full head coverage, registration with T1w image, and the accuracy of fiber tract segmentation. Numeric values are assigned on a scale of 0–3, indicating absent, mild, moderate, and severe levels of each type of artifact, respectively. The reviewers assign an overall QC score indicating the image can be recommended for use (1) or for exclusion (0). Exclusion is recommended if any of the five categories are rated as severe (a value of 3). Image preprocessing steps for the current study include: (1) eddy current and motion correction, (2) B0-distortion correction, (3) field inhomogeneity correction, (4) resampling to 1.7-mm3 resolution, and (5) alignment to atlas space (Hagler et al. 2019). Following general preprocessing, diffusion tensor imaging metrics are calculated to estimate white matter tracts in each participant using a probabilistic atlas-based method, known as AtlasTrack (Hagler et al. 2009). Fractional anisotropy (FA) and mean diffusivity (MD) — metrics that may reflect impairments in myelination, poor axon organization, or differences in axon caliber — were then quantified for each tract of interest (Le Bihan 2003). Specifically, FA reflects the degree of anisotropy of water diffusion (e.g., restricted/directional diffusion of water) and is unitless with a value range of 0 (i.e., isotropic) to 1 (i.e., anisotropic). MD reflects the overall diffusion of water with a unit of mm2/sec. In the current study, we performed an a priori region-of-interest analysis focused on a whole-brain summary measure across all white matter tracts (i.e., global estimates across the whole brain, referred to as “all fibers”) as well as seven white matter tracts of interest. These seven white matter tracts included the anterior thalamic radiation (ATR), the cingulum (which is further delineated into the cingulate body [CGC] and hippocampal portions [CGH]), the corpus callosum (CC), the inferior fronto-occipital fasciculus (IFO), the superior longitudinal fasciculus (SLF), and the uncinate fasciculus (UNC). These were chosen because they carry connections between the PFC to limbic brain regions that are vital for emotional processing and regulation (Catani and Thiebaut de Schotten 2008, Coenen et al. 2012, Gabard-Durnam et al. 2014, Kamali et al. 2014, Philippi et al. 2009, Von Der Heide et al. 2013) and have also been identified by recent meta-analysis as potential transdiagnostic biological markers of mental health risk (Hinton et al. 2019, Jenkins et al. 2016, Vanes and Dolan 2021) (additional details in Appendix Table 2).
Emotional Behavior
Emotional behavioral questionnaire data were collected from the child’s caregiver at both the baseline visit (ages 9–10 years) as well as 1-year later (ages 10–11 years) using the Child Behavior Checklist (CBCL) (Achenbach and Rescorla 2001). The CBCL asks 113 items about the child’s behavior over the 6 months prior to each study visit, with each item endorsed with a three-point Likert-type scale of 0 = “not true,” 1 = “somewhat or sometimes true,” or 2 = “very true or often true.” Scores were then calculated for total problems, internalizing and externalizing problems, as well as various subscale scores, including those of interest for the current study: anxious/depressed, withdrawn/depressed, rule-breaking behaviors, aggressive behaviors, and attention problems (additional details in Appendix Figure 4). For all outcomes of interest, higher scores reflect more behavioral problems. This measure shows good test-retest reliability (Pearson’s r = 0.88; min = 0.8, max = 0.94) and internal consistency (Cronbach’s alpha = 0.8; min = 0.63, max = 0.97) (Achenbach and Rescorla 2001). To account for age and sex differences and to improve model diagnostics, t scores (mean of 50, standard deviation of 10) on the CBCL are used as outcomes in our analyses.
POTENTIAL CONFOUNDING AND PRECISION VARIABLES
For each aim, we used a directed acyclic graph (DAG) to a priori define potential confounding variables (i.e., those variables known to both predict the outcomes of interest here and also likely to influence where people live, thus their exposure to ambient air pollutants estimated at the residence) (Hernan et al. 2002). Questionnaire data on caregivers’ sociodemographics were completed at the baseline visit using REDCap (Research Electronic Data Capture), which is a secure, web-based software platform (Harris et al. 2009, 2019). These sociodemographic variables included parental education (<high school diploma [<HS], high school diploma/general education degree [HS/GED], some college, bachelor’s degree, postgraduate degree), total household income (<$50,000 [<$50k], ≥$50,000 and <$100,000 [≥$50k & <$100k], ≥$100,000 [≥$100k)], or don’t know/refuse), parental employment status (working, stay-at-home parent, unemployed, other), maternal age at birth (years), tobacco exposure during pregnancy (yes/no), and second-hand smoke exposure at ages 9–10 years (yes/no). Questionnaire data based on caregiver report at the time of the baseline visit were also obtained about the child, including age at the time of visit (in months), sex at birth, race/ethnicity (Non-Hispanic Black, Non-Hispanic White, Hispanic, Other including American Indian/Native American, Alaska Native, Native Hawaiian, Guamanian, Samoan, Other Pacific Islander, Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, Other Asian, or Other Race), body mass index (calculated from obtained measures of weight [kg] and height [cm]), pubertal development (pre-, early, mid-, and late- and post-) (Petersen et al. 1988), asthma diagnosis (yes/no), average days per week of physical activity (Barch et al. 2018), and average hours of screen time use per week (Paulus et al. 2019, Paulich et al. 2021). Beyond sociodemographic and/or lifestyle factors, we also considered several available additional environmental exposure estimates as potential confounding variables. These included perceived neighborhood quality, as assessed by caregivers derived from three questions modified from the PhenX Neighborhood Safety protocol (scale of 1–5, with higher values representing greater safety) (Echeverria et al. 2004, Mujahid et al. 2007), as well as geocode-derived characterization of the primary residential address at baseline, including categorization of urbanization (based on 2010 Census definitions of urban, urban area, urban-rural) (US Census Bureau 2010), population density (persons per 1-km2) (Center for International Earth Science Information Network [CIESIN] Columbia University 2016, Fan et al. 2021), proximity to major roads (in meters) (Fan et al. 2021), and nighttime sound (in decibels) (Mennitt et al. 2014). Three additional precision covariates were also identified for the first analysis on brain MRI, including the child’s handedness (Oldfield 1971), head motion during the MRI scan (measured by frame displacement in millimeters) (Hagler et al. 2019), and MRI scan manufacturer (Siemens, GE, or Philips), as these factors have been known to influence MRI brain outcomes, but are not likely to influence where an individual lives, and thus are not confounders. Using these variables and software designed to assist in the creation and interpretation of DAGs, known as DAGitty (Textor et al. 2016), we identified a minimally sufficient adjustment set of variables for our primary analyses. For Aim 1, this included the child’s age, sex, race and ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, and urban classification as well as the addition of MRI manufacturer, head motion, and child handedness as precision variables for MRI white matter outcomes in Aim 1. For Aim 2, CBCL t scores are normalized on age and sex, thus the minimally sufficient adjustment set included the child’s race and ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, and urban classification. We identified a second sufficient adjustment set for each aim, which we used as part of an additional sensitivity analysis to determine whether an alternative set of adjustment variables would affect our results. Given the potential theoretical overlap of some predictors in the exposure model and the identified geospatially derived covariates of the minimally sufficient adjustment set, such as the distance to roadways, population density, and urban classification, we conducted an additional sensitivity analysis using our primary minimally sufficient adjustment set excluding these three geospatially derived variables. Additional details can be found in the Appendix (see section, A Priori Directed Acyclic Graph [DAG] Approach to Identify Confounders) and Appendix Figures 5 and 6.
DATA ANALYSIS
NONRESPONSE ANALYSIS
Since missing data can introduce potential bias in both parameter estimation and reduce the generalizability of the results, we examined the proportion of missingness, as well as the missing data mechanisms and missing data patterns. The proportion of missingness of the exposure (i.e., no residential address provided) at ages 9–10 years old was approximately 5.5% of the ABCD study cohort, whereas the proportion of missingness of the outcome for Aim 1 was 6.2% and for Aim 2 was 2.9% (Appendix Tables 3 and 4). Several spatial confounders (i.e., distance to roadway, population density, and urban classification) were also based on residential address, and therefore, also showed ~6% missingness. All other confounders had ≤0.5% missingness. Slight sociodemographic differences were seen between children missing and not missing MRI and CBCL outcomes (Appendix Tables 5 and 6) as well as those missing and not missing exposure estimates (i.e., no residential information provided) (Appendix Tables 7 and 8), albeit the magnitude of these effects were exceptionally small. Given this very small fraction of missingness as well as the small effect sizes noted in sociodemographic factors as they pertained to item-response missingness, we chose not to impute missing data (Dong and Peng 2013). Moreover, with our very low percentage of missingness, we also chose to implement a multilevel model with maximum likelihood estimation that can handle missing data and can produce unbiased parameter estimates and standard errors under the missing-at-random assumption (Ibrahim and Molenberghs 2009).
SELECTION BIAS AND LOSS-TO-FOLLOW-UP ANALYSIS
Given that the emotional outcome analyses included two waves of data (at ages 9–10 years and at 1-year follow-up), selection bias can also occur in the current study if the likelihood of missing data at the 1-year follow-up is related to either the exposure or outcome. As such, we explored the association between sociodemographic and exposure data between children that did and did not have missing 1-year follow-up emotional outcome data (i.e., retention in CBCL outcomes) using logistic regression. We found no significant differences in the exposure, confounding variables, or outcome in those children missing emotional outcome data as compared to those with complete 1-year follow-up data (Appendix Table 9).
To further examine selection bias, we compared the children who were included in our final analyses versus those of the larger ABCD study cohort. Compared to the enrolled ABCD study population, children included in the current analyses on brain MRI outcomes were more likely to be Hispanic and White and have undergone brain scanning on a Siemens MRI machine as compared to the full ABCD study cohort (Appendix Table 10). Children included in the analyses on emotional behavior were exposed to slightly higher levels of PM2.5, were from more densely populated areas, and were more likely to come from households with incomes of less than $100,000 as compared to the full ABCD study cohort (Appendix Table 11).
EXPOSURE–OUTCOME ASSOCIATIONS
We performed both single-pollutant and two-pollutant models using a linear mixed model (LMM) approach with the similar structure for both PFC white matter and CBCL emotional outcomes.
Air Pollution Exposure and PFC White Matter Structural Connectivity at 9–10 Years
We performed multilevel LMM to study the association of annualized exposure with microstructure outcomes across all tracts (i.e., global measure) and each of the seven white matter tracts at ages 9–10 years while accounting for the nested structure of participants within 21 research sites. Nonlinearity in associations between air pollution exposures and white matter microstructure outcomes was assessed using a Type III analysis of variance (Satterthwaite method) comparing models with natural cubic spline terms for air pollution exposure estimates to models with linear terms for air pollution exposure estimates (P < 0.05). Initial examination of the data suggested the associations between the PM2.5 and white matter microstructure outcomes in most tracts may be nonlinear, whereas NO2 associations with the outcomes were linear. Moreover, examination of exposure and outcomes suggested PM2.5 and NO2 associations may also be different between hemispheres for bilateral tracts. Thus, our final multilevel mixed-effect models implemented natural cubic splines to capture potential nonlinearity in associations between PM2.5 and the outcome as well as interaction terms between each air pollution exposure and brain hemisphere term, with nested random intercepts at the subject level and the study site level and fixed effects for all potential confounding variables identified as part of the minimally sufficient adjustment set described above. Natural cubic splines for PM2.5 were fitted with two knots at 7.05 and 8.31 μg/m3, derived from tertiles of exposure. Given the many tests performed, we corrected for multiple comparisons using false discovery rate (FDR) procedures. Next, we performed LMMs stratified by brain hemisphere to further describe the associations between exposure and white matter microstructure outcomes in each tract and in each hemisphere.
Air Pollution Exposure and Emotional Behaviors at 9–10 Years and at 1-Year Follow-Up
CBCL t scores (i.e., normalized, with a mean = 50 and SD = 10 by age and sex) were found to be nonnormal. Given the ABCD Study aimed to recruit a neurotypical sample, most individuals reported few, if any, problems at either time point, with <12.4% of participants having a t score ≥63 and less than 3.2% of participants having a subscale t score ≥70, which would be deemed of potential clinical relevance (Appendix Figure 7). Given that most of the sample fell within the normal range and CBCL outcomes are inherently continuous, we chose to investigate exposure effects on continuous t-score estimates of emotional problems to minimize loss of information as well as ensure we could capture air pollution effects on behavioral problems across the entire spectrum (i.e., subclinical and clinical) of emotional health. Despite nonnormal CBCL outcome distributions, initial examination of the data found linear assumptions to have otherwise been met between the exposure and the outcome. Thus, to investigate the relationship between the annualized exposure and emotional behaviors (i.e., CBCL t scores) at ages 9–10 years as well as at a 1-year follow-up visit, we utilized a generalized LMM model with gamma distribution assumption and an identity link that again accounted for the nested structure of the multiple measurements of participants within the 21 research sites. Specifically, separate models were conducted for each outcome of interest (i.e., total problems, internalizing, externalizing, aggressive, anxious/depressed, attention problems, rule-breaking, and withdrawn/depressed) that again included the appropriate nested structure of the data (i.e., nested participant random intercept by study site) along with the main fixed effects of the annualized exposure of interest at 9–10 years (i.e., PM2.5 and NO2), time (coded as 0 for 9–10 years and 1 for follow-up at ages 10–11 years), and an exposure-by-time interaction term (i.e., PM2.5-by-time, NO2-by-time), while adjusting for all potentially confounding variables identified as part of the minimally sufficient adjustment set described above. We took a hypothesis-based approach to selecting these eight emotional outcomes; thus, we did not correct for multiple statistical tests for these outcomes.
Sensitivity Analyses of Two-Pollutant Models: Alternative Minimally Sufficient Sets, Urban Classification, Duration of Residency, and Prenatal Exposure
In identifying our primary minimally sufficient adjustment set of variables, a secondary minimally sufficient set of confounders was identified that suggested accounting for nighttime noise in place of population density, urbanicity, and distance to roadways. For completeness, final two-pollutant models were reproduced using the second minimally sufficient set of confounders. In addition to adjusting our models for urbanicity, we also conducted a sensitivity analysis of two-pollutant models excluding participants from rural areas. As previously mentioned, we also conducted an additional sensitivity analysis of two-pollutant models using our primary minimally sufficient adjustment set excluding any geospatially derived confounders (see details above in Potential Confounding and Precision Variables section).
In addition, we conducted additional sensitivity analyses of two-pollutant models to account for duration of residency and prenatal exposure to PM2.5 and NO2. As previously mentioned, we have chosen to focus on the 1-year average exposure estimate assigned at ages 9–10 to fill an important gap in the existing literature as to how exposure during early adolescence may relate to neurological outcomes during this critical transition period.
Moreover, both experimental animal studies and human studies provide evidence that childhood air pollution exposure may have effects on brain structure and function that are independent of prenatal exposure (Fonken et al. 2011, Allen et al. 2014, Peterson et al. 2015b, Mortamais et al. 2017, Babadjouni et al. 2018) and averaging across long periods of development may prevent detection of how exposure impacts brain and behavior during critical windows of neuromaturation (Jorcano et al. 2019). However, we acknowledge the importance of accounting for differences in duration at the baseline residential address as well as any potential role of prenatal exposure on our outcomes. Thus, we also performed three additional two-pollutant model sensitivity analyses that (1) adjusted for duration at baseline address (in years), (2) included duration at baseline address as an effect modifier (i.e., moderator, creating a two-way pollutants-by-duration interaction for Aim1 and relevant three-way and two-way interactions between pollutants, time, and residential address duration for Aim 2), and (3) adjusted for prenatal exposure in our modeling strategy for the subsample of the participants that have these data. It is important to note that because the ABCD Study was not by design a birth cohort study, birth addresses and residential history information, including parents’ estimates of duration at baseline address, were collected by retrospective recall at a 2-year follow-up visit. Of those reporting a primary address at ages 9–10 years, the mean (and standard deviation) of years living at that residence was 5.44 years (3.76). Of those enrolled in the ABCD Study, 7,848 participants (66.2%) of the study sample reported valid residential information dating back to the child’s birth year (see additional details in AppendixSensitivity Analysis of Prenatal Exposure and Duration of Residency section). Prenatal exposure was assigned by averaging across the daily exposure estimates of PM2.5 and NO2 for 9 months of pregnancy based on the child’s birth date.
STATISTICAL PACKAGES
All analyses were performed using R Statistical Software (v4.1.0 (R Core Team 2021) and v4.2.2 (R Core Team 2022). For Aim 1, LMMs were fitted using the lmer() function in the R lme4 package. For Aim 2, LMMs were fitted using glmer() command in lme4 package in R (Bates et al. 2015). For all other R packages used in data cleaning, creating descriptive tables, model assumption checking, and missing data analysis, please see the relevant R scripts.
RESULTS
DESCRIPTIVE ANALYSIS
Participant characteristics of the overall ABCD study cohort as well as final analytic samples for Aim 1 and Aim 2 are shown in Table 1. Annual exposure levels among ABCD study participants varied by site (Table 2) with the mean PM2.5 exposure of 7.66 μg/m3 (SD: 1.56 μg/m3; range: 1.72–15.90 μg/m3) and mean NO2 exposure of 18.61 ppb (SD: 5.75 ppb; range: 0.73–37.94 ppb). Annual average PM2.5 and NO2 exposure levels at ages 9–10 years were weakly correlated across the entire sample (Spearman’s correlation coefficient = 0.20) (Figure 3).
Correlation matrix of prenatal and annual average exposure at ages 9–10 years (baseline visit). Correlation values represent Spearman’s r correlations from complete pairwise comparisons between annual average PM2.5, annual average NO2, prenatal PM2.5, and prenatal NO2 using the full ABCD study dataset and those with high-quality prenatal exposure data (i.e., prenatal air pollution estimates for subjects without errors in reporting of prenatal addresses or errors in reporting percentage of time spent at address during the prenatal period). Sample sizes for comparisons: annual air pollution estimates: N = 11,189; annual air pollution estimates with prenatal air pollution: n = 7,504; prenatal air pollution: n = 7,848.
ABCD = Adolescent Brain Cognitive Development; AI/AN = American Indian/Native American; GED = general educational development; HS = high school; MRI = magnetic resonance imaging; m = meters; mm = millimeters; NH/PI = Native Hawaiian/Pacific Islander; NA = not applicable.
aValues shown are either mean (standard deviation) or N (% frequency).
b “Other” race/ethnicity category includes subjects who were parent-identified as American Indian/Native American, Alaska Native, Native Hawaiian, Guamanian, Samoan, Other Pacific Islander, Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, Other Asian, or Other Race.
Table 2.
Annual Average Air Pollutant Concentrations for PM2.5 and NO2 at Ages 9–10 Years (baseline visit) Within the ABCD Study
PM2.5 = particulate matter ≤ 2.5 micrometers; NO2 = nitrogen dioxide; SD = standard deviation; IQR = interquartile range; ppb = parts per billion.
Study Site Abbreviations: CHLA = Children’s Hospital of Los Angeles, CUB = University of Colorado Boulder, FIU = Florida International University, LIBR = Laureate Institute for Brain Research, MUSC = Medical University of South Carolina, OHSU = Oregon Health and Science University, ROC = University of Rochester, SRI = SRI International, UCLA = University of California, Los Angeles, UCSD = UC San Diego, UFL = University of Florida, UMB = University of Maryland Baltimore, UMICH = University of Michigan, UMN = University of Minnesota, UPMC = University of Pittsburgh, UTAH = University of Utah, UVM = University of Vermont, UWM = University of Wisconsin-Milwaukee, VCU = Virginia Commonwealth University, WUSTL = Washington University in St. Louis, YALE = Yale University.
The mean prenatal PM2.5 exposure among participants with prenatal addresses (n = 7,848) was 10.98 μg/m3 (SD: 2.46 μg/m3; range: 2.03–23.64 μg/m3) whereas the mean prenatal NO2 exposure was 26.39 ppb (SD: 9.78 ppb; range: 1.03–63.89 ppb) (data not shown). Prenatal PM2.5 and NO2 exposure levels were weakly correlated across the entire sample (Spearman’s correlation coefficient = 0.37). Moderate correlations were seen between annual averages at 9–10 years and prenatal exposures and are presented in Figure 3.
Given the study design of primarily targeting schools in a study catchment area 50 miles from the research institute, the proportion of ABCD study participants residing in urban versus rural areas (based on urbanization classification from the 2010 Census) varied as a function of site (Appendix Table 12). Distributions of annual exposure levels by urban classification for each study site are shown in Appendix Figures 8 and 9.
AIR POLLUTION AND WHITE MATTER MICROSTRUCTURE
The mean, standard deviations, and ranges for PFC white matter structural connectivity outcomes, as measured by FA and MD values, are presented for each of the white matter tracts of interest as well as summarized across the entire brain (i.e., mean across all white matter fibers) in Table 3.
Table 3.
White Matter Microstructure by Tract and Hemisphere at Ages 9–10 Years (baseline visit)a
SD = standard deviation; ATR = anterior thalamic radiation; CGC = cingulum (body); CGH = cingulum (hippocampal); IFO = inferior fronto-occipital fasciculus; SLF = superior longitudinal fasciculus; UNC = uncinate fasciculus; CC = corpus callosum.
aDescriptive statistics created from Aim 1 dataset (n = 7,546) (Table 1), after data cleaning to remove subjects that did not pass imaging QC requirements.
bFA reflects degree of anisotropy of diffusion (e.g., restricted/directional diffusion of water); unitless with a value range of 0 (i.e., isotropic) to 1 (i.e., anisotropic).
cMD reflects overall diffusion of water; unit is mm2/sec.
dAnatomical details about each tract can be found in Appendix Table 2. Note that in Appendix Table 2 CING includes both CGC and CGH.
eAll fibers reflect a summary measure across the entire brain.
In both single and multipollutant models, we found significant PM2.5-by-hemisphere and NO2-by-hemisphere interactions for the whole brain and for nearly all white matter tracts for both FA and MD values (Table 4 and Appendix Table 13). These interactions motivated to investigate hemisphere-stratified associations to estimate and quantify differing associations between air pollution and white matter outcomes in each hemisphere using a two-pollutant model. Hemisphere-stratified two-pollutant models revealed nonlinear associations between PM2.5 and FA in the right superior longitudinal fasciculus (SLF) and linear, negative associations between NO2 and FA in the right anterior thalamic radiation (ATR), left uncinate (UNC), and the corpus callosum (CC) (Table 5 and Figure 4). In the right SLF, increased exposure to PM2.5 was associated with increased FA values at lower exposure levels (3.0 to 7.7 μg/m3) but decreased FA values at higher exposure levels (above 7.7 μg/m3). In the right ATR, the left UNC, and CC, increased exposure to NO2 was associated with decreases in FA (Table 5 and Figure 3). Hemisphere-stratified two-pollutant models also revealed nonlinear negative associations between PM2.5 exposure and MD in the whole brain white matter summary estimate in the left hemisphere, and in multiple tracts, including the left cingulate cortex hippocampal portions (CGH), left ATR, left SLF, bilateral UNC, and the CC (Table 5 and Figure 4). In the whole brain estimate and in most tracts, more pronounced negative associations were observed at higher concentrations of PM2.5 exposure, with an increase in exposure from 7.7–12.4 μg/m3 associated with percent decreases in MD ranging from –0.76% to –1.06%, as compared to percent decreases associated with lower levels exposure (ranging from –0.19% to –0.526%). An exception was observed in the left SLF, where percent decreases in MD were comparable at lower exposure levels (–0.785%) and higher exposure levels (–0.700%).
Associations between annual average PM2.5 and NO2 exposure and PFC white matter fractional anisotropy (FA) at ages 9–10 years (baseline visit). (A) Visualization of white matter tracts showing significant effects of annual average PM2.5 and NO2 at ages 9–10 years from hemisphere-specific two-pollutant models predicting fractional anisotropy (FA) at ages 9–10 years, adjusting for a minimally sufficient set of covariates, including: child’s age, sex, race/ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, urban classification, and MRI precision variables (manufacturer, head motion, and child handedness). Anatomical details about each tract can be found in Appendix Table 2. (B) Plots of non-linear associations between annual average PM2.5 and FA at ages 9–10 years and linear associations between annual average NO2 and FA at ages 9–10 years for significant white matter tracts from hemisphere-specific two-pollutant LMMs adjusted for minimally sufficient set of covariates. Shaded bands represent 95% confidence intervals. Significant tracts for PM2.5 effects were identified with a Type III analysis of variance (Satterthwaite method) comparing models with both a spline PM2.5 term and a linear term for NO2 to models with a linear term for NO2 only (i.e., a “null” model). Significant tracts for NO2 were identified using the P value associated with the linear NO2 beta coefficient in models with both a spline PM2.5 term and a linear term for NO2. This P value was calculated using the Satterthwaite method for computing degrees of freedom and t statistics. Significance level was set a priori at P < 0.05 for all tests. Abbreviations: A = anterior; P = posterior; R = right hemisphere; L = left hemisphere; ATR = anterior thalamic radiation (green); SLF = superior longitudinal fasciculus (light blue); UNC = uncinate fasciculus (darker blue); CC = corpus callosum (red).
Table 4.
Adjusted Annual Average Two-Pollutant Models Including Pollutant-by-Hemisphere Interaction Terms for 1-Year Average Air Pollution Exposure and PFC White Matter Connectivity at Ages 9–10 Years (baseline visit)a
aMarginal R2 (R2m) and conditional R2 (R2c) from two-pollutant LMMs with pollutant-by-hemisphere interaction terms for both annual average air pollutant exposures at ages 9–10 years, a natural cubic splines term for PM2.5, a linear term for NO2, and adjusted for minimally sufficient set of covariates including child’s age, sex, race/ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, urban classification, and MRI precision variables (manufacturer, head motion, and child handedness).
bAnatomical details about each tract can be found in Appendix Table 2. Note that in Appendix Table 2 CING includes both CGC and CGH.
cFDR-corrected P value from Type III analysis of variance (Satterthwaite method) comparing models with no air pollution terms (“null” models) to models with both baseline PM2.5 and baseline NO2 (with hemisphere interaction terms) (included to indicate the overall significance of associations between the air pollutants in this study and the white matter outcomes). Significance level was set a priori at P < 0.05.
dFDR-corrected P values from Type III analysis of variance (Satterthwaite method) comparing models with no hemisphere-by-pollutant interaction (with hemisphere included as an additive covariate) to models with a hemisphere-by-pollutant interaction term for the indicated air pollutant. Significance level was set a priori at P < 0.05.
eFA reflects degree of anisotropy of diffusion (e.g., restricted/directional diffusion of water); unitless with a value range of 0 (i.e., isotropic) to 1 (i.e., anisotropic).
fAll fibers reflect a summary measure across the entire brain.
gCC crosses both hemispheres; marginal R2 (R2m) and conditional R2 (R2c) from LMM two-pollutant models with no pollutant-by-hemisphere interaction terms, and with a natural cubic splines term for PM2.5, a linear term for NO2, and adjusted for minimally sufficient set of covariates.
hMD reflects overall diffusion of water; unit is mm2/sec.
Table 5.
Percent Change or Beta Estimates for Fractional Anisotropy and Mean Diffusivity at Ages 9–10 Across Increments of 1-Year Annual Average PM2.5 and NO2 Exposure at Ages 9–10 Years (baseline visit)a
SD = standard deviation; CI = confidence interval; ATR = anterior thalamic radiation; CGC = cingulum (body); CGH = cingulum (hippocampal); IFO = inferior fronto-occipital fasciculus; SLF = superior longitudinal fasciculus; UNC = uncinate fasciculus; CC = corpus callosum.
aEffect sizes of air pollution terms (annual average PM2.5 and NO2 at ages 9–10 years) from hemisphere-specific two-pollutant LMMs including natural cubic splines terms for PM2.5, linear terms for NO2, and adjusted for minimally sufficient sets including: child’s age, sex, race/ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, urban classification, and MRI precision variables (MRI manufacturer, head motion, and child handedness). Given the nonlinear associations between PM2.5 and white matter outcomes, percent change between mean PM2.5 exposure (7.7 μg/m3) and exposure levels ± 3 standard deviations (3.0 μg/m3 to 12.4 μg/m3) from the mean are reported to indicate the differing associations observed at lower and higher levels of exposure. P values for PM2.5 effects were identified with a Type III analysis of variance (Satterthwaite method) comparing models with both a spline PM2.5 term and a linear term for NO2 to models with only a linear term for NO2 (i.e., a “null” model). For NO2, a linear beta coefficient is reported. P values for NO2 effects were calculated in using the Satterthwaite method for computing degrees of freedom and t statistics. Significance at * P < 0.05, ** P < 0.01, and *** P < 0.001.
bFA reflects degree of anisotropy of diffusion (e.g., restricted/directional diffusion of water); unitless with a value range of 0 (i.e., isotropic) to 1 (i.e., anisotropic).
cAnatomical details about each tract can be found in Appendix Table 2. Note that in Appendix Table 2 CING includes both CGC and CGH.
dMD reflects overall diffusion of water; unit is mm2/sec.
eAll fibers reflects a summary measure across the entire brain.
In sensitivity analyses of two-pollutant models utilizing the secondary minimally sufficient adjustment set, results remained nearly identical for all FA and MD outcomes (Appendix Table 14), except for the association between NO2 and FA in the CC, which was no longer statistically significant(data not shown). Negative associations were still observed between PM2.5 and MD in significant tracts. At higher levels of PM2.5 exposure, the magnitude of negative associations was comparable using both adjustment sets. In models with the secondary minimally sufficient set, the magnitude of negative associations at lower levels of PM2.5 exposure was more pronounced across all tracts (ranging from –0.28% to –1.09%) (Appendix Table 14), as compared to models with the primary minimally sufficient adjustment set (ranging from –0.19% to –0.79%) (Table 5). Negative, linear associations between NO2 remained similar with both adjustment sets.
In sensitivity analyses excluding participants from rural areas (n = 6,924), results for MD outcomes were nearly identical in direction and magnitude (Appendix Table 15), except that the association between PM2.5 and MD in the left SLF was no longer significant (data not shown). Results for FA outcomes were also nearly identical in magnitude and direction to the primary analysis (Appendix Table 15), except that the association between PM2.5 and FA in the right SLF was no longer significant (data not shown). Additional negative associations were observed between NO2 and FA in the left IFO and in the white matter summary estimate in both hemispheres (Appendix Table 15). In all tracts, the negative associations between NO2 and FA were slightly more pronounced in magnitude (Appendix Table 15) as compared to associations between NO2 and FA in the primary analysis (Table 5).
In sensitivity analyses adjusting for the primary minimally sufficient adjusting set but excluding the identified geospatially derived confounders, results for associations between PM2.5 and FA and MD outcomes remained nearly identical. In these models, negative associations between PM2.5 and MD became slightly more pronounced in some tracts and slightly less pronounced in other tracts, with no clear pattern in differences between the two sets of models (Appendix Table 16). The magnitude of the significant association between PM2.5 and FA in the right SLF became slightly more pronounced in these models as compared to primary models (i.e., more positive at lower levels of PM2.5 and more negative at higher levels of PM2.5 exposure) (Appendix Table 16). The associations between NO2 and FA in the CC, the right ATR, and the left UNC were no longer significant(data not shown).
In additional sensitivity analyses adjusting for residential duration, all FA and MD outcomes remained significant as identified in the primary analyses (Table 6). In all tracts with significant associations for FA and MD, all associations remained similar in both direction and magnitude to the primary analysis. In these residential duration sensitivity models, additional negative associations were observed between NO2 and FA in the white matter summary estimate in both hemispheres, again of a similar magnitude as observed in other significant tracts in these models (Table 6).
Table 6.
Percent Change Estimates or Beta Coefficients for Fractional Anisotropy and Mean Diffusivity at Ages 9–10 Years Across Increments of 1–Year Annual Average PM2.5 and NO2 Exposure at Ages 9–10 Years (baseline visit), Adjusting for Residential Durationa
SD = standard deviation; ATR = anterior thalamic radiation; CGC = cingulum (body); CGH = cingulum (hippocampal); IFO = inferior fronto–occipital fasciculus; SLF = superior longitudinal fasciculus; UNC = uncinate fasciculus; CC = corpus callosum.
aEffect sizes of air pollution terms (annual average PM2.5 and NO2 at ages 9–10 years) from hemisphere–specific two–pollutant LMMs including natural cubic splines terms for PM2.5, linear terms for NO2, and adjusted for minimally sufficient sets including child’s age, sex, race/ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, urban classification, and MRI precision variables (MRI manufacturer, head motion, and child handedness) and duration of time living at baseline address (years). Given the nonlinear associations between PM2.5 and white matter outcomes, percent change between mean PM2.5 exposure (7.7 μg/m3) and exposure levels ± 3 standard deviations (3.0 μg/m3 to 12.4 μg/m3) from the mean are reported to indicate the differing associations observed at lower and higher levels of exposure. P values for PM2.5 effects were identified with a Type III analysis of variance (Satterthwaite method) comparing models with both a spline PM2.5 term and a linear term for NO2 to models with only a linear term for NO2 (i.e., a “null” model). For NO2, a linear beta coefficient is reported. P values for NO2 effects were calculated in using the Satterthwaite method for computing degrees of freedom and t statistics. Significance at * P < 0.05 ** P < 0.01 *** P < 0.001.
bFA reflects degree of anisotropy of diffusion (e.g., restricted/directional diffusion of water); unitless with a value range of 0 (i.e., isotropic) to 1 (i.e., anisotropic).
cAnatomical details about each tract can be found in Appendix Table 2. Note that in Appendix Table 2 CING includes both CGC and CGH.
dAll fibers reflect a summary measure across the entire brain.
eMD reflects overall diffusion of water; unit is mm2/sec.
In two-pollutant sensitivity analyses examining address duration as an effect modification (i.e., moderator) of the exposure, no significant interactions between either PM2.5 or NO2 exposure and address duration were seen for FA outcomes (Appendix Table 17). However, for MD outcomes, significant PM2.5 -by-address duration interactions were noted for the whole brain white matter summary estimate (i.e., all fibers) in both hemispheres, and for multiple tracts, including the left ATR and bilateral CGC (Appendix Table 17). Because a spline term was used for PM2.5 (and beta coefficients from these spline terms are not easily interpretable), in order to understand these interactions we examined associations between PM2.5 exposure and each outcome at different levels of residential address duration (including the first [2 years], second [5 years] and third quartiles [9 years] for residential address duration) (Appendix Table 18 and Appendix Figure 10). The magnitude of associations between PM2.5 and MD at different quartiles of residential duration was variable. For those reporting living 2 years at the residential location, PM2.5 was associated with greater decreases in MD values at higher exposure levels (7.7 to 12.4 μg/m3) as compared to lower exposure levels (3.0 to 7.7 μg/m3) and differences in effects between PM2.5 exposure levels were somewhat pronounced (2- to 16-fold differences). For those reporting living 5 years at the residential location, PM2.5 was still associated with greater decreases in MD values at higher exposure levels as compared to lower, but the differences were less pronounced (0.6- to 6-fold differences). For those reporting living 9 years at the residential location, the estimated percent changes in MD were very similar at lower and higher exposure levels with no notable pattern between PM2.5 exposure levels (Appendix Table 18 and Appendix Figure 10).
Significant NO2-by-address duration interactions were also seen for MD outcomes (Appendix Table 17 and Appendix Table 19), including whole brain white matter summary estimate (i.e., all fibers) in both hemispheres, as well as the ATR, CGC, and SLF in the left hemisphere. The NO2-by-residential address duration interaction terms for all of these tracts were positive (ranging from 0.00009 to 0.00015) (Appendix Table 19). These interactions are visualized in Appendix Figure 11, where, for those reporting 2 years at their residence, associations between NO2 and MD appear negative. For those reporting 5 years at their residence, associations appear to be flat, slightly positive, or slightly negative, with no clear pattern observed between tracts (Appendix Figure 11). However, for those reporting living at the residence for 9 years, associations between NO2 exposure and MD appear to be positive (Appendix Figure 11).
In a sensitivity analysis of two-pollutant models utilizing a smaller subsample of subjects with available prenatal air pollution exposure estimates and adjusting for known prenatal exposure to PM2.5, NO2, and address duration, NO2 associations with FA remained significant for the right ATR, left UNC, and the CC; PM2.5 associations with MD remained significant for the white matter summary estimate in the left hemisphere of the brain, left SLF, left UNC, and CC (Table 7). The magnitude of negative associations between NO2 and FA remained similar, but the magnitude of negative associations between PM2.5 and MD were more pronounced in models adjusting for prenatal exposure (Table 7). In these sensitivity models, additional negative associations were observed between NO2 and FA in the white matter summary estimate in both hemispheres, in the left ATR, the left CGC, bilaterally in the IFO, and the right UNC (Table 7), and these associations were similar in magnitude to associations observed in the primary analysis. Overall, adjusting for prenatal exposure produced more significant negative associations with a stronger magnitude between PM2.5 and MD and more significant negative associations with similar magnitude between NO2 and FA. However, prenatal PM2.5 and prenatal NO2 did not relate to FA or MD in any of these white matter outcomes (Appendix Table 20). Because the dataset with available data on prenatal air pollution is much smaller than the main analysis dataset (n = 4,920), an exploratory follow-up analysis was performed, including only the subjects in this smaller prenatal air pollution dataset, but excluding the prenatal air pollution exposure estimates as covariates. This analysis yielded the same set of significant associations as the models with prenatal air pollution exposure estimates included (data not shown). Thus, from this subsequent analysis, it is possible that the additional significant associations observed in prenatal sensitivity models were due to selection bias and not due to confounding effects accounted for by the inclusion of the prenatal air pollution exposure estimates.
Table 7.
Percent Change Estimates or Beta Coefficients for Fractional Anisotropy and Mean Diffusivity at Ages 9–10 Years Across Increments of 1-Year Annual Average PM2.5 and NO2 Exposure at Ages 9-–10 Years (baseline visit), Adjusting for Prenatal Exposure and Residential Durationa
SD = standard deviation; ATR = anterior thalamic radiation; CGC = cingulum (body); CGH = cingulum (hippocampal); IFO = inferior fronto-occipital fasciculus; SLF = superior longitudinal fasciculus; UNC = uncinate fasciculus; CC = corpus callosum.
aEffect sizes of air pollution terms (annual average PM2.5 and NO2 at ages 9–10 years) from hemisphere-specific two-pollutant LMMs including natural cubic splines terms for PM2.5, linear terms for NO2, and adjusted for minimally sufficient sets including: child’s age, sex, race/ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, urban classification, MRI precision variables (MRI manufacturer, head motion, and child handedness), duration of time living at baseline address (years), prenatal PM2.5 exposure, and prenatal NO2 exposure. Given the nonlinear associations between PM2.5 and white matter outcomes, percent change between mean PM2.5 exposure (7.7 μg/m3) and exposure levels ± 3 standard deviations (3.0 μg/m3 to 12.4 μg/m3) from the mean are reported to indicate the differing associations observed at lower and higher levels of exposure. P values for PM2.5 effects were identified with a Type III analysis of variance (Satterthwaite method) comparing models with both a spline PM2.5 term and a linear term for NO2 to models with only a linear term for NO2 (i.e., a “null” model). For NO2, a linear beta coefficient is reported. P values for NO2 effects were calculated using the Satterthwaite method for computing degrees of freedom and t statistics. Significance at *P < 0.05 ** P< 0.01 ***P < 0.001.
b FA reflects degree of anisotropy of diffusion (e.g., restricted/directional diffusion of water); unitless with a value range of 0 (i.e., isotropic) to 1 (i.e., anisotropic).
c Anatomical details about each tract can be found in Appendix Table 2. Note that in Appendix Table 2 cingulum includes both CGC and CGH.
d All fibers reflects a summary measure across the entire brain.
e MD reflects overall diffusion of water; unit is mm2/sec.
AIR POLLUTION AND EMOTIONAL OUTCOMES
Descriptive statistics for internalizing, externalizing, and total problems as well as anxiety/depression, withdrawal/depression, attention, rule-breaking, and aggression scores at ages 9–10 years and 10–11 years are reported in Table 8. Overall, as expected, slight increases were seen in internalizing problems (i.e., anxiety/depression and withdrawn depression), whereas slight decreases were seen in externalizing problems (i.e., attention, rule breaking, aggression) over time (Table 8). Single- and two-pollutant results were similar in direction and magnitude (Table 9 and Appendix Tables 21 and 22). Thus, we report the two-pollutant model results in greater detail, as these models are the most complete models and better account for co-exposure. Based on the main effect of exposures in our models, the annual PM2.5 and NO2 exposure was not associated with emotional outcomes at ages 9–10 years, however, the differential effect of 1-year exposure was associated with some of the emotional outcomes at the 1-year follow-up (Table 9). In contrast to our hypothesis, 1-year exposure was negatively associated with emotional outcomes. Specifically, looking at the interaction effects, a one-unit increase in PM2.5 exposure was differentially associated on average with a 0.26 decrease in internalizing problems, a 0.15 decrease in anxiety/depression, and a 0.07 decrease in aggressive problems at the 1-year follow-up. Similarly, a one-unit increase in NO2 exposure was differentially associated on average with a 0.18 decrease in internalizing and a 0.16 decrease in total problems at the 1-year follow-up. Importantly, CBCL scores are normalized for age and sex, the air pollution effects are less than 0.5 SD difference in the outcome, suggesting the effects are extremely small and may not be clinically meaningful in a real-world setting, as they may reflect only a one-point change in reported problems.
Table 8.
Child Behavior Checklist (CBCL) t Scores at Ages 9–10 Years (baseline visit) and at 1-Year Follow-Up
a Scores and subscales are scaled based on age and sex with a mean of 50 and SD of 10.
Table 9.
Adjusted Associations Between 1-Year Annual Average Exposure at Ages 9–10 Years and CBCL Outcomesa at Ages 9–10 Years (baseline visit) and at 1-Year Follow-Upb
CBCL = Child behavior checklist; SD = standard deviation; CI = confidence interval.
a CBCL outcomes are t scores normalized on age and sex with a mean of 50 and SD of 10.
Number of observations = 17,592; Number of groups: subjects within study sites = 9,334; Study sites = 21.
Significance at *P < 0.05 ** P < 0.01 *** P < 0.001.
bEstimates from two-pollutant (annual average PM2.5 and NO2 at ages 9–10 years) LMMs adjusted for minimally sufficient set including child’s race/ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, and urban classification.
In sensitivity analyses of the two-pollutant models utilizing the alternative minimally sufficient adjustment set, results remained nearly identical, except the differential association between NO2 and internalizing behavior at the 1-year follow-up did not remain statistically significant (Appendix Table 23).
In sensitivity analyses excluding participants from rural areas (n = 8,570), results remained largely similar, except the differential effect of NO2 exposure on total problems was not significant and internalizing problems became significant at the 1-year follow-up, while the differential effect of PM2.5 exposure for total problems and withdrawn/depression at the 1-year follow-up was marginally significant (Appendix Table 24).
A similar pattern was also seen in the identified associations from two-pollutant models when adjusting for the primary minimally sufficient adjustment, excluding the identified geospatially derived confounders, except the differential effect of PM2.5 and aggression was no longer significant (Appendix Table 25).
In two-pollutant analyses adjusting for how many years the child resided at the residential address, the magnitude and significance of results remained similar, except the interaction between NO2 and time for total problems was no longer significant (Table 10). In two-pollutant analyses examining address duration as an effect modification (i.e., moderator) of the exposure, three-way exposure-by-time-by-address duration interaction terms were not significant (data not shown). Therefore, the three-way interaction variables were removed to create a more parsimonious model. In the two-pollutant analyses including two-way interaction terms for residential duration, the main findings remained similar for PM2.5-by-time for internalizing, anxiety/depression, and aggressive problems, but NO2-by-time for internalizing problems was no longer statistically significant (Appendix Table 26). In addition, significant two-way interactions for PM2.5-by-address duration were seen for internalizing problems, as well as anxiety/depression, withdrawal/depression, and attention subscales. To further understand these interactions, we again examined associations between the exposure and each outcome, using the first (2 years), second (5 years), and third quartile (9 years) for residential address duration. Interaction plots showed that for each of these outcomes, higher PM2.5 exposure was associated with more emotional problems in those living at the address for 9 years, whereas those reporting living at the residential address for 2 years showed a negative association (i.e., higher PM2.5 associated with less emotional problems) (Appendix Figure 12).
Table 10.
Associations Between 1-Year Annual Average Exposure at Ages 9–10 Years (baseline visit) and CBCL Outcomesa at Ages 9–10 Years (baseline visit) and at 1-Year Follow-Up, Adjusting for Residential Durationb
bEstimates from two-pollutant (annual average PM2.5 and NO2 at ages 9–10 years) LMMs adjusted for residential address duration (in years) in addition to minimally sufficient set including child’s race/ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, and urban classification.
Similarly, after accounting for prenatal exposure in the two-pollutant models, the magnitude and significance of PM2.5 exposure by time results were nearly identical, however, NO2 exposure was no longer found to relate to changes in CBCL scores at the 1-year follow-up (Table 11). In the two-pollutant models, prenatal PM2.5 and prenatal NO2 exposures did not significantly relate to any of the emotional outcomes examined (Appendix Table 27).
Table 11.
Associations Between 1-Year Annual Average Exposure at Ages 9–10 Years (baseline visit) and CBCL Outcomesa at Ages 9–10 Years (baseline visit) and at 1-year Follow-Up, Adjusting for Prenatal Exposures and Residential Durationb
CBCL = child behavior checklist; SD = standard deviation; CI = confidence interval.
a CBCL outcomes are t scores normalized on age and sex with a mean of 50 and SD of 10. Number of observations = 16,925; Number of groups: subjects within study sites = 8,975; Study sites = 21.
Significance at * P< 0.05, **P < 0.01, and ***P < 0.001.
bEstimates from two-pollutant (annual average PM2.5 and NO2 at ages 9–10 years) LMMs adjusted for prenatal exposure to PM2.5 and NO2, residential address duration (in years), and minimally sufficient set including child’s race/ethnicity, total household income, child’s physical activity, child’s screen time use, distance to roadways, perceived neighborhood quality, population density, and urban classification.
DISCUSSION AND CONCLUSIONS
Our objective was to characterize associations between annual ambient PM2.5 and NO2 exposure during early adolescence and prefrontal white matter structural connectivity as well as emotional behavior using a large, multisite nationwide study of 9–10-year-old participants in the United States.
MAIN FINDINGS
Air Pollution and White Matter Development
This study was the first large, multisite US study to find associations between annual PM2.5 and NO2 exposures and altered microstructure within white matter tracts that innervate the prefrontal cortex in early adolescence. We found that higher PM2.5 exposure was associated with decreases in mean diffusivity (MD) in the left hemisphere across the entire brain as well as specific tracts, including the left anterior thalamic radiations, left cingulum (hippocampal portion), left superior longitudinal fasciculus, as well as bilateral uncinate fasciculus, and corpus callosum. We also found that higher NO2 exposure was associated with decreases in fractional anisotropy (FA) in the right anterior thalamic radiations, left uncinate, and the bilateral corpus callosum. In contrast, higher PM2.5 exposure was also associated with nonlinear increases in FA in the right superior longitudinal fasciculus. Taken together, these findings suggest that exposure to ambient air pollution during early adolescence may contribute to microarchitectural differences in fronto-parietal and limbic white matter circuitry, which is important for attention (superior longitudinal fasciculus), emotional processing (uncinate, cingulum, and superior longitudinal fasciculus), and memory (cingulum and superior longitudinal fasciculus) (Von Der Heide et al. 2013, Bubb et al. 2018, Nakajima et al. 2020, Senova et al. 2020). The robustness of these findings was supported by similar findings in sensitivity analyses using an alternative adjustment set of confounders and accounting for the average length of living at the residential address at study enrollment. Similarly, overall effects between NO2 and FA, as well as PM2.5 effects for the left hemisphere, as well as the left superior longitudinal fasciculus, left uncinate, and corpus callosum remained significant after accounting for prenatal exposure. Lastly, annual average exposure effects on MD varied by length of residential duration, with stronger negative associations noted for PM2.5 at lower exposures (3.0 to 7.7 μg/m3) for those reporting living at that residential location longer (e.g., 9 years as compared to 2 years). For NO2, increasing the length of residential duration resulted in a more positive association between NO2 and MD, albeit most main effects of NO2 on MD were not deemed significant. As mentioned, diffusion tensor imaging in the in vivo brain allows for the assessment of white matter integrity, reflecting various cellular processes including changes in axonal diameter, fiber density, myelin structure, and support cells (Beaulieu 2002). Thus, these findings suggest that PM2.5 exposure is linked to potential increases in cellular barriers in white matter (reflected by decreases in MD), whereas NO2 exposure is linked to reduced directionality of restriction, which could be reflective of axon diameter or myelination of fibers. Given that our two-pollutant modeling approach accounts for potential confounding of co-exposure, these findings suggest that PM2.5 and NO2 may have unique impacts on the cellular composition of key white matter tracts during childhood.
Beyond the current study, three additional studies to date have implemented diffusion tensor imaging to understand how air pollution impacts white matter microstructural development. Peterson and colleagues (Peterson et al. 2022) examined prenatal PM2.5 exposure and diffusion tensor imaging in 259 children 6–14.7 years of age from the Columbia Center for Children’s Environmental Health longitudinal cohort study of African American and Dominican women in New York City. Specifically, this study estimated prenatal PM2.5 exposure based on the residence of the mother as well as acquired polycyclic aromatic hydrogen (PAH) samples through personal monitoring over 48 hours during the third trimester of pregnancy in 727 African American and Dominican mothers from northern Manhattan (1998–2006). In contrast to the current study, prenatal PM2.5 was related to greater diffusion within the white matter of posterior bundles and anterior corona radiata, but there was a similar negative association between PAH and less diffusion within the internal capsule. Although these findings reported by Peterson and colleagues accounted for postnatal urinary PAH levels, more recent exposure levels at the time of brain imaging were not assessed. In another study by Lubczyńska and colleagues, both prenatal and postnatal exposure (e.g., birth to MRI visit) based on residential location was found to be associated with diffusion tensor imaging outcomes in 2,954 children ages 9–12 years old from the Generation R Study, which is a population-based birth cohort from Rotterdam, Netherlands (2002–2006) (Lubczyńska et al. 2020). In multipollutant models using a deletion/substitution/addition algorithm, different effects were noted depending on the exposure period and pollutant examined. Prenatal findings showed that higher concentration of PM2.5 was associated with significantly lower FA across the entire brain, and a higher concentration of elemental silicon was related to significantly higher global MD. Childhood exposures, however, showed that higher NOx was associated with lower global FA, whereas higher global MD values were seen for elemental zinc and the oxidative potential of PM. However, when accounting for both prenatal and childhood periods of exposure, only the significant associations of silicon during pregnancy and zinc during childhood were found to relate to higher global MD values. Moreover, associations between various exposures and microstructure did not pass multiple comparison corrections in specific white matter tracts. More recently, a follow-up study examining periods of air pollution vulnerability using the Generation R study was published by Binter and colleagues (Binter et al. 2022). Higher levels of PM2.5 exposure from conception to 4 years old, as well as higher levels of NO2 exposure from 3.6 to 4.8 years old, are related to lower FA and higher MD values for global brain estimates in early adolescence. Taken together, the current study and these additional studies suggest that ambient air pollution exposure during development is linked to differences in in vivo patterns of white matter microstructure in childhood. Although the directionality of white matter microstructure findings is mixed, the noted differences might stem from divergent study samples, timing of exposure assessment during development, and/or differences in MRI techniques. For example, PM2.5 exposure levels were higher in the previous studies as compared to the current study (average of 16.5–17.4 μg/m3 vs. 7.7 μg/m3). Relatively low levels of exposure during childhood may result in increasing cellular processes (reflected by decreases in MD, as seen in our study), such as activation or infiltration of microglial cells, whereas higher levels may lead to cellular or myelin disruption (reflected by increases in MD, which have been noted in other studies). Alternatively, these differing findings may reflect distinct diffusion-MRI acquisition parameters. The previous studies utilized single-shell DTI sequences, whereas the current study utilized a multi-shell HARDI sequence with various b values, allowing for increased sensitivity and specificity (Mori and Tournier 2014) and may better capture the restricted compartment as compared to single-shell sequences using lower b values (Palmer et al. 2022). Lastly, while the existing studies and the current study examined 9-month averages of prenatal PM2.5 exposure, each study to date assessed postnatal exposure differently. Although Lubczyńska and colleagues were able to account for lifetime exposure, it is feasible that a single postnatal estimate may have dampened stronger exposure effects during more discrete periods of development, including 1-year exposure at the time of assessment. Similarly, Peterson and colleagues were unable to account for the potential confounding of other various periods of postnatal exposure beyond age 5 with urinary PAH levels. The current study is also limited to other periods of exposure and their potential confounding effects (see additional details in the Limitations section below). Nonetheless, this emerging literature suggests that ambient air pollution exposure during development may impact white matter structural connectivity.
These findings build upon emerging literature showing that outdoor ambient air pollution exposure, especially to PM2.5 as well as NOx, are linked to differences in the size and shape of the brain (Peterson et al. 2015a, Guxens et al. 2018, Cserbik et al. 2020, Lubczyńska et al. 2021, Guxens et al. 2022, Peterson et al. 2022), as well as brain function (Pujol et al. 2016b, Pérez-Crespo et al. 2022, Cotter et al. 2023). In fact, our team has conducted additional studies using the ABCD Study showing that 1-year annualized PM2.5 exposure relates to cortical thickness (Cserbik et al. 2020), subcortical gray matter microarchitecture (Sukumaran et al. 2023), as well as disruption of longitudinal changes in functional brain development from ages 9–12 years (Cotter et al. 2023). In these previous ABCD studies as well as the current study, air pollution associations differed in the left versus right hemispheres of the brain. Given both notable known neuroanatomical differences between the left and right sides of the brain as well as lateralization of several brain functions (Toga and Thompson 2003), it is biologically plausible that exposure effects may vary between the two hemispheres. While the exact mechanisms underlying inherent central nervous system asymmetries remain largely unknown (Lubben et al. 2021), structural and functional differences have been widely noted between the two cerebral hemispheres at the macroscopic, microstructure, and molecular levels (Mundorf et al. 2021, Ocklenburg 2022, Ocklenburg et al. 2022). Genomewide association and epigenetic studies also suggest brain asymmetries occur via differences in functional genetic pathways of microtubule regulation, neurogenesis, and axonogenesis, which are involved in both neuronal development and organizational differences in gene expression between the two hemispheres (Lubben et al. 2021). Thus, these known asymmetries in brain structure and function may contribute to the current observed hemisphere-specific patterns and some previous studies showing hemispheric differences between air pollution exposure and brain outcomes during development (Büchel et al. 2004, Butter 2006, Bonekamp et al. 2007, Bava et al. 2010, Peterson et al. 2015b, Guxens et al. 2018, Cserbik et al. 2020). However, the functional implications, if any, of these hemispheric differences noted in air pollution and brain imaging studies remain to be determined. Air pollution effects may appear more prominent on one side of the brain as compared to the other because of underlying evolutionary, hereditary, and/or developmental factors contributing to brain asymmetry, without any specific unilateral biological effects driving these differences. In other words, known underlying differences in brain asymmetry may render air pollution associations more salient on the right versus left side for some brain regions. Ongoing studies focused on brain asymmetry in the broader field of neuroscience will ultimately be important to help clarify potential specificity and inference of observed unilateral associations seen between air pollution and brain outcomes to date.
While diffusion tensor imaging does not directly measure cellular processes, our findings are congruent with animal and cellular exposure models that show air pollution impacts brain structure and function. Depending on particle size, particulate matter can impact the brain via nasal and respiratory pathways (Genc et al. 2012). In the respiratory pathway, particles may pass from the lungs into the bloodstream to infiltrate the blood–brain barrier or create systematic secondary effects through inflammation. Microglia, the resident immune cells of the central nervous system, are some of the first cells to respond to pollutants and cause inflammatory activation (Gomez-Budia et al. 2020). Resting microglia have small somas with thin processes, while activated microglia in animal models and cell cultures exhibit increases in somal size, with shorter processes (Subhramanyam et al. 2019). The negative associations between PM2.5 and mean diffusivity observed in this study suggest an increased barrier to water diffusion, which is congruent with the potential increases in the number or size of support cells in white matter tracts. To better test this hypothesis, our team implemented novel biophysical models of diffusion to investigate how PM2.5 exposure may relate to different types of diffusion within a given brain region in early adolescence, also utilizing the ABCD study cohort (Burnor et al. 2021). Those analyses showed that PM2.5 effects were related to increases in restricted, spherical diffusion in these white matter pathways (i.e., intracellular isotropic diffusion), which is also in agreement with the potential influx of microglia cell number and/or size (Burnor et al. 2021). In addition to oxidative stress and neuroinflammation, NO2 has been found in animal studies to disrupt the myelin sheath (Li et al. 2021). Thus, given that MD quantifies the magnitude of diffusion, whereas FA is not dependent on the amount of diffusion but the overall directionality of diffusion, our findings suggest PM2.5 may be increasing the number of cellular boundaries via changes in white matter composition, whereas NO2 may influence cellular processes that contribute to unidirectional water diffusion, such as axonal organization and/or myelination, as detected by diffusion tensor imaging.
Air Pollution and Emotional Development
Internalizing behaviors, characterized by behavior inwardly directed toward oneself, include negative emotions such as anxiety and depression (Achenbach and Rescorla 2001). Externalizing behavior is directed outward, including aggression and defiance (Achenbach and Rescorla 2001). Many of these externalizing behaviors decrease in frequency and magnitude with age (Achenbach and Rescorla 2001). As such, the current study did see small increases in internalizing behaviors and decreases in externalizing behaviors over the 1-year follow-up period. In contrast to our hypothesis, however, higher levels of annual average PM2.5 and NO2 exposure were not associated with a greater number of emotional problems at 9–10 years or 1 year later. Rather, higher PM2.5 and NO2 exposures were associated with slightly fewer problems over time on some of the CBCL outcomes. These unexpected effects largely remained unchanged in all sensitivity analyses conducted, including adjusting for various spatial confounders, excluding participants from rural areas, including a fixed effect and effect modification of residential duration, and accounting for prenatal exposures. Although the directions of the relationships between the pollutants and longitudinal CBCL outcomes are counterintuitive, it is important to consider the magnitude of the parameter estimates in such a large sample, rather than the statistical significance of these findings. That is, higher pollution exposure was associated with less than a single unit decrease on any given scale. Given the three-point Likert scale used to endorse 113 items on the CBCL (i.e., 0 = not true, 1 = somewhat or sometimes true, 2 = very true or often true), a one-unit change is likely clinically negligible and may not have real-world implications. Thus, the small effect sizes seen are likely to have very little, if any, clinical significance for the onset of psychopathology in children and adolescents.
Another finding that emerged from one of our sensitivity analyses was that residential address duration modified the association detected between 1-year PM2.5 exposure and emotional behaviors in some outcomes. Specifically, a small positive association was seen between 1-year estimates of PM2.5 and internalizing, anxiety/depression, depression/withdrawal, and attention problems for children whose parents reported living at the primary address longer (i.e., 9 years as compared to 2 years). However, it is important to note that when looking at the range of each of the outcomes, the magnitude of all these parameter estimates was again very small. Moreover, address duration did not modify the significant negative interactions noted between higher 1-year exposures and less emotional outcomes for the 1-year follow-up. It should also be noted that residential duration information in the current study may be subject to recall bias, and while intriguing, exposure estimates were not linked beyond the annual average. Thus, while residential duration may serve as a proxy for longer-term exposure to annualized PM2.5 at the same address, it does not accurately capture spatial and temporal differences in exposure at that exact residential address over time, or account for previous exposure levels for those individuals who lived at their residence for fewer years. Nonetheless, these findings suggest future research is warranted to further understand whether longer periods of residential exposure, beyond an annual average, may influence emotional outcomes in childhood and adolescence.
Our study focuses on childhood exposure at ages 9–10 years old, a developmental period currently underrepresented in the literature. About 26% of studies on pollution-related mental health problems cover this age range (Zundel et al. 2022), despite the high incidence of psychiatric diagnoses that emerge across adolescence (Kessler et al. 2005, Kessler et al. 2007, Solmi et al. 2022). Yet even these air pollution studies focused on mental health outcomes in youth have reported mixed findings. Some of the earliest longitudinal research in this area comes from the Columbia Center for Children’s Environmental Health longitudinal cohort study of African American and Dominican women in New York City. These essential studies found that prenatal exposure to airborne PAHs, which come from fossil fuel combustion, was linked to greater symptoms of anxiety, depression, and attention problems at ages 4–5 and 6–7 years-old children (Perera et al. 2011, 2012), as well as to ADHD symptoms at age 9 (Perera et al. 2014b) and to delayed emotional regulation throughout childhood (Margolis et al. 2016). However, in a more recent study that included using either the Strength and Difficulties Questionnaire or the CBCL in eight European population–based birth cohorts, prenatal and postnatal air pollution, including PM2.5 and NO2 exposure, were not found to relate to depression, anxiety, or aggression in >13,000 children ages 7–11 years old (Jorcano et al. 2019). In fact, higher postnatal exposure was linked with overall lower odds of having symptoms in the borderline/clinical range when assessed cross-sectionally with the CBCL, albeit the results did not reach statistical significance. Similar findings were also noted when using the quantitative scores of the symptom scales (Jorcano et al. 2019) as implemented in the current analysis. A similar study assessing ADHD symptoms in children 3–10 years old using these same eight European birth cohorts also found no association, or even decreased risk, between prenatal air pollution exposure and ADHD (Forns et al. 2018). Lastly, higher exposure to some PM components in pregnancy and childhood was also found to relate to fewer internalizing, externalizing, and attention problems on the CBCL in >5,000 adolescents (ages 13–16 years old) from the Generation R birth cohort (Kusters et al. 2022). The authors noted that similar associations were found regardless of whether the parent or youth reported these problems (Kusters et al. 2022). Thus, our current longitudinal findings are in line with these more recent studies based on multisites. Like these studies, it seems very unlikely that the significant effects found in the current study are reflective of a protective effect given both (1) the absence of any postulated mechanism for a protective element of air pollution exposure, as well as (2) the extremely small magnitude of change, detected in part to our large sample size and the resulting statistical power. It is feasible that both the previous findings as well as the current results could be due to residual negative confounding (Forns et al. 2018, Jorcano et al. 2019, Kusters et al. 2022), although it is important to note that in each case the analyses adjusted for many essential sociodemographic variables (i.e., income, caregiver educational attainment, etc.) that are known to be associated with air pollution exposure and mental health in children. Additional research is needed to decipher which unexplained factor(s) are likely to be at play and if they vary across various cities within Western populations (e.g., the United States, Germany, Italy, and Spain).
Developmental Context of Current Finding
Putting our current results in the larger context of the literature, the importance of window(s) of exposure and the timing of behavioral assessment continue to prevail as important factors to consider in determining what role air pollution may play in terms of risk for mental health problems. That is, while our current findings suggest that neither prenatal nor annual average of air pollution exposure is associated with more emotional problems at ages 9–10 years or increases in behavioral problems 1 year later, it is still feasible that air pollution exposure during development may contribute toward an individual’s risk for mental health problems later in adolescence or early adulthood. The development and progression of mental health disorders emerge over time (Kessler et al. 2005, 2007, Solmi et al. 2022). Thus, although air pollution–related changes in CBCL were not considered clinically meaningful, the current study is limited in that only small changes in emotional behavior in children are likely to be seen in the short timeframe studied (i.e., 1-year follow-up). As such, expanding upon the analyses presented in this report, we have expanded to determine if the expected negative association between PM2.5 and NO2 exposure at 9–10 years emerges over time by adding a third wave of CBCL data through ages 11–12 years; however, similar results were seen (Campbell et al. 2024). However, it is important to note that the peak incidence of psychopathology and psychiatric diagnoses seen in adolescence typically occurs in mid-adolescence, around age 14.5 years, whereas the median age of onset is 18 years (Solmi et al. 2022), which is above the upper limit of ages we have examined. Therefore, it is still feasible that exposure during the early adolescent period may ultimately predispose an individual to risk for developing psychopathology later in adolescence or early adulthood. As previously mentioned, this idea is supported by recent findings that higher levels of air pollution during childhood and adolescence predict later onset of major depressive disorder (Roberts et al. 2019) and other internalizing, externalizing, and thought disorder symptoms at age 18 years (Reuben et al. 2021). Moreover, given the effect modification of residential duration on PM2.5 and emotional behavior at ages 9–10 years, the results of the current study may suggest that while 1-year PM2.5 and NO2 exposure at 9–10 years does not meaningfully impact the relative risk of emotional problems, it is feasible that longer periods of exposure during development may have harmful effects. Lastly, even if associations do not exist at a population level, PM2.5 and NO2 exposure during this time may still have harmful effects in more susceptible children, due to either genetic risk or due to co-exposure to other adverse environmental threats. Thus, more research that takes a more integrated neural exposome approach to understanding adolescent environmental exposures and risk for psychopathology (Tamiz et al. 2022) is warranted.
Regardless if ambient air pollution exposure is related to overt emotional problems in early adolescence, it is clear from the results in Aim 1 of the current study, alongside the emerging literature (Pujol et al. 2016a, 2016b, Mortamais et al. 2017, Guxens et al. 2018, 2022, Mortamais et al. 2019, Lubczyńska et al. 2020, Burnor et al. 2021, Lubczyńska et al. 2021, Binter et al. 2022, Pérez-Crespo et al. 2022, Cotter et al. 2023, Sukumaran et al. 2023), that ambient air pollution may influence ongoing brain development and plasticity across adolescence. It is feasible that these detectable brain differences by MRI may be early neural biomarkers of exposure-related risk before any overt changes in behavior. The white matter tracts found to be linked to exposure in the current study reflect crucial structural brain connections between the prefrontal cortex and limbic brain regions involved in emotional processing and regulation (Catani and Thiebaut de Schotten 2008, Philippi et al. 2009, Coenen et al. 2012, Von Der Heide et al. 2013, Gabard-Durnam et al. 2014, Kamali et al. 2014). For example, the uncinate fasciculus connects the orbitofrontal cortex and amygdala, allowing for these brain regions to work together for memory and emotional processing, whereas the superior longitudinal fasciculus connects frontal, parietal, and opercular brain regions vital for perceiving and paying attention to emotional information (Catani and Thiebaut de Schotten 2008, Catani et al. 2013). Similarly, the cingulum white matter bundle connects the cingulate cortex, medial frontal cortex, and parahippocampal area to also process attention, emotion, and memory (Catani and Thiebaut de Schotten 2008, Catani et al. 2013). The other implicated white matter structures include the corpus callosum — the major connection between the two hemispheres of the brain — as well as the anterior thalamic radiation, which is a part of the larger prefrontal-cortico-thalamic network essential for processing external information and integration of that information in executing motor and behavioral actions in our everyday lives (Coenen et al. 2012). Neurological differences in these white matter tracts may, therefore, either reflect and/or contribute to altered behavior or cognitive control of emotions. In recent meta-analytic studies, widespread differences in both FA and MD outcomes have also been noted in many of these same white matter tracts among individuals diagnosed with various mental health disorders (Jenkins et al. 2016, Hinton et al. 2019, Vanes and Dolan 2021). Given these white matter alterations are shared among various emotional disorders, it has been proposed that differences in white matter microstructure may reflect a common brain phenotype or a transdiagnostic biological marker of mental health risk. For these reasons, it is possible that the associations between air pollution exposure and white matter findings in the current study may reflect an early brain biomarker of risk for mental health problems later in life. Active follow-up of ABCD study cohort participants through early adulthood will ultimately allow us to formally test this hypothesis.
LIMITATIONS
The research presented in this report is based on a large, diverse sample of children across the United States, utilized state-of-the-art exposure modeling to estimate residential air pollution exposure at the individual level with high spatiotemporal precision (1-km2 resolution), and collected harmonized brain imaging and emotional behavior from participants across research locations. The current study is likely to have good internal validity given the low percentage of missingness and the use of optimal statistical models, as confirmed by rigorous model checking. The current research also performed several sensitivity analyses, including using alternative adjustment sets for confounders, controlling for address duration, and accounting for prenatal exposure. Therefore, this study is highly valuable in assessing how current levels of ambient air pollution in the United States are associated with white matter microstructure and emotional problems in children. Nonetheless, several limitations of the current research need to be acknowledged.
The ABCD Study currently lacks air pollution exposure beyond the prenatal estimates and annual estimates that correspond with the 9–10-year age period. The current findings build upon previous studies showing that 1 year of exposure is associated with brain outcomes in late childhood in both the ABCD Study (Cserbik et al. 2020, Burnor et al. 2021) and the BRain dEvelopment and Air polluTion ultrafine particles in scHool childrEn (BREATHE) cohort in Spain (Pujol et al. 2016a, 2016b). In sensitivity analyses, we adjusted for and examined potential effect modification of exposure effects based on residential address length. These exploratory analyses suggest that longer time at the residential address may potentially contribute to stronger PM2.5 and NO2 effects on PFC structural connectivity at lower concentrations, as well as the hypothesized small, but positive associations between PM2.5 and more emotional problems. However, as previously mentioned, while residential duration may serve as a proxy for longer-term exposure to annualized PM2.5 at ages 9–10 years old, it does not accurately capture spatial and/or temporal differences in exposure over time, nor can we account for previous exposure levels for those individuals who lived at their residence for fewer years. Thus, key questions remain as to how exposures during earlier periods of life may have unique or cumulative effects over time, as well as how other periods of exposure may contribute to the current findings. Moreover, PM2.5 composition varies by geographical location and can include elemental carbon, sulfate, nitrate, ammonium, hydrogen ions, low- and moderate-volatility organic compounds, and different mixtures of metals (Bell et al. 2007). Analyses of how PM2.5 component makeup relates to the current findings will be especially useful in light of evidence that certain PM2.5 components may be more detrimental to health than others (Bell and Ebisu 2012). The current study was also limited in exposure assessment, as it did not include personal, home, or school-based measures of air pollution using real-time monitors. This exposure assessment limitation is especially important to note as children are likely to spend a substantial amount of time at locations other than their home address, such as school. Moreover, the current study was limited by assessing only outdoor ambient air quality and did not account for indoor air exposure, which might drastically influence personal exposure levels (Morawska et al. 2017). We also acknowledge the limitations of relying on address histories collected by retrospective recall. Moreover, only 66% of caregivers provided valid residential data that dated back to the child’s birth year, resulting in a smaller and less representative subsample of the ABCD study population to conduct sensitivity analyses adjusting for prenatal exposure.
Although a recent study found similar results (i.e., negative association between exposure and CBCL) regardless of whether the parent or youth reported on the CBCL (Kusters et al. 2022), the current study is limited by utilizing only a caregiver’s report of the child’s emotional problems. That is, informant differences (i.e., parent, youth, and teacher) have been noted in using the CBCL items to assess emotional and behavioral problems in youth and can vary depending on context (i.e., home versus school) (Achenbach et al. 1987). Therefore, it is feasible that caregiver reports of emotional problems in the current study may contribute to misclassification bias, especially for internalizing behaviors that may be less apparent to others. Additional studies are warranted that also incorporate additional reporters, as well as those that implement more objective measures of mental health, such as clinician-based interviews.
The cross-sectional outcome for white matter microstructure and two waves of emotional outcomes is also a limitation, especially since brain maturation is an ongoing developmental process (Fuhrmann et al. 2015, Lebel et al. 2019, Bethlehem et al. 2022) and mental health problems emerge during adolescence (Kessler et al. 2005, 2007, Solmi et al. 2022). Additional waves of MRI data and mental health assessments over longer periods of follow-up would give us additional insight into how ambient air pollution may influence trajectories of white matter microstructure as well as help us decipher if air pollution during development contributes to risk for mental health disorders later in life.
Confounding
The ABCD Study enrollment process was dynamically monitored to ensure the study met target sex, socioeconomic, ethnic, and racial diversity (Garavan et al. 2018). Moreover, although we had an overall low missingness of key variables for our main analysis, and loss-to-follow-up was minimal for the 1-year follow-up period, we cannot rule out other factors that might contribute to selection bias to participate in the ABCD Study. Moreover, participation was limited to the 21 study sites, which may contribute to ecological confounding effects. Although we aimed to be mindful of both choosing and adjusting for key confounders, we cannot rule out the possibility of residual confounding.
Spatial confounding has emerged as an important factor to consider in spatial statistics, albeit it is seldom defined and with conflicting definitions (Urdangarin et al. 2023). In the context of the current study, spatial confounding may bias the fixed effect estimates if there are unobserved variables with a spatial pattern. Alternatively, spatial confounding in the current study could also exist, given the potential collinearity between many of our address-based covariates and our random effect variable of study site location. In the current study we aimed to minimize spatial confounding by (1) implementing hybrid spatiotemporal modeling, which incorporates geospatially derived covariates, to estimate individual-level exposure based on the child’s residential address, (2) conducting sensitivity analyses with and without adjusting for a wide range of spatial covariates (i.e., urban classification, noise, population density, etc.), and (3) accounting for study site and other individual-level sociodemographic variables that may vary by location during statistical hypothesis testing. Nonetheless, residual, or unmeasured spatial confounding, may persist. Future studies can reduce this type of confounding by identifying and measuring additional potential confounders, using a method or study design that can further minimize this bias (i.e., quasi-experimental methods if appropriate), or considering other emerging methods proposed to further reduce potential spatial confounding (Urdangarin et al. 2023).
Generalizability
Although the ABCD Study sought to recruit a sample that mirrored the US population by design, the ABCD cohort is not a fully representative US cohort (Compton et al. 2019, Heeringa and Berglund 2020). The ABCD cohort has an underrepresentation of Asian, American Indian/Alaskan Native, and Native Hawaiian/Pacific Islander, as well as an overrepresentation of families with higher total household incomes (~64% with ≥$50,000) and those with highly educated caregivers (~59% with bachelor’s degree or higher). Moreover, our final analytic samples varied in some demographic characteristics from the overall ABCD study cohort. For example, our analytic samples were comprised of slightly higher proportions of non-Hispanic whites and those from higher socioeconomic status backgrounds as compared to the larger ABCD study cohort. Thus, it is important to note that the results presented here are generalizable to populations like our sample but may not be generalizable to all children in the United States. For instance, the burden of health effects related to outdoor air pollution exposure is not ubiquitous. Individuals who are poorer, have completed fewer years of education, and are from historically minoritized communities are exposed to higher levels of pollution within the United States (Hajat et al. 2015, Tessum et al. 2022), and these groups may be especially more susceptible to air pollution–related health threats because of compounded effects of disadvantage (i.e., psychosocial stress, access to resources, proximity to industrial or roadway emission sources, etc.) (Munoz-Pizza et al. 2020, Cortes-Ramirez et al. 2021). Additional studies are needed to assess the current findings within these populations underrepresented in the ABCD study cohort.
FUTURE DIRECTIONS
Given the novelty of the current results, replication of results in other cohorts is warranted. Moreover, the current study focused on MRI-based measures of white matter microstructure, but additional studies are necessary to implement a multimodal approach, or a combination of MRI methods, to more fully understand how air pollution exposure may impact the developing brain. Specifically, investigating how air pollution relates to structural, diffusion, and functional MRI metrics during development may help to elucidate which large-scale brain systems may be most vulnerable. Given that air pollution is a mixture of chemicals and gases that vary by sources and geographical locations, additional studies are also necessary to examine how PM components and sources of air pollution may contribute to the current findings.
IMPLICATIONS OF FINDINGS
The result from the current study suggests that current levels of exposure to PM2.5 and NO2 are associated with differences in white matter microstructure in early adolescence. We observed this association at levels of PM2.5 and NO2 above the 2021 World Health Organization guidelines but falling largely below the current EPA standards. While associations between NO2 and white matter microstructure were linear, associations were much stronger at PM2.5 levels above 7.7 μg/m3. Given that these levels of ambient air pollution are common in urban cities across the United States, policymakers should consider the neurological health effects in deriving current legislation standards.
DATA AVAILABILITY STATEMENT
Data used in the preparation of this article were obtained from the Adolescent Brain Cognitive Development (ABCD) Study (https://abcdstudy.org), held in the National Institute of Mental Health (NIMH) Data Archive (NDA). This is a multisite, longitudinal study designed to recruit more than 10,000 children aged 9–10 and follow them over 10 years into early adulthood. The ABCD Study is supported by the National Institutes of Health and additional federal partners under award numbers U01DA041048, U01DA050989, U01DA051016, U01DA041022, U01DA051018, U01DA051037, U01DA050987, U01DA041174, U01DA041106, U01DA041117, U01DA041028, U01DA041134, U01DA050988, U01DA051039, U01DA041156, U01DA041025, U01DA041120, U01DA051038, U01DA041148, U01DA041093, U01DA041089, U24DA041123, U24DA041147. A full list of supporters is available at https://abcdstudy.org/federal-partners.html. A listing of participating sites and a complete listing of the study investigators can be found at https://abcdstudy.org/consortium_members/. ABCD consortium investigators designed and implemented the study and/or provided data, but did not necessarily participate in the analysis or writing of this report. This manuscript reflects the views of the authors and may not reflect the opinions or views of the National Institutes of Health or ABCD consortium investigators.
The ABCD data repository grows and changes over time. The ABCD data used in this report came from NIMH Data Archive Digital Object Identifiers 10.15154/1523041 and 10.15154/8873-zj65.
ACKNOWLEDGMENTS
A special thank you to Kimberly Felix for helping with administrative tasks related to the current research project and Shermaine Abad, MPP, for conducting the geospatial linkages at the University of California, San Diego. Elizabeth Romero, MPH, and Gabrielle Major, MPH, also provided editorial assistance in preparing tables and text for the final report. Devyn Cotter, MA, created the white matter tract figures presented in the final report.
The ABCD Study is conducted by the Children’s Hospital of Los Angeles, University of Colorado Boulder, Florida International University, Laureate Institute for Brain Research, Medical University of South Carolina, Oregon Health and Science University, University of Rochester, SRI International, University of California Los Angeles, University of California San Diego, University of Florida, University of Maryland Baltimore, University of Michigan, University of Minnesota, University of Pittsburgh, University of Utah, University of Vermont, University of Wisconsin-Milwaukee, Virginia Commonwealth University, Washington University in St. Louis, and Yale. We gratefully acknowledge the contribution of the children and their caregivers for their participation. A special acknowledgment to Dr. Wes Thompson, Dr. Chun Fan, and Shermaine Abad for assisting with overseeing the linkage of exposure data to the residential addresses at the ABCD Data Analysis and Informatics Resource Center (DAIRC) at the University of California, San Diego.
Additional support for this work was made possible by NIEHS R01-ES032295 and R01-ES031074.
Footnotes
* A list of abbreviations and other terms appears at the end of this volume.
HEI QUALITY ASSURANCE STATEMENT
The conduct of this study was subjected to an independent audit by Westat staff members, including experienced quality assurance (QA) auditors with expertise in statistical modeling, epidemiology, exposure assessment, and geographic information systems analysis. The Westat QA audit team consisted of Mr. Michael Giangrande and Ms. Rebecca Birch. These staff members are highly experienced in quality assurance oversight across various relevant domains. The QA oversight program consisted of a remote audit of the final report and the data processing steps. Key details of the dates of the audit and the reviews performed are listed below.
Date: June 2024 – September 2024
Remarks: The Herting et al. study underwent an independent quality assurance (QA) audit by two Westat auditors with quality assurance oversight experience and expertise relevant to exposure assessment, air quality monitoring and modeling, epidemiological methods, geospatial analysis, and statistical analysis. The Westat QA audit of the final Herting et al. report focused on adherence to the study protocol, appropriateness of the documentation of the study methods (e.g., data processing, exposure modeling, and statistical modeling), whether study assumptions and limitations were adequately addressed, and whether the investigators’ conclusions were reasonable given the study findings and in consideration of the limitations. The QA team also evaluated whether the report was easy to understand and performed an editorial review.
The Westat QA audit team provided a written report to HEI and the study investigators. The Westat QA auditors concluded that the study was well conducted in accordance with the study protocol and that the report was well written. The auditors also provide HEI and the investigator team with specific recommendations for improvement. Areas of QA feedback included clarification of variable derivation, providing visualization of monitoring sites, and editorial comments and suggestions.
Herting et al. responded to the QA recommendations and incorporated the feedback from the QA auditors in a final report that HEI provided to Westat. The Westat QA audit team attests that the final report appears to be representative of the study conducted.
Given record-level human subjects data used in the current project, R code is available from the authors for individuals with the necessary data use certificate (DUC) from the NIMH Data Archive (NDA) for access to the ABCD shared data.
ABOUT THE AUTHORS
Megan M. Herting is an associate professor and Director of the Herting Neuroimaging Laboratory at the Department of Population and Public Health Sciences, University of Southern California, Los Angeles, California, USA. She received her PhD from Oregon Health & Science University, Portland, Oregon, USA. She is the Principal Investigator with over 15 years of experience in behavioral neuroscience and developmental neuroimaging. She is a co-principal investigator of the Children’s Hospital Los Angeles’ research site and an affiliate of the ABCD Study Consortium. Her research interests include brain and cognitive development in healthy and at-risk populations, including several ongoing investigations in children, adolescents, and young adults.
Elisabeth Burnor is a public health epidemiologist. She received her master’s degree with a focus on Environmental Health Sciences from the University of Washington, Seattle, Washington, USA. Her role in this project was data management and statistical analyses for Aim 1 as well as preparation of reports during her time as a research coordinator in the Herting Neuroimaging Laboratory within the Department of Population and Public Health Sciences, University of Southern California, Los Angeles, California, USA. Her research interests include environmental epidemiology, infectious disease epidemiology, and public health equity.
Hedyeh Ahmadi is a biostatistician in the Herting Neuroimaging Laboratory at the Department of Population and Public Health Sciences, University of Southern California, Los Angeles, California, USA. She received her PhD in Measurement, Evaluation, and Statistics from Columbia University, New York City, New York, USA, and her MS in Statistics from the University of California, Irvine, California, USA. As a statistical consultant, her research interests include but are not limited to analyzing and managing salivary and medical data, education/psychology data, air pollution data, and neuroimaging data, as well as performing power analysis for cross-sectional and longitudinal studies in regression and structural equation modeling framework. Her role in this project was to oversee longitudinal data management, conduct analyses for Aim 2, consult on analysis for Aim 1, and prepare the final report.
Jim Gauderman is a professor and the chair of the Biostatistics Division in the Department of Population and Public Health Sciences, University of Southern California, Los Angeles, California, USA. He received his master’s in science from the University of Southern California in 1988 and his PhD from the University of Southern California in 1992. His research interests included both developing statistical methods for genetic-epidemiological analysis of pedigree data, as well as the design and analysis of studies relating health outcomes to environmental exposures. His role in the current project was as one of the two lead biostatisticians, in which he provided expertise in statistical methods development and applied statistical analyses to the research team.
Sandra P. Eckel is an associate professor in the Division of Biostatistics in the Department of Population and Public Health Sciences, University of Southern California, Los Angeles, California, USA. She received her PhD from Johns Hopkins University, Baltimore, Maryland, USA. Her role in this project was one of the two lead biostatisticians to provide expertise in statistical methods development and applied statistical analyses in environmental epidemiology. Her research interests are in statistical methods and applications in environmental epidemiology with a special focus on child outcomes.
Joel Schwartz is a professor in the Department of Epidemiology, Harvard University, Cambridge, Massachusetts, USA. He received his PhD from Brandeis University, Waltham, Massachusetts, USA. His research interests include modeling exposure to air pollutants using multiple methods, including machine learning, fusing data from satellite remote sensing, chemical transport models, land use data, and meteorology. His role in this project was to provide neural network estimates of PM2.5 and NO2 exposure on a 1×1-km grid across the United States. He also provided expertise in analyzing and interpreting findings in the larger context of the field of environmental epidemiology.
Kiros T. Berhane is a professor and chair in the Department of Biostatistics, Columbia Mailman School of Public Health, New York City, New York, USA. He received his MS from the University of Guelph, Ontario, Canada, and his PhD from the University of Toronto, Ontario, Canada. His research interests focus on the development of statistical methods for complex and correlated data structures, as well as their application in a wide range of public health topics, with a special focus on the health impacts of environmental factors. His role as mentor for the current project was to provide expertise in the development of statistical methods for environmental research and their application to examine the effects of air pollution on children’s health outcomes.
Rob S. McConnell is a professor, the director of the Southern California Environmental Health Sciences Center, and the chair of the Division of Environmental Health in the Department of Population and Public Health Sciences, University of Southern California, Los Angeles, California, USA. He received his MD from the University of California, San Francisco, California, USA. His research interests are focused on the relationship of air pollution to obesity and its metabolic consequences, and the role of air pollution’s neurotoxic effects as it pertains to autism. His role as a mentor in the project was to provide expertise on the effects of air pollution on neurodevelopment using epidemiological datasets.
Jiu-Chiuan Chen is an associate professor in the Department of Population and Public Health Sciences with a secondary adjunct position in the Department of Radiology, School of Medicine, University of Southern California, Los Angeles, California, USA. He received his MD from Taipei Medical University, Taipei, Taiwan, and his ScD from Harvard University School of Public Health, Boston, Massachusetts, USA. His research endeavors focus on investigating whether and how exposures to neurotoxic air pollutants and other physical characteristics of outdoor environments (e.g., ambient air temperature and neighborhood greenspace) affect neurobehavioral development and brain aging. His role as a mentor for the current project was to provide expertise on translational environmental health research and multidimensional longitudinal data analyses.
OTHER PUBLICATIONS RESULTING FROM THIS RESEARCH
Burnor E, Cserbik D, Cotter DL, Palmer CE, Ahmadi H, Eckel SP, et al. 2021. Association of outdoor ambient fine particulate matter with intracellular white matter microstructural properties among children. JAMA Netw Open 4:e2138300; doi:10.1001/jamanetworkopen.2021.38300.
Campbell CE, Cotter DL, Bottenhorn KL, Burnor E, Ahmadi H, Gauderman WJ, et al. 2024. Air pollution and emotional behavior in adolescents across the US. J Environ Res 240:117390; doi:10.1016/j.envres.2023.117390.
REFERENCES
Abid Z, Roy A, Herbstman JB, Ettinger AS. 2014. Urinary polycyclic aromatic hydrocarbon metabolites and attention/deficit hyperactivity disorder, learning disability, and special education in US children aged 6 to 15. J Environ Public Health
2014:628508; doi:10.1155/2014/628508.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Achenbach TM, McConaughy SH, Howell CT. 1987. Child/adolescent behavioral and emotional problems: Implications of cross-informant correlations for situational specificity. Psychol Bull
101(2):213–232.
[PubMed] [Google Scholar]
Achenbach TM, Rescorla LA. 2001. Manual for the ASEBA School-Age Forms and Profiles. Burlington, VT, USA, Research Center for Children, Youth and Families, University of Vermont. [Google Scholar]
Ahmed SM, Mishra GD, Moss KM, Yang IA, Lycett K, Knibbs LD. 2022. Maternal and childhood ambient air pollution exposure and mental health symptoms and psychomotor development in children: An Australian population-based longitudinal study. Environ Int
158:107003; doi:10.1016/j.envint.2021.107003.
[DOI] [PubMed] [Google Scholar]
Allen JL, Liu X, Weston D, Prince L, Oberdörster G, Finkelstein JN, et al. 2014. Developmental exposure to concentrated ambient ultrafine particulate matter air pollution in mice results in persistent and sex-dependent behavioral neurotoxicity and glial activation. Toxicol Sci
140(1):160–178; doi:10.1093/toxsci/kfu059.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Auchter AM, Hernandez Mejia M, Heyser CJ, Shilling PD, Jernigan TL, Brown SA, et al. 2018. A description of the ABCD organizational structure and communication framework. Dev Cogn Neurosci
32:8–15; doi:10.1016/j.dcn.2018.04.003.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Babadjouni R, Patel A, Liu Q, Shkirkova K, Lamorie-Foote K, Connor M, et al. 2018. Nanoparticulate matter exposure results in neuroinflammatory changes in the corpus callosum. PLoS One
13(11):e0206934; doi:10.1371/journal.pone.0206934.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Bansal E, Hsu HH, de Water E, Martínez-Medina S, Schnaas L, Just AC, et al. 2021. Prenatal PM2.5 exposure in the second and third trimesters predicts neurocognitive performance at age 9–10 years: A cohort study of Mexico City children. Environ Res
202:111651; doi:10.1016/j.envres.2021.111651.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Barch DM, Albaugh MD, Avenevoli S, Chang L, Clark DB, Glantz MD, et al. 2018. Demographic, physical and mental health assessments in the adolescent brain and cognitive development study: Rationale and description. Dev Cogn Neurosci
32:55–66; doi:10.1016/j.dcn.2017.10.010.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Baune BT, Fuhr M, Air T, Hering C. 2014. Neuropsychological functioning in adolescents and young adults with major depressive disorder: A review. Psychiatry Res
218(3):261–271; doi:10.1016/j.psychres.2014.04.052.
[DOI] [PubMed] [Google Scholar]
Bava S, Thayer R, Jacobus J, Ward M, Jernigan TL, Tapert SF. 2010. Longitudinal characterization of white matter maturation during adolescence. Brain Res
1327:38–46; doi:10.1016/j.brainres.2010.02.066.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Beaulieu C. 2002. The basis of anisotropic water diffusion in the nervous system - a technical review. NMR Biomed
15(7-8):435–455; doi:10.1002/nbm.782.
[DOI] [PubMed] [Google Scholar]
Bell ML, Dominici F, Ebisu K, Zeger SL, Samet JM. 2007. Spatial and temporal variation in PM(2.5) chemical composition in the United States for health effects studies. Environ Health Perspect
115(7):989–995; doi:10.1289/ehp.9621.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Bell ML, Ebisu K. 2012. Environmental inequality in exposures to airborne particulate matter components in the United States. Environ Health Perspect
120(12):1699–1704; doi:10.1289/ehp.1205201.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Benes FM. 1989. Myelination of cortical-hippocampal relays during late adolescence. Schizophr Bull
15(4):585–593; doi:10.1093/schbul/15.4.585.
[DOI] [PubMed] [Google Scholar]
Bethlehem RAI, Seidlitz J, White SR, Vogel JW, Anderson KM, Adamson C, et al. 2022. Brain charts for the human lifespan. Nature
604(7906):525–533; doi:10.1038/s41586-022-04554-y.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Binter AC, Kusters MSW, van den Dries MA, Alonso L, Lubczyńska MJ, Hoek G, et al. 2022. Air pollution, white matter microstructure, and brain volumes: Periods of susceptibility from pregnancy to preadolescence. Environ Pollut
313:120109; doi:10.1016/j.envpol.2022.120109.
[DOI] [PubMed] [Google Scholar]
Block ML, Elder A, Auten RL, Bilbo SD, Chen H, Chen JC, et al. 2012. The outdoor air pollution and brain health workshop. Neurotoxicology
33(5):972–984; doi:10.1016/j.neuro.2012.08.014.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Bø R, Billieux J, Gjerde LC, Eilertsen ME, Landrø NI. 2017. Do executive functions predict binge-drinking patterns? Evidence from a longitudinal study in young adulthood. Front Psychol
8:489; doi:10.3389/fpsyg.2017.00489.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Boda E, Rigamonti AE, Bollati V. 2020. Understanding the effects of air pollution on neurogenesis and gliogenesis in the growing and adult brain. Curr Opin Pharmacol
50:61–66; doi:10.1016/j.coph.2019.12.003.
[DOI] [PubMed] [Google Scholar]
Bonekamp D, Nagae LM, Degaonkar M, Matson M, Abdalla WMA, Barker PB, et al. 2007. Diffusion tensor imaging in children and adolescents: Reproducibility, hemispheric, and age-related differences. NeuroImage
34(2):733–742; doi:10.1016/j.neuroimage.2006.09.020.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Brockmeyer S, D’Angiulli A. 2016. How air pollution alters brain development: The role of neuroinflammation. Transl Neurosci
7(1):24–30; doi:10.1515/tnsci-2016-0005.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Büchel C, Raedler T, Sommer M, Sach M, Weiller C, Koch MA. 2004. White matter asymmetry in the human brain: A diffusion tensor MRI study. Cereb Cortex
14(9):945–951; doi:10.1093/cercor/bhh055. Epub 2004 Apr 27.
[DOI] [PubMed] [Google Scholar]
Burnor E, Cserbik D, Cotter DL, Palmer CE, Ahmadi H, Eckel SP, et al. 2021. Association of Outdoor Ambient Fine Particulate Matter With Intracellular White Matter Microstructural Properties Among Children. JAMA Netw Open
4(12):e2138300; doi:10.1001/jamanetworkopen.2021.38300.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Butter ME. 2006. Are women more vulnerable to environmental pollution?
J Hum Ecol
20:221–226; doi:10.1080/09709274.2006.11905931. [Google Scholar]
Campbell CE, Cotter DL, Bottenhorn KL, Burnor E, Ahmadi H, Gauderman WJ, et al. 2024. Air pollution and emotional behavior in adolescents across the US. J Environ Res
240:117390; doi:10.1016/j.envres.2023.117390. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cardinal RN, Parkinson JA, Hall J, Everitt BJ. 2002. Emotion and motivation: The role of the amygdala, ventral striatum, and prefrontal cortex. Neurosci Biobehav Rev
26(3):321–352; doi:10.1016/s0149-7634(02)00007-6.
[DOI] [PubMed] [Google Scholar]
Carlson M, Eldreth CD, Chuang Y-F, Eaton WW. 2012. Mental disorders across the life span and the role of executive function networks. In: Public Mental Health (Eaton WW, ed). New York: Oxford University Press, 245–268. [Google Scholar]
Casey BJ. 2015. Beyond simple models of self-control to circuit-based accounts of adolescent behavior. Annu Rev Psychol
66:295–319; doi:10.1146/annurev-psych-010814-015156.
[DOI] [PubMed] [Google Scholar]
Casey BJ, Cannonier T, Conley MI, Cohen AO, Barch DM, Heitzeg MM, et al. 2018. The Adolescent Brain Cognitive Development (ABCD) study: Imaging acquisition across 21 sites. Dev Cogn Neurosci
32:43–54; doi:10.1016/j.dcn.2018.03.001. Epub 2018 Mar 14.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Castagna A, Mascheroni E, Fustinoni S, Montirosso R. 2022. Air pollution and neurodevelopmental skills in preschool- and school-aged children: A systematic review. Neurosci Biobehav Rev
136:104623; doi:10.1016/j.neubiorev.2022.104623.
[DOI] [PubMed] [Google Scholar]
Catani M, Dell’acqua M, Thiebaut de Schotten M. 2013. A revised limbic system model for memory, emotion and behaviour. Neurosci Biobehav Rev
37(8):1724–1737; doi:10.1016/j.neubiorev.2013.07.001.
[DOI] [PubMed] [Google Scholar]
Catani M, Thiebaut de Schotten M. 2008. A diffusion tensor imaging tractography atlas for virtual in vivo dissections. Cortex
44(8):1105–1132; doi:10.1016/j.cortex.2008.05.004.
[DOI] [PubMed] [Google Scholar]
CIESIN (Center for International Earth Science Information Network). 2016. Gridded Population of the World, Version 4 (GPWv4): Population Density Adjusted to Match 2015 Revision UN WPP Country Totals. Columbia University, Palisades, NY, NASA Socioeconomic Data and Applications Center (SEDAC). [Google Scholar]
Clifford A, Lang L, Chen R, Anstey KJ, Seaton A. 2016. Exposure to air pollution and cognitive functioning across the life course--A systematic literature review. Environ Res
147:383–398; doi:10.1016/j.envres.2016.01.018.
[DOI] [PubMed] [Google Scholar]
Coenen VA, Panksepp J, Hurwitz TA, Urbach H, Mädler B. 2012. Human medial forebrain bundle (MFB) and anterior thalamic radiation (ATR): Imaging of two major subcortical pathways and the dynamic balance of opposite affects in understanding depression. J Neuropsychiatry Clin Neurosci
24(2):223–236; doi:10.1176/appi.neuropsych.11080180.
[DOI] [PubMed] [Google Scholar]
Cohen AJ, Brauer M, Burnett R, Anderson HR, Frostad J, Estep K, et al. 2017. Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: An analysis of data from the Global Burden of Diseases Study 2015. Lancet
389(10082):1907–1918; doi:10.1016/S0140-6736(17)30505-6.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Compton WM, Dowling GJ, Garavan H. 2019. Ensuring the best use of data: The adolescent brain cognitive development study. JAMA Pediatr
173(9):809–810; doi:10.1001/jamapediatrics.2019.2081.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Cortes-Ramirez J, Wilches-Vega JD, Paris-Pineda OM, Rod JE, Ayurzana L, Sly PD. 2021. Environmental risk factors associated with respiratory diseases in children with socioeconomic disadvantage. Heliyon
7(4):e06820; doi:10.1016/j.heliyon.2021.e06820.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Cory-Slechta DA, Merrill A, Sobolewski M. 2023. Air pollution–related neurotoxicity across the life span. Annu Rev Pharmacol Toxicol
63(1):143–163; doi:10.1146/annurev-pharmtox-051921-020812.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Cory-Slechta DA, Sobolewski M. 2023. Neurotoxic effects of air pollution: An urgent public health concern. Nat Rev Neurosci
24(3):129–130; doi:10.1038/s41583-022-00672-8.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Cotter DL, Campbell CE, Sukumaran K, McConnell R, Berhane K, Schwartz J, et al. 2023. Effects of ambient fine particulates, nitrogen dioxide, and ozone on maturation of functional brain networks across early adolescence. Environ Int
177:108001; doi:10.1016/j.envint.2023.108001.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Cserbik D, Chen JC, McConnell R, Berhane K, Sowell ER, Schwartz J, et al. 2020. Fine particulate matter exposure during childhood relates to hemispheric-specific differences in brain structure. Environ Int
143:105933; doi:10.1016/j.envint.2020.105933.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Cullen KR, Klimes-Dougan B, Muetzel R, Mueller BA, Camchong J, Houri A, et al. 2010. Altered white matter microstructure in adolescents with major depression: A preliminary study. J Am Acad Child Adolesc Psychiatry
49(2):173–183.e171; doi:10.1097/00004583-201002000-00011.
[DOI] [PMC free article] [PubMed] [Google Scholar]
de Prado Bert P, Mercader EMH, Pujol J, Sunyer J, Mortamais M. 2018. The Effects of Air Pollution on the Brain: A Review of Studies Interfacing Environmental Epidemiology and Neuroimaging. Curr Environ Health Rep
5(3):351–364; doi:10.1007/s40572-018-0209-9.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Di Q, Amini H, Shi L, Kloog I, Silvern R, Kelly J, et al. 2019. An ensemble-based model of PM2.5 concentration across the contiguous United States with high spatiotemporal resolution. Environ Int
130:104909; doi:10.1016/j.envint.2019.104909.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Di Q, Amini H, Shi L, Kloog I, Silvern R, Kelly J, et al. 2020. Assessing NO2 concentration and model uncertainty with high spatiotemporal resolution across the contiguous United States using ensemble model averaging. Environ Sci Technol
54(3):1372–1384; doi:10.1021/acs.est.9b03358.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Di Q, Kloog I, Koutrakis P, Lyapustin A, Wang Y, Schwartz J. 2016. Assessing PM2.5 exposures with high spatiotemporal resolution across the continental United States. Environ Sci Technol
50(9):4712–4721; doi:10.1021/acs.est.5b06121.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Dominici F, Schwartz J, Di Q, Braun D, Choirat C, Zanobetti A. 2019. Assessing adverse health effects of long-term exposure to low levels of ambient air pollution: Phase 1. Res Rep Health Eff Inst (200):1–51. [PMC free article] [PubMed] [Google Scholar]
Dong Y, Peng CY. 2013. Principled missing data methods for researchers. Springerplus
2(1):222; doi:10.1186/2193-1801-2-222.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Dores AK, Fick GH, MacMaster FP, Williams JVA, Bulloch AGM, Patten SB. 2021. Outdoor air pollution and depression in Canada: A population-based cross-sectional study from 2011 to 2016. Int J Environ Res Public Health
18(5); doi:10.3390/ijerph18052450. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dosenbach NU, Nardos B, Cohen AL, Fair DA, Power JD, Church JA, et al. 2010. Prediction of individual brain maturity using fMRI. Science
329(5997):1358–1361; doi:10.1126/science.1194144.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Echeverria SE, Diez-Roux AV, Link BG. 2004. Reliability of self-reported neighborhood characteristics. J Urban Health
81(4):682–701; doi:10.1093/jurban/jth151.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Fair DA, Cohen AL, Dosenbach NU, Church JA, Miezin FM, Barch DM, et al. 2008. The maturing architecture of the brain’s default network. Proc Natl Acad Sci U S A
105(10):4028–4032; doi:10.1073/pnas.0800376105.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Fair DA, Cohen AL, Power JD, Dosenbach NU, Church JA, Miezin FM, et al. 2009. Functional brain networks develop from a “local to distributed” organization. PLoS Comput Biol
5(5):e1000381; doi:10.1371/journal.pcbi.1000381.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Fair DA, Dosenbach NU, Church JA, Cohen AL, Brahmbhatt S, Miezin FM, et al. 2007. Development of distinct control networks through segregation and integration. Proc Natl Acad Sci U S A
104(33):13507–13512; doi:10.1073/pnas.0705843104.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Fan B, Wang T, Wang W, Zhang S, Gong M, Li W, et al. 2019. Long-term exposure to ambient fine particulate pollution, sleep disturbance and their interaction effects on suicide attempts among Chinese adolescents. J Affect Disord
258:89–95; doi:10.1016/j.jad.2019.08.004.
[DOI] [PubMed] [Google Scholar]
Fan CC, Marshall A, Smolker H, Gonzalez MR, Tapert SF, Barch DM, et al. 2021. Adolescent Brain Cognitive Development (ABCD) study Linked External Data (LED): Protocol and practices for geocoding and assignment of environmental data. Dev Cogn Neurosci
52:101030; doi:10.1016/j.dcn.2021.101030.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Fan SJ, Heinrich J, Bloom MS, Zhao TY, Shi TX, Feng WR, et al. 2020. Ambient air pollution and depression: A systematic review with meta-analysis up to 2019. Sci Total Environ
701:134721; doi:10.1016/j.scitotenv.2019.134721.
[DOI] [PubMed] [Google Scholar]
Feldstein Ewing SW, Chang L, Cottler LB, Tapert SF, Dowling GJ, Brown SA. 2018. Approaching retention within the ABCD study. Dev Cogn Neurosci
32:130–137; doi:10.1016/j.dcn.2017.11.004.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Fonken LK, Xu X, Weil ZM, Chen G, Sun Q, Rajagopalan S, et al. 2011. Air pollution impairs cognition, provokes depressive-like behaviors and alters hippocampal cytokine expression and morphology. Mol Psychiatry
16(10):987–995, 973; doi:10.1038/mp.2011.76.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Fordyce TA, Leonhard MJ, Chang ET. 2018. A critical review of developmental exposure to particulate matter, autism spectrum disorder, and attention deficit hyperactivity disorder. J Environ Sci Health A Tox Hazard Subst Environ Eng
53(2):174–204; doi:10.1080/10934529.2017.1383121.
[DOI] [PubMed] [Google Scholar]
Forns J, Dadvand P, Foraster M, Alvarez-Pedrerol M, Rivas I, López-Vicente M, et al. 2016. Traffic-related air pollution, noise at school, and behavioral problems in Barcelona schoolchildren: A cross-sectional Study. Environ Health Perspect
124(4):529–535; doi:10.1289/ehp.1409449.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Forns J, Sunyer J, Garcia-Esteban R, Porta D, Ghassabian A, Giorgis-Allemand L, et al. 2018. Air pollution exposure during pregnancy and symptoms of attention deficit and hyperactivity disorder in children in Europe. Epidemiology
29(5):618–626; doi:10.1097/EDE.0000000000000874.
[DOI] [PubMed] [Google Scholar]
Fuhrmann D, Knoll LJ, Blakemore SJ. 2015. Adolescence as a sensitive period of brain development. Trends Cogn Sci
19(10):558–566; doi:10.1016/j.tics.2015.07.008.
[DOI] [PubMed] [Google Scholar]
Gabard-Durnam LJ, Flannery J, Goff B, Gee DG, Humphreys KL, Telzer E, et al. 2014. The development of human amygdala functional connectivity at rest from 4 to 23 years: A cross-sectional study. Neuroimage
95:193–207; doi:10.1016/j.neuroimage.2014.03.038.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Garavan H, Bartsch H, Conway K, Decastro A, Goldstein RZ, Heeringa S, et al. 2018. Recruiting the ABCD sample: Design considerations and procedures. Dev Cogn Neurosci
32:16–22; doi:10.1016/j.dcn.2018.04.004.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Gee DG, Humphreys KL, Flannery J, Goff B, Telzer EH, Shapiro M, et al. 2013. A developmental shift from positive to negative connectivity in human amygdala-prefrontal circuitry. J Neurosci
33(10):4584–4593; doi:10.1523/JNEUROSCI.3446-12.2013.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Genc S, Zadeoglulari Z, Fuss SH, Genc K. 2012. The adverse effects of air pollution on the nervous system. J Toxicol
2012:782462; doi:10.1155/2012/782462.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Ghashghaei HT, Hilgetag CC, Barbas H. 2007. Sequence of information processing for emotions based on the anatomic dialogue between prefrontal cortex and amygdala. Neuroimage
34(3):905–923; doi:10.1016/j.neuroimage.2006.09.046.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Giedd JN, Rapoport JL. 2010. Structural MRI of pediatric brain development: What have we learned and where are we going?
Neuron
67(5):728–734; doi:10.1016/j.neuron.2010.08.040.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Giedd JN, Snell JW, Lange N, Rajapakse JC, Casey BJ, Kozuch PL, et al. 1996a. Quantitative magnetic resonance imaging of human brain development: Ages 4-18. Cereb Cortex
6(4):551–560; doi:10.1093/cercor/6.4.551.
[DOI] [PubMed] [Google Scholar]
Giedd JN, Vaituzis AC, Hamburger SC, Lange N, Rajapakse JC, Kaysen D, et al. 1996b. Quantitative MRI of the temporal lobe, amygdala, and hippocampus in normal human development: Ages 4–18 years. J Comp Neurol
366(2):223–230; doi:10.1002/(SICI)1096-9861(19960304)366:2<223::AID-CNE3>3.0.CO;2-7.
[DOI] [PubMed] [Google Scholar]
Gogtay N, Nugent TF, 3rd, Herman DH, Ordonez A, Greenstein D, Hayashi KM, et al. 2006. Dynamic mapping of normal human hippocampal development. Hippocampus
16(8):664–672; doi:10.1002/hipo.20193.
[DOI] [PubMed] [Google Scholar]
Goldman-Rakic PS. 1987. Development of cortical circuitry and cognitive function. Child Dev
58(3):601–622.
[PubMed] [Google Scholar]
Gomez-Budia M, Konttinen H, Saveleva H, Korhonen P, Jalava PI, Kanninen KM, et al. 2020. Glial smog: Interplay between air pollution and astrocyte-microglia interactions. Neurochem Int
136:104715; doi:10.1016/j.neuint.2020.104715.
[DOI] [PubMed] [Google Scholar]
Guxens M, Lubczyńska MJ, Muetzel RL, Dalmau-Bueno A, Jaddoe VWV, Hoek G, et al. 2018. Air pollution exposure during fetal life, brain morphology, and cognitive function in school-age children. Biol Psychiatry
84(4):295–303; doi:10.1016/j.biopsych.2018.01.016.
[DOI] [PubMed] [Google Scholar]
Guxens M, Lubczyńska MJ, Pérez-Crespo L, Muetzel RL, El Marroun H, Basagana X, et al. 2022. Associations of air pollution on the brain in children: A brain imaging study. Res Rep Health Eff Inst(209):1–61. [PMC free article] [PubMed] [Google Scholar]
Hagler DJ, Jr, Ahmadi ME, Kuperman J, Holland D, McDonald CR, Halgren E, et al. 2009. Automated white-matter tractography using a probabilistic diffusion tensor atlas: Application to temporal lobe epilepsy. Hum Brain Mapp
30(5):1535–1547; doi:10.1002/hbm.20619.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Hagler DJ, Jr, Hatton S, Cornejo MD, Makowski C, Fair DA, Dick AS, et al. 2019. Image processing and analysis methods for the Adolescent Brain Cognitive Development Study. Neuroimage
202:116091; doi:10.1016/j.neuroimage.2019.116091.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Hajat A, Hsia C, O’Neill MS. 2015. Socioeconomic disparities and air pollution exposure: A global review. Curr Environ Health Rep
2(4):440–450; doi:10.1007/s40572-015-0069-5.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Hare TA, Tottenham N, Galvan A, Voss HU, Glover GH, Casey BJ. 2008. Biological substrates of emotional reactivity and regulation in adolescence during an emotional go-nogo task. Biol Psychiatry
63(10):927–934; doi:10.1016/j.biopsych.2008.03.015.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Hariri AR, Bookheimer SY, Mazziotta JC. 2000. Modulating emotional responses: Effects of a neocortical network on the limbic system. Neuroreport
11(1):43–48; doi:10.1097/00001756-200001170-00009.
[DOI] [PubMed] [Google Scholar]
Harris MH, Gold DR, Rifas-Shiman SL, Melly SJ, Zanobetti A, Coull BA, et al. 2016. Prenatal and childhood traffic-related air pollution exposure and childhood executive function and behavior. Neurotoxicol Teratol
57:60–70; doi:10.1016/j.ntt.2016.06.008.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Harris PA, Taylor R, Minor BL, Elliott V, Fernandez M, O’Neal L, et al. 2019. The REDCap consortium: Building an international community of software platform partners. J Biomed Inform
95:103208; doi:10.1016/j.jbi.2019.103208.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. 2009. Research electronic data capture (REDCap): A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform
42(2):377–381; doi:10.1016/j.jbi.2008.08.010.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Heller AS, Casey BJ. 2016. The neurodynamics of emotion: Delineating typical and atypical emotional processes during adolescence. Dev Sci
19(1):3–18; doi:10.1111/desc.12373.
[DOI] [PubMed] [Google Scholar]
Hernan MA, Hernandez-Diaz S, Werler MM, Mitchell A. 2002. Causal knowledge as a prerequisite for confounding evaluation: An application to birth defects epidemiology. Am J Epidemiol
155(2):176–184; doi:10.1093/aje/155.2.176.
[DOI] [PubMed] [Google Scholar]
Herting MM, Johnson C, Mills KL, Vijayakumar N, Dennison M, Liu C, et al. 2018. Development of subcortical volumes across adolescence in males and females: A multisample study of longitudinal changes. Neuroimage
172:194–205; doi:10.1016/j.neuroimage.2018.01.020.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Herting MM, Younan D, Campbell CE, Chen JC. 2019. Outdoor air pollution and brain structure and function from across childhood to young adulthood: A methodological review of brain MRI studies. Front Public Health
7:332; doi:10.3389/fpubh.2019.00332.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Hinton KE, Lahey BB, Villalta-Gil V, Meyer FAC, Burgess LL, Chodes LK, et al. 2019. White matter microstructure correlates of general and specific second-order factors of psychopathology. Neuroimage Clin
22:101705; doi:10.1016/j.nicl.2019.101705.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Huang-Pollock C, Shapiro Z, Galloway-Long H, Weigard A. 2017. Is poor working memory a transdiagnostic risk factor for psychopathology?
J Abnorm Child Psychol
45(8):1477–1490; doi:10.1007/s10802-016-0219-8.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Huttenlocher PR. 1990. Morphometric study of human cerebral cortex development. Neuropsychologia
28(6):517–527; doi:10.1016/0028-3932(90)90031-i.
[DOI] [PubMed] [Google Scholar]
Ibrahim JG, Molenberghs G. 2009. Missing data methods in longitudinal studies: A review. Test (Madr)
18(1):1–43; doi:10.1007/s11749-009-0138-x.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Jedrychowski WA, Perera FP, Camann D, Spengler J, Butscher M, Mroz E, et al. 2015. Prenatal exposure to polycyclic aromatic hydrocarbons and cognitive dysfunction in children. Environ Sci Pollut Res Int
22(5):3631–3639; doi:10.1007/s11356-014-3627-8.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Jenkins LM, Barba A, Campbell M, Lamar M, Shankman SA, Leow AD, et al. 2016. Shared white matter alterations across emotional disorders: A voxel-based meta-analysis of fractional anisotropy. Neuroimage Clin
12:1022–1034; doi:10.1016/j.nicl.2016.09.001.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Jorcano A, Lubczyńska MJ, Pierotti L, Altug H, Ballester H, Cesaroni G, et al. 2019. Prenatal and postnatal exposure to air pollution and emotional and aggressive symptoms in children from 8 European birth cohorts. Environ Int
131:104927; doi:10.1016/j.envint.2019.104927.
[DOI] [PubMed] [Google Scholar]
Kamali A, Flanders AE, Brody J, Hunter JV, Hasan KM. 2014. Tracing superior longitudinal fasciculus connectivity in the human brain using high-resolution diffusion tensor tractography. Brain Struct Funct
219(1):269–281; doi:10.1007/s00429-012-0498-y.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Karcher NR, Barch DM. 2021. The ABCD study: Understanding the development of risk for mental and physical health outcomes. Neuropsychopharmacology
46(1):131–142; doi:10.1038/s41386-020-0736-6.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Kessler RC, Angermeyer M, Anthony JC, DE Graaf R, Demyttenaere K, Gasquet I, et al. 2007. Lifetime prevalence and age-of-onset distributions of mental disorders in the World Health Organization’s World Mental Health Survey Initiative. World Psychiatry
6(3):168–176.
[PMC free article] [PubMed] [Google Scholar]
Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE. 2005. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry
62(6):593–602; doi:10.1001/archpsyc.62.6.593.
[DOI] [PubMed] [Google Scholar]
Khan A, Plana-Ripoll O, Antonsen S, Brandt J, Geels C, Landecker H, et al. 2019. Environmental pollution is associated with increased risk of psychiatric disorders in the US and Denmark. PLoS Biol
17(8):e3000353; doi:10.1371/journal.pbio.3000353.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Kim SS, Vuong AM, Dietrich KN, Chen A. 2021. Proximity to traffic and exposure to polycyclic aromatic hydrocarbons in relation to Attention Deficit Hyperactivity Disorder and conduct disorder in US children. Int J Hyg Environ Health
232:113686; doi:10.1016/j.ijheh.2020.113686.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Kusters MSW, Essers E, Muetzel E, Ambrós A, Tiemeier H, Guxens M. 2022. Air pollution exposure during pregnancy and childhood, cognitive function, and emotional and behavioral problems in adolescents. Environ Res
214(Pt 2):113891; doi:10.1016/j.envres.2022.113891.
[DOI] [PubMed] [Google Scholar]
LaMantia AS, Rakic P. 1990. Axon overproduction and elimination in the corpus callosum of the developing rhesus monkey. J Neurosci
10(7):2156–2175; doi:10.1523/JNEUROSCI.10-07-02156.1990.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Landrigan PJ, Fuller R, Acosta NJR, Adeyi O, Arnold R, Basu N, et al. 2018. The Lancet Commission on pollution and health. Lancet
391(10119):462–512; doi:10.1016/S0140-6736(17)32345-0.
[DOI] [PubMed] [Google Scholar]
Lantrip C, Isquith PK, Koven NS, Welsh K, Roth RM. 2016. Executive function and emotion regulation strategy use in adolescents. Appl Neuropsychol Child
5(1):50–55; doi:10.1080/21622965.2014.960567.
[DOI] [PubMed] [Google Scholar]
Latham RM, Kieling C, Arseneault L, Botter-Maio Rocha T, Beddows A, Beevers SD, et al. 2021. Childhood exposure to ambient air pollution and predicting individual risk of depression onset in UK adolescents. J Psychiatr Res
138:60–67; doi:10.1016/j.jpsychires.2021.03.042.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Le Bihan D. 2003. Looking into the functional architecture of the brain with diffusion MRI. Nat Rev Neurosci
4(6):469–480; doi:10.1038/nrn1119.
[DOI] [PubMed] [Google Scholar]
Lebel C, Treit S, Beaulieu C. 2019. A review of diffusion MRI of typical white matter development from early childhood to young adulthood. NMR Biomed
32(4):e3778; doi:10.1002/nbm.3778.
[DOI] [PubMed] [Google Scholar]
LeWinn KZ, Connolly CG, Wu J, Drahos M, Hoeft F, Ho TC, et al. 2014. White matter correlates of adolescent depression: Structural evidence for frontolimbic disconnectivity. J Am Acad Child Adolesc Psychiatry
53(8):899–909, 909.e891–897; doi:10.1016/j.jaac.2014.04.021.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Li D, Ji S, Guo Y, Sang N. 2021. Ambient NO2 exposure sex-specifically impairs myelin and contributes to anxiety and depression-like behaviors of C57BL/6J mice. J Hazard Mater
416:125836; doi:10.1016/j.jhazmat.2021.125836.
[DOI] [PubMed] [Google Scholar]
Liao M, Yang F, Zhang Y, He Z, Su L, Li L. 2014. White matter abnormalities in adolescents with generalized anxiety disorder: A diffusion tensor imaging study. BMC Psychiatry
14:41; doi:10.1186/1471-244X-14-41.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Likhtik E, Pelletier JG, Paz R, Paré D. 2005. Prefrontal control of the amygdala. J Neurosci
25(32):7429–7437; doi:10.1523/JNEUROSCI.2314-05.2005.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Loftus CT, Ni Y, Szpiro AA, Hazlehurst MF, Tylavsky FA, Bush NR, et al. 2020. Exposure to ambient air pollution and early childhood behavior: A longitudinal cohort study. Environ Res
183:109075; doi:10.1016/j.envres.2019.109075.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Lubben N, Ensink E, Coetzee G, Labrie V. 2021. The enigma and implications of brain hemispheric asymmetry in neurodegenerative diseases. Brain Commun
3(3):fcab211; doi:10.1093/braincomms/fcab211.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Lubczyńska MJ, Muetzel RL, El Marroun H, Basagana X, Strak M, Denault W, et al. 2020. Exposure to air pollution during pregnancy and childhood, and white matter microstructure in preadolescents. Environ Health Perspect
128(2):27005; doi:10.1289/EHP4709.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Lubczyńska MJ, Muetzel RL, El Marroun H, Hoek G, Kooter IM, Thomson EM, et al. 2021. Air pollution exposure during pregnancy and childhood and brain morphology in preadolescents. Environ Res
198:110446; doi:10.1016/j.envres.2020.110446.
[DOI] [PubMed] [Google Scholar]
Luciana M, Bjork JM, Nagel BJ, Barch DM, Gonzalez R, Nixon, et al. 2018. Adolescent neurocognitive development and impacts of substance use: Overview of the adolescent brain cognitive development (ABCD) baseline neurocognition battery. Dev Cogn Neurosci
32:67–79; doi:10.1016/j.dcn.2018.02.006.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Margolis AE, Herbstman JB, Davis KS, Thomas VK, Tang D, Wang Y, et al. 2016. Longitudinal effects of prenatal exposure to air pollutants on self-regulatory capacities and social competence. J Child Psychol Psychiatry
57(7):851–860; doi:10.1111/jcpp.12548.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Margolis AE, Liu R, Conceicao VA, Ramphal B, Pagliaccio D, DeSerisy ML, et al. 2022. Convergent neural correlates of prenatal exposure to air pollution and behavioral phenotypes of risk for internalizing and externalizing problems: Potential biological and cognitive pathways. Neurosci Biobehav Rev
137:104645; doi:10.1016/j.neubiorev.2022.104645.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Mennitt D, Sherrill K, Fristrup K. 2014. A geospatial model of ambient sound pressure levels in the contiguous United States. J Acoust Soc Am
135(5):2746–2764; doi:10.1121/1.4870481.
[DOI] [PubMed] [Google Scholar]
Mills KL, Goddings AL, Herting MM, Meuwese R, Blakemore SJ, Crone EA, et al. 2016. Structural brain development between childhood and adulthood: Convergence across four longitudinal samples. Neuroimage
141:273–281; doi:10.1016/j.neuroimage.2016.07.044.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Morawska L, Ayoko GA, Bae GN, Buonanno G, Chao CYH, Clifford S, et al. 2017. Airborne particles in indoor environment of homes, schools, offices and aged care facilities: The main routes of exposure. Environ Int
108:75–83; doi:10.1016/j.envint.2017.07.025.
[DOI] [PubMed] [Google Scholar]
Mori S, Tournier J-D. 2014. Moving Beyond DTI: High Angular Resolution Diffusion Imaging (HARDI). In: Introduction to Diffusion Tensor Imaging and Higher Order Models, Oxford, UK: Academic Press: 65–78. [Google Scholar]
Morris RH, Counsell SJ, McGonnell IM, Thornton C. 2021. Early life exposure to air pollution impacts neuronal and glial cell function leading to impaired neurodevelopment. Bioessays
43(9):e2000288; doi:10.1002/bies.202000288.
[DOI] [PubMed] [Google Scholar]
Mortamais M, Pujol J, Martínez-Vilavella G, Fenoll R, Reynes C, Sabatier R, et al. 2019. Effects of prenatal exposure to particulate matter air pollution on corpus callosum and behavioral problems in children. Environ Res
178:108734; doi:0.1016/j.envres.2019.108734.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Mortamais M, Pujol J, van Drooge BL, Macià D, Martínez-Vilavella G, Reynes C, et al. 2017. Effect of exposure to polycyclic aromatic hydrocarbons on basal ganglia and attention-deficit hyperactivity disorder symptoms in primary school children. Environ Int
105:12–19; doi:10.1016/j.envint.2017.04.011.
[DOI] [PubMed] [Google Scholar]
Mujahid MS, Diez Roux AV, Morenoff JD, Raghunathan T. 2007. Assessing the measurement properties of neighborhood scales: From psychometrics to ecometrics. Am J Epidemiol
165(8):858–867; doi:10.1093/aje/kwm040.
[DOI] [PubMed] [Google Scholar]
Mundorf A, Peterburs J, Ocklenburg S. 2021. Asymmetry in the central nervous system: A clinical neuroscience perspective. Front Syst Neurosci
15:733898; doi:10.3389/fnsys.2021.733898.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Munoz-Pizza DM, Villada-Canela M, Reyna MA, Texcalac-Sangrador JL, Osornio-Vargas AR. 2020. Air pollution and children’s respiratory health: A scoping review of socioeconomic status as an effect modifier. Int J Public Health
65(5):649–660; doi:10.1007/s00038-020-01378-3.
[DOI] [PubMed] [Google Scholar]
Myhre O, Låg M, Villanger GD, Oftedal B, Øvrevik J, Holme JA, et al. 2018. Early life exposure to air pollution particulate matter (PM) as risk factor for attention deficit/hyperactivity disorder (ADHD): Need for novel strategies for mechanisms and causalities. Toxicol Appl Pharmacol
354:196–214; doi:10.1016/j.taap.2018.03.015.
[DOI] [PubMed] [Google Scholar]
Nagel BJ, Bathula D, Herting M, Schmitt C, Kroenke CD, Fair, et al. 2011. Altered white matter microstructure in children with attention-deficit/hyperactivity disorder. J Am Acad Child Adolesc Psychiatry
50(3):283–292; doi:10.1016/j.jaac.2010.12.003.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Nakajima R, Kinoshita M, Shinohara H, Nakada M. 2020. The superior longitudinal fascicle: Reconsidering the fronto-parietal neural network based on anatomy and function. Brain Imaging Behav
14:2817–2830; doi:10.1007/s11682-019-00187-4.
[DOI] [PubMed] [Google Scholar]
Nephew BC, Nemeth A, Hudda N, Beamer G, Mann P, Petitto J, et al. 2020. Traffic-related particulate matter affects behavior, inflammation, and neural integrity in a developmental rodent model. Environ Res
183:109242; doi:10.1016/j.envres.2020.109242.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Newbury JB, Arseneault L, Beevers S, Kitwiroon N, Roberts S, Pariante CM, et al. 2019. Association of air pollution exposure with psychotic experiences during adolescence. JAMA Psychiatry
76(6):614–623; doi:10.1001/jamapsychiatry.2019.0056.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Newman NC, Ryan P, Lemasters G, Levin L, Bernstein D, Hershey GK, et al. 2013. Traffic-related air pollution exposure in the first year of life and behavioral scores at 7 years of age. Environ Health Perspect
121(6):731–736; doi:10.1289/ehp.1205555.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Ochsner KN, Ray RD, Cooper JC, Robertson ER, Chopra S, Gabrieli JD, et al. 2004. For better or for worse: Neural systems supporting the cognitive down- and up-regulation of negative emotion. Neuroimage
23(2):483–499; doi:10.1016/j.neuroimage.2004.06.030.
[DOI] [PubMed] [Google Scholar]
Ocklenburg S. 2022. Laterality. In: Encyclopedia of Behavioral Neuroscience, 2nd edition (Second Edition; Sala SD, ed), Amsterdam, Netherlands: Elsevier, 350–356. [Google Scholar]
Ocklenburg S, Peterburs J, Mundorf A. 2022. Hemispheric asymmetries in the amygdala: A comparative primer. Prog Neurobiol
214:102283; doi:10.1016/j.pneurobio.2022.102283.
[DOI] [PubMed] [Google Scholar]
Oldfield RC. 1971. The assessment and analysis of handedness: The Edinburgh inventory. Neuropsychologia
9(1):97–113; doi:10.1016/0028-3932(71)90067-4.
[DOI] [PubMed] [Google Scholar]
Palmer CE, Pecheva D, Iversen JR, Hagler DJ, Jr, Sugrue L, Nedelec P, et al. 2022. Corrigendum to “Microstructural development from 9 to 14 years: Evidence from the ABCD study” [Dev. Cognit. Neurosci. 53 (2022) 101044]. Dev Cogn Neurosci
54:101063; doi:10.1016/j.dcn.2022.101063.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Paulich KN, Ross JM, Lessem JM, Hewitt JK. 2021. Screen time and early adolescent mental health, academic, and social outcomes in 9- and 10-year-old children: Utilizing the Adolescent Brain Cognitive Development (ABCD) study. PLoS One
16(9) e0256591; doi:10.1371/journal.pone.0256591.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Paulus MP, Squeglia LM, Bagot K, Jacobus J, Kuplicki R, Breslin FJ, et al. 2019. Screen media activity and brain structure in youth: Evidence for diverse structural correlation networks from the ABCD study. Neuroimage
185:140–153; doi:10.1016/j.neuroimage.2018.10.040.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Paus T, Keshavan M, Giedd JN. 2008. Why do many psychiatric disorders emerge during adolescence?
Nat Rev Neurosci
9(12):947–957; doi:10.1038/nrn2513.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Perera F, Weiland K, Neidell M, Wang S. 2014a. Prenatal exposure to airborne polycyclic aromatic hydrocarbons and IQ: Estimated benefit of pollution reduction. J Public Health Policy
35(3):327–336; doi:10.1057/jphp.2014.14.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Perera FP, Chang HW, Tang D, Roen EL, Herbstman J, Margolis A, et al. 2014b. Early-life exposure to polycyclic aromatic hydrocarbons and ADHD behavior problems. PLoS One
9(11):e111670; doi:10.1371/journal.pone.0111670.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Perera FP, Tang D, Wang S, Vishnevetsky J, Zhang B, Diaz D, et al. 2012. Prenatal polycyclic aromatic hydrocarbon (PAH) exposure and child behavior at age 6–7 years. Environ Health Perspect
120(6):921–926; doi:10.1289/ehp.1104315.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Perera FP, Wang S, Vishnevetsky J, Zhang B, Cole KJ, Tang D, et al. 2011. Polycyclic aromatic hydrocarbons-aromatic DNA adducts in cord blood and behavior scores in New York City children. Environ Health Perspect
119(8):1176–1181; doi:10.1289/ehp.1002705.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Pérez-Crespo L, Kusters MSW, López-Vicente M, Lubczyńska MJ, Foraster M, White T, et al. 2022. Exposure to traffic-related air pollution and noise during pregnancy and childhood, and functional brain connectivity in preadolescents. Environ Int
164:107275; doi:10.1016/j.envint.2022.107275.
[DOI] [PubMed] [Google Scholar]
Pessoa L. 2009. How do emotion and motivation direct executive control?
Trends Cogn Sci
13(4):160–166; doi:10.1016/j.tics.2009.01.006.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Petersen AC, Crockett L, Richards M, Boxer A. 1988. A self-report measure of pubertal status: Reliability, validity, and initial norms. J Youth Adolesc
17(2):117–133; doi:10.1007/BF01537962.
[DOI] [PubMed] [Google Scholar]
Peterson BS, Bansal R, Sawardekar S, Nati C, Elgabalawy ER, Hoepner LA, et al. 2022. Prenatal exposure to air pollution is associated with altered brain structure, function, and metabolism in childhood. J Child Psychol Psychiatry
63(11):1316–1331; doi:10.1111/jcpp.13578.
[DOI] [PubMed] [Google Scholar]
Peterson BS, Rauh VA, Bansal R, Hao X, Toth Z, Nati G, et al. 2015a. Effects of prenatal exposure to air pollutants (polycyclic aromatic hydrocarbons) on the development of brain white matter, cognition, and behavior in later childhood. JAMA Psychiatry
72(6):531–540; doi:10.1001/jamapsychiatry.2015.57.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Peterson D, Mahajan R, Crocetti D, Mejia A, Mostofsky S. 2015b. Left-hemispheric microstructural abnormalities in children with high-functioning autism spectrum disorder. Autism Research
8(1):61–72; doi:10.1002/aur.1413.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Phan KL, Fitzgerald DA, Nathan PJ, Moore GJ, Uhde TW, Tancer ME. 2005. Neural substrates for voluntary suppression of negative affect: A functional magnetic resonance imaging study. Biol Psychiatry
57(3):210–219; doi:10.1016/j.biopsych.2004.10.030.
[DOI] [PubMed] [Google Scholar]
Philippi CL, Mehta S, Grabowski T, Adolphs R, Rudrauf D. 2009. Damage to association fiber tracts impairs recognition of the facial expression of emotion. J Neurosci
29(48):15089–15099; doi:10.1523/JNEUROSCI.0796-09.2009.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Pujol J, Fenoll R, Macià D, Martínez-Vilavella G, Alvarez-Pedrerol M, Rivas I, et al. 2016a. Airborne copper exposure in school environments associated with poorer motor performance and altered basal ganglia. Brain Behav
6(6):e00467; doi:10.1002/brb3.467.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Pujol J, Martinez-Vilavella G, Macia D, Fenoll R, Alvarez-Pedrerol M, Rivas I, et al. 2016b. Traffic pollution exposure is associated with altered brain connectivity in school children. Neuroimage
129:175–184; doi:10.1016/j.neuroimage.2016.01.036.
[DOI] [PubMed] [Google Scholar]
Puzzo I, Seunarine K, Sully K, Darekar A, Clark C, Sonuga-Barke EJS, et al. 2018. Altered white-matter microstructure in conduct disorder is specifically associated with elevated callous-unemotional traits. J Abnorm Child Psychol
46(7):1451–1466; doi:10.1007/s10802-017-0375-5.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Qin S, Young CB, Supekar K, Uddin LQ, Menon V. 2012. Immature integration and segregation of emotion-related brain circuitry in young children. Proc Natl Acad Sci U S A
109(20):7941–7946; doi:10.1073/pnas.1120408109.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Quirk GJ, Beer JS. 2006. Prefrontal involvement in the regulation of emotion: Convergence of rat and human studies. Curr Opin Neurobiol
16(6):723–727; doi:10.1016/j.conb.2006.07.004.
[DOI] [PubMed] [Google Scholar]
R Core Team. 2021. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. [Google Scholar]
R Core Team. 2022. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. [Google Scholar]
Ramaiah P, Altalbawy FMA, Margiana R, Kumar NB, Kahar F, Jalil AT, et al. 2023. The association between prenatal exposure to polycyclic aromatic hydrocarbons and childhood intelligence: A systematic review of observational studies. Environ Sci Pollut Res Int
30(8):19592–19601; doi:10.1007/s11356-023-25309-w.
[DOI] [PubMed] [Google Scholar]
Rasnick E, Ryan PH, Bailer AJ, Fisher T, Parsons PJ, Yolton K, et al. 2021. Identifying sensitive windows of airborne lead exposure associated with behavioral outcomes at age 12. Environ Epidemiol
5(2):e144; doi:10.1097/EE9.0000000000000144.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Reuben A, Arseneault L, Beddows A, Beevers SD, Moffitt TE, Ambler A, et al. 2021. Association of air pollution exposure in childhood and adolescence with psychopathology at the transition to adulthood. JAMA Netw Open
4(4) e217508; doi:10.1001/jamanetworkopen.2021.7508.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Roalf DR, de la Garza AG, Rosen A, Calkins ME, Moore TM, Quarmley M, et al. 2020. Alterations in white matter microstructure in individuals at persistent risk for psychosis. Mol Psychiatry
25(10):2441–2454; doi:10.1038/s41380-019-0360-1.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Roberts S, Arseneault L, Barratt B, Beevers S, Danese A, Odgers CL, et al. 2019. Exploration of NO2 and PM2.5 air pollution and mental health problems using high-resolution data in London-based children from a UK longitudinal cohort study. Psychiatry Res
272:8–17; doi:10.1016/j.psychres.2018.12.050.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Senova S, Fomenko A, Gondard E, Lozano AM. 2020. Anatomy and function of the fornix in the context of its potential as a therapeutic target. J Neurol Neurosurg Psychiatry
91(5):547–559; doi:10.1136/jnnp-2019-322375.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Shih P, Chiang TL, Wu CD, Shu BC, Lung FW, Guo YL. 2022. Air pollution during the perinatal period and neurodevelopment in children: A national population study in Taiwan. Dev Med Child Neurol; doi:10.1111/dmcn.15430. [DOI] [PubMed] [Google Scholar]
Snyder HR, Miyake A, Hankin BL. 2015. Advancing understanding of executive function impairments and psychopathology: Bridging the gap between clinical and cognitive approaches. Front Psychol
6:328; doi:10.3389/fpsyg.2015.00328.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Solaimani P, Saffari A, Sioutas C, Bondy SC, Campbell A. 2017. Exposure to ambient ultrafine particulate matter alters the expression of genes in primary human neurons. Neurotoxicology
58:50–57; doi:10.1016/j.neuro.2016.11.001.
[DOI] [PubMed] [Google Scholar]
Solmi M, Radua J, Olivola M, Croce E, Soardo L, Salazar de Pablo G, et al. 2022. Age at onset of mental disorders worldwide: Large-scale meta-analysis of 192 epidemiological studies. Mol Psychiatry
27(1):281–295; doi:10.1038/s41380-021-01161-7.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Somerville LH, Jones RM, Ruberry EJ, Dyke JP, Glover G, Casey BJ. 2013. The medial prefrontal cortex and the emergence of self-conscious emotion in adolescence. Psychol Sci
24(8):1554–1562; doi:10.1177/0956797613475633.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Stingone JA, Pandey OP, Claudio L, Pandey G. 2017. Using machine learning to identify air pollution exposure profiles associated with early cognitive skills among US children. Environ Pollut
230:730–740; doi:10.1016/j.envpol.2017.07.023.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Suades-González E, Gascon M, Guxens M, Sunyer J. 2015. Air pollution and neuropsychological development: A review of the latest evidence. Endocrinology
156(10):3473–3482; doi:10.1210/en.2015-1403.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Subhramanyam CS, Wang C, Hu Q, Dheen ST. 2019. Microglia-mediated neuroinflammation in neurodegenerative diseases. Semin Cell Dev Biol
94:112–120; doi:10.1016/j.semcdb.2019.05.004.
[DOI] [PubMed] [Google Scholar]
Sukumaran K, Cardenas-Iniguez C, Burnor E, Bottenhorn KL, Hackman DA, McConnell R, et al. 2023. Ambient fine particulate exposure and subcortical gray matter microarchitecture in 9- and 10-year-old children across the United States. iScience
26(3):106087; doi:10.1016/j.isci.2023.106087.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Tamiz AP, Koroshetz WJ, Dhruv NT, Jett DA. 2022. A focus on the neural exposome. Neuron
110(8):1286–1289; doi:10.1016/j.neuron.2022.03.019.
[DOI] [PubMed] [Google Scholar]
Tamnes CK, Herting MM, Goddings AL, Meuwese R, Blakemore SJ, Dahl RE, et al. 2017. Development of the cerebral cortex across adolescence: A multisample study of inter-related longitudinal changes in cortical volume, surface area, and thickness. J Neurosci
37(12):3402–3412; doi:10.1523/JNEUROSCI.3302-16.2017.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Tessum MW, Anenberg SC, Chafe ZA, Henze DK, Kleiman G, Kheirbek I, et al.2022. Sources of ambient PM2.5 exposure in 96 global cities. Atmos Environ (1994)
286:119234; doi:10.1016/j.atmosenv.2022.119234.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Textor J, van der Zander B, Gilthorpe MS, Liskiewicz M, Ellison GT. 2016. Robust causal inference using directed acyclic graphs: The R package ‘dagitty.’
Int J Epidemiol
45(6):1887–1894; doi:10.1093/ije/dyw341.
[DOI] [PubMed] [Google Scholar]
Tottenham N, Hare TA, Casey BJ. 2011. Behavioral assessment of emotion discrimination, emotion regulation, and cognitive control in childhood, adolescence, and adulthood. Front Psychol
2:39; doi:10.3389/fpsyg.2011.00039.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Uban KA, Horton MK, Jacobus J, Heyser C, Thompson WK, Tapert SF, et al. 2018. Biospecimens and the ABCD study: Rationale, methods of collection, measurement and early data. Dev Cogn Neurosci
32:97–106; doi:10.1016/j.dcn.2018.03.005.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Urdangarin A, Goicoa T, Ugarte MD. 2023. Evaluating recent methods to overcome spatial confounding. Rev Mat Complut
36(May):333–360; doi:10.1007/s13163-022-00449-8. [Google Scholar]
Vanes LD, Dolan RJ. 2021. Transdiagnostic neuroimaging markers of psychiatric risk: A narrative review. Neuroimage Clin
30:102634; doi:10.1016/j.nicl.2021.102634.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Volk HE, Perera F, Braun JM, Kingsley SL, Gray K, Buckley, et al. 2021. Prenatal air pollution exposure and neurodevelopment: A review and blueprint for a harmonized approach within ECHO. Environ Res
196:110320; doi:10.1016/j.envres.2020.110320.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Volkow ND, Koob GF, Croyle RT, Bianchi DW, Gordon JA, Koroshetz WJ, et al. 2018. The conception of the ABCD study: From substance use to a broad NIH collaboration. Dev Cogn Neurosci
32
4–7; doi:10.1016/j.dcn.2017.10.002.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Von Der Heide RJ, Skipper LM, Klobusicky E, Olson IR. 2013. Dissecting the uncinate fasciculus: Disorders, controversies and a hypothesis. Brain
136(6):1692–1707; doi:10.1093/brain/awt094.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Wakschlag LS, Perlman SB, Blair RJ, Leibenluft E, Briggs-Gowan MJ, Pine DS. 2018. The neurodevelopmental basis of early childhood disruptive behavior: Irritable and callous phenotypes as exemplars. Am J Psychiatry
175(2):114–130; doi:10.1176/appi.ajp.2017.17010045.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Wang P, Zhao Y, Li J, Zhou Y, Luo R, Meng X, et al. 2021. Prenatal exposure to ambient fine particulate matter and early childhood neurodevelopment: A population-based birth cohort study. Sci Total Environ
785:147334; doi:10.1016/j.scitotenv.2021.147334.
[DOI] [PubMed] [Google Scholar]
Woodward NC, Haghani A, Johnson RG, Hsu TM, Saffari A, Sioutas C, et al. 2018. Prenatal and early life exposure to air pollution-induced hippocampal vascular leakage and impaired neurogenesis in association with behavioral deficits. Transl Psychiatry
8(1):261; doi:10.1038/s41398-018-0317-1.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Wright P, Albarracin D, Brown RD, Li H, He G, Liu YG. 2008. Dissociated responses in the amygdala and orbitofrontal cortex to bottom-up and top-down components of emotional evaluation. Neuroimage
39(2):894–902; doi:10.1016/j.neuroimage.2007.09.014.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Yang T, Wang J, Huang J, Kelly FJ, Li G. 2023. Long-term exposure to multiple ambient air pollutants and association with incident depression and anxiety. JAMA Psychiatry; doi:10.1001/jamapsychiatry.2022.4812. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yap MB, Allen NB, Sheeber L. 2007. Using an emotion regulation framework to understand the role of temperament and family processes in risk for adolescent depressive disorders. Clin Child Fam Psychol Rev
10(2):180–196; doi:10.1007/s10567-006-0014-0.
[DOI] [PubMed] [Google Scholar]
Yi C, Wang Q, Qu Y, Niu J, Oliver BG, Chen H. 2022. In-utero exposure to air pollution and early-life neural development and cognition. Ecotoxicol Environ Saf
238:113589; doi:10.1016/j.ecoenv.2022.113589.
[DOI] [PubMed] [Google Scholar]
Yolton K, Khoury JC, Burkle J, LeMasters G, Cecil K, Ryan P. 2019. lifetime exposure to traffic-related air pollution and symptoms of depression and anxiety at age 12 years. Environ Res
173:199–206; doi:10.1016/j.envres.2019.03.005.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Younan D, Tuvblad C, Franklin M, Lurmann F, Li L, Wu J, et al. 2018. Longitudinal analysis of particulate air pollutants and adolescent delinquent behavior in Southern California. J Abnorm Child Psychol
46(6):1283–1293; doi:10.1007/s10802-017-0367-5.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Yuchi W, Brauer M, Czekajlo A, Davies HW, Davis Z, Guhn M, et al. 2022. Neighborhood environmental exposures and incidence of attention deficit/hyperactivity disorder: A population-based cohort study. Environ Int
161:107120; doi:10.1016/j.envint.2022.107120.
[DOI] [PubMed] [Google Scholar]
Yurgelun-Todd DA, Killgore WD. 2006. Fear-related activity in the prefrontal cortex increases with age during adolescence: A preliminary fMRI study. Neurosci Lett
406(3):194–199; doi:10.1016/j.neulet.2006.07.046.
[DOI] [PubMed] [Google Scholar]
Zundel CG, Ryan P, Brokamp C, Heeter A, Huang Y, Strawn JR, et al. 2022. Air pollution, depressive and anxiety disorders, and brain effects: A systematic review. Neurotoxicology
93:272–300; doi:10.1016/j.neuro.2022.10.011.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Res Rep Health Eff Inst.
Commentary by Review Committee
Research Report 225, Air Pollution Exposure, Prefrontal Connectivity, and Emotional Behavior in Early Adolescence, by M.M. Herting et al.
Dr. Megan M. Herting’s 3-year study, “Air Pollution Exposure and Prefrontal Connectivity in Early Adolescence,” began in June 2021. Total expenditures were $499,985. The draft Investigators’ Report from Herting and colleagues was received for review in June 2023. A revised report, received in January 2024, was accepted for publication in March 2024. During the review process, the HEI Review Committee and the investigators had the opportunity to exchange comments and clarify issues in both the Investigators’ Report and the Review Committee’s Commentary. As a co-investigator of the Herting Investigators’ Report, Dr. Kiros T. Berhane was not involved in its evaluation by the Review Committee.
This report has not been reviewed by public or private party institutions, including those that support the Health Effects Institute; therefore, it may not reflect the views of these parties; no endorsements by them should be inferred.
Early life exposure to air pollution is associated with altered brain development, neurodevelopmental disorders (such as autism and attention deficit hyperactivity disorder), and behavioral and mental health problems (Cory-Slechta et al. 2023). Neurodevelopmental disabilities affect about 8.5% of children in the United States, and the prevalence is increasing (Zablotsky et al. 2023). Neurodevelopmental disorders are characterized by lifelong impairments in cognition and behavior — often coinciding with psychiatric illnesses — and can predispose affected individuals to neurodegenerative diseases later in life (Cory-Slechta et al. 2023). Considerable air pollution research has focused on the prenatal and early life periods because it is a time of rapid nervous system development and heightened vulnerability. However, brain development continues through early adulthood, and less is known about how air pollution exposure might affect children in preadolescence.
To evaluate the effects of early adolescent exposure to air pollution on neurodevelopment, Dr. Megan Herting of the University of Southern California submitted an application to HEI titled “Air Pollution Exposure and Prefrontal Connectivity in Early Adolescence” in response to HEI’s Request for Applications 18-2: Walter A. Rosenblith New Investigator Award. This award was established to provide support for an outstanding new investigator at the assistant professor level to conduct research in the area of air pollution and health; it is unrestricted with respect to the specific topic of research. Dr. Herting proposed to examine whether exposure to two major air pollutants, outdoor fine particulate matter (PM2.5*)and nitrogen dioxide (NO2), are associated with changes in structural connectivity in the prefrontal cortex of the brain — an area that is involved in numerous high-level cognitive processes — and with measures of emotional health over a 1-year period in children who are transitioning to adolescence.
HEI’s Research Committee recommended funding Dr. Herting’s application because it focused on an understudied period of neurodevelopment and leveraged an existing nationwide cohort study with rich data. The Committee appreciated that prenatal exposures would be considered in addition to the primary analysis of current exposures. The study started in 2020.
This Commentary provides the HEI Review Committee’s independent evaluation of the study. It is intended to aid the sponsors of HEI and the public by highlighting both the strengths and limitations of the study and by placing the results presented in the Investigators’ Report into a broader scientific and regulatory context.
SCIENTIFIC AND REGULATORY BACKGROUND
Air pollutants such as PM2.5 and NO2 can have substantial negative effects on brain health. PM2.5 consists of suspended particles with a diameter ≤2.5 μm, which when inhaled, can reach deep into the lungs and enter the circulatory system. PM is emitted directly from combustion sources (such as smokestacks, vehicle exhaust, and wildfires) but is also formed by chemical reactions in the atmosphere (Li et al. 2022). NO2 is a gas, the majority of which is rapidly formed by chemical reactions following fuel combustion-related nitric oxide emissions (CARB 2024). Exposure to these pollutants has been associated with altered brain structure and function, including changes to gray and white matter (Guxens et al. 2022; Peterson et al. 2022).
Gray and white matter are two distinct brain areas that serve different structural and functional roles. Gray matter primarily consists of neuronal cell bodies located at the outer areas of the brain and is predominantly involved in processing and interpreting information. In contrast, white matter is located deeper in the brain and consists of myelinated axons that connect neurons to facilitate communication between different brain regions (Filley and Fields 2016). White matter is organized into anatomically distinct bundles known as tracts; these different tracts are linked to the performance of specific cognitive functions. Damage to or disconnection of the white matter tracts can result in cognitive dysfunction (Ribeiro et al. 2024).
The effects of air pollution exposure on the brain have been studied using advanced neuroimaging techniques, including structural, diffusion tensor, and functional magnetic resonance imaging (MRI). Structural MRI provides detailed images of brain anatomy, enabling measurement of gray and white matter volumes. Diffusion tensor MRI maps structural differences in water diffusion along axons and can detect changes in white matter integrity and connectivity. Lastly, functional MRI measures changes in blood flow to assess brain activity (Op de Beeck and Nakatani 2019).
Regulatory efforts have sought to mitigate the health effects of PM2.5 and NO2. The United States Environmental Protection Agency (US EPA) has established National Ambient Air Quality Standards, limiting the 3-year annual average PM2.5 concentrations to 9 μg/m3 and the annual average NO2 concentrations to 53 parts per billion (US EPA 1971, 2024a). The 2019 US EPA Integrated Science Assessment for PM classified the associations between long-term PM2.5 exposure and nervous system outcomes as “likely to be causal” (US EPA 2019). It is also important to note that the particles in PM2.5 measurements can include metals such as lead, a regulated criteria pollutant known to be causally related to multiple adverse nervous system effects (US EPA 2024b). In contrast, the US EPA determined there was inadequate evidence to infer a causal relation between NO2 and neurodevelopment due to inconsistencies in the reported literature (US EPA 2016). Additional research is warranted to help determine whether further safeguards are needed to protect health.
STUDY OBJECTIVES
The study aimed to accomplish the following:
Assess the association between 1-year average air pollution exposure and prefrontal cortex white matter connectivity at ages 9–10 (baseline).
Assess the effect of 1-year average air pollution exposure on emotional problems during adolescence at baseline (ages 9–10) and 1 year later (ages 10–11).
Herting and colleagues hypothesized that higher air pollution exposures would be related to decreased prefrontal cortex white matter connectivity and increased emotional behavior problems. To evaluate this hypothesis, they leveraged data from the Adolescent Brain Cognitive Development (ABCD) Study, a nationally representative multicenter cohort of children with detailed information on demographic and social factors. They assessed neurodevelopment by using brain imaging to detect differences in prefrontal cortex white matter connectivity at baseline and by using caregiver-reported emotional behavior problems at baseline and 1 year later. They evaluated whether outdoor PM2.5 and NO2 exposure estimates at the participant’s home address in the preceding year and during the 9-month prenatal period were associated with differences in neurodevelopment.
The original project aims also included a diagnostic assessment for schizophrenia and affective disorders, and a mediation analysis to assess whether air pollution influenced emotional behaviors through changes in prefrontal cortex connectivity. However, due to issues beyond the investigators’ control, the diagnostic assessments were unavailable. Furthermore, the Research Committee and investigators jointly determined not to proceed with the mediation analyses because it was unclear whether associations between air pollution and emotional behaviors were clinically meaningful.
SUMMARY OF METHODS AND STUDY DESIGN
ABCD STUDY POPULATION
The ABCD study is the largest longitudinal study of brain development in the United States. It is organized and conducted by a consortium of 13 federal collaborators and 21 research institutes and coordinated by directors at the University of California, San Diego (Auchter et al. 2018, Volkow et al. 2018). The study enrolled 11,880 sociodemographically diverse children, ages 9–10, in the United States between 2016 and 2018. Children were recruited from a combination of (1) randomly selected schools within a 50-mile radius of the participating research institutes (Commentary Figure 1) across 17 states, (2) mailings using address information from birth records, and (3) nonrandom outreach methods for 10% of participants.
Children were excluded if they had a history of major neurological, psychiatric, or developmental disorders, substance use disorders, traumatic brain injury, preterm birth, very low birth weight, birth complications, or current use of antipsychotic or mood-stabilizing medications. Children were also excluded if the family planned to move far from an ABCD Study site. (Garavan et al. 2018; Volkow et al. 2018). Child height, weight, handedness, caregiver-reported sociodemographic, lifestyle, and child health information were collected at the baseline visit.
EXPOSURE ASSESSMENT
Herting and colleagues estimated daily ambient PM2.5 and NO2 concentrations at 1-km resolution across the continental United States by hybrid-machine learning models that combined satellite-based aerosol optical depth, land-use information, chemical transport models, meteorological data, and other inputs (Di et al. 2019, 2020). They validated the models using 2000–2016 monitoring data from the US EPA, the Interagency Monitoring of Protected Visual Environments (IMPROVE), the Clean Air Status and Trends Network (CASTNET), and other local and regional monitoring networks. The total number of monitoring sites for PM2.5 and NO2 was 2,156 and 912 sites, respectively, and the coefficient of determination (R2) between modeled and monitored pollutant concentrations was 0.86 and 0.79, respectively. One-year average air pollution concentrations were linked to the geocoded home address reported at the first study visit (2016–2018), and the preceding 9-month average air pollution concentrations were linked to the geocoded home address on the date of birth (Fan et al. 2021).
NEURODEVELOPMENTAL ASSESSMENT
Brain Connectivity
At each of the 21 research institutes, white matter microstructure in the prefrontal cortex was ascertained by diffusion tensor MRI brain imaging using advanced 3-Tesla scanners (Siemens Prisma, General Electric 750, and Phillips). The MRI scans were harmonized across institutions using detailed quality control and image preprocessing.
The white matter microstructures were estimated using two metrics, mean diffusivity and fractional anisotropy, calculated by AtlasTrack software (Hagler et al. 2009). Mean diffusivity measures how easily water molecules move through tissues, and fractional anisotropy measures the directionality of the water movement. Both metrics are nonspecific brain markers with opposite interpretations; higher levels of fractional anisotropy and lower levels of mean diffusivity generally indicate better white matter organization and integrity, whereas lower levels of fractional anisotropy and higher levels of mean diffusivity suggest potential disease processes. Structural factors that influence mean diffusivity and fractional anisotropy of white matter include the number, density, size, organization, and degree of myelination of nerve fibers (Winston 2012).
Seven white matter tracts — the anterior thalamic radiation, cingulum-cingulate body, cingulum-hippocampal portions, corpus callosum, inferior fronto-occipital fasciculus, superior longitudinal fasciculus, and uncinate fasciculus — were evaluated. Because white matter microstructure changes can be specific to brain hemisphere, the tracts were evaluated separately for the left and right hemispheres, except for the corpus callosum, which traverses both.
Emotional Behavior
Information on emotional behavior was reported by caregivers using the Child Behavior Checklist (Achenbach and Rescorla 2001). The 113-question checklist assesses child behavior over the past 6 months and provides age- and sex-normalized scores for several measures of behavioral problems. This study focused on the measures of total problems, internalizing behavior, externalizing behavior, anxious-depressed behaviors, withdrawn-depressed behaviors, rule-breaking, aggressive behaviors, and attention problems. For all eight measures, higher scores indicate more behavioral problems.
MAIN ANALYSES
Herting and colleagues evaluated the associations between air pollution exposure and neurodevelopmental outcome using single- and two-pollutant multilevel mixed-effect models that accounted for the nested study design (i.e., the fact that children recruited to a specific study site will have more similarities than children recruited from the other study sites) and the small fraction of missing data (6%).
In the models evaluating brain connectivity, they assessed interactions between exposure and the brain hemisphere because exposures might not affect both sides of the brain equally, and they reported stratified results for statistically significant interactions. They also evaluated nonlinearities in exposures by incorporating natural cubic splines and used false discovery rate correction to control for multiple comparisons. The models used to evaluate emotional behavior included Child Behavior Checklist scores from baseline and follow-up. These analyses were not corrected for multiple comparisons because the eight checklist measures were selected a priori.
Confounders were selected by constructing a priori–directed acyclic graphs. They considered 22 sociodemographic, lifestyle, health, neighborhood-level, and MRI-based factors. The minimally sufficient set of confounders for the white matter microstructure outcomes included child age, sex, race and ethnicity, physical activity, screen time, household income, road proximity to home, perceived neighborhood quality, population density, and urbanicity. To increase the precision of regression estimates, they also adjusted for the MRI scanner manufacturer, head motion during the MRI scan, and child handedness. The minimally sufficient set of confounders for the age- and sex-normalized Child Behavior Checklist scores included child race and ethnicity, physical activity, screen time, household income, road proximity to home, perceived neighborhood quality, population density, and urbanicity.
ADDITIONAL ANALYSES
Additional analyses were conducted to assess the robustness of the results due to various factors. Specifically, the investigators (1) excluded adjustment for road proximity to home, population density, and urbanicity because these might be partially controlled for in the exposure models, (2) adjusted for nighttime noise instead of road proximity to home, population density, and urbanicity, (3) excluded participants from rural areas, (4) accounted for duration in the current home as both an adjustment variable and effect modifier, and (5) accounted for prenatal exposure.
SUMMARY OF KEY RESULTS
POPULATION CHARACTERISTICS AND EXPOSURE
The ABCD cohort (N = 11,840) characteristics are shown in Commentary Table 1. Sample size varied by outcome, with 7,546 children having MRI data and 9,815 children having Child Behavior Checklist scores. Population characteristics, however, were similar across the subsamples.
Commentary Table 1.
ABCD Cohort Population Characteristics (N = 11,840)
Data presented as N(%) or mean (standard deviation).
Mean 1-year PM2.5 exposure at baseline (N = 11,840) was 7.7 μg/m3 (Commentary Table 1) and ranged from 1.7 to 15.9 μg/m3. Mean 1-year NO2 exposure was 18.6 ppb and ranged from 0.7 to 37.9 ppb. Mean prenatal exposures (n = 7,848) were higher, at 11.0 μg/m3 and 26.4 ppb for PM2.5 and NO2, respectively. Exposure concentrations varied by study site and urbanicity. Correlations between exposure measurements were weak to moderate, ranging from 0.20 for 1-year PM2.5 and NO2, to 0.64 for 1-year and prenatal NO2.
MAIN ANALYSES
Air Pollution and Brain Connectivity
Using two-pollutant models, Herting and colleagues found that associations between air pollution and white matter connectivity in the prefrontal cortex were pollutant-specific: PM2.5 was associated with better markers of brain connectivity, whereas NO2 was associated with worse markers of brain connectivity. Many of the detected associations were specific to white matter tracts on either the right or left sides of the brain, as evidenced by statistical interactions by brain hemisphere. The associations were also detected in the corpus callosum, which spans both sides of the brain.
Specifically, PM2.5 exposure was associated with less mean diffusivity in six of the seven evaluated white matter tracts, particularly in the left hemisphere (Commentary Figure 2A). These associations were nonlinear: PM2.5 concentrations above the mean exposure of 7.7 μg/m3 were associated with larger decreases in mean diffusivity than PM2.5 concentrations below the mean. In one white matter tract, PM2.5 was associated with higher fractional anisotropy at exposures below the mean but slightly decreased fractional anisotropy at exposures above the mean (Commentary Figure 2B). These findings were not consistent with the study hypothesis.
Associations between annual PM2.5 and NO2 exposure and prefrontal cortex white matter connectivity at ages 9–10. PM2.5 was associated with no change or higher connectivity (A and B), whereas NO2 was associated with lower connectivity (C). Higher mean diffusivity and lower fractional anisotropy indicate disease states. Nonsignificant changes are not shown. ATR = anterior thalamic radiation; CGH = cingulum (hippocampal); SLF = superior longitudinal fasciculus; UNC = uncinate fasciculus.
In contrast, higher NO2 exposure was linearly associated with lower levels of fractional anisotropy in three white matter tracts (Commentary Figure 2B). These findings were consistent with the study hypothesis and might indicate that NO2 exposure might lead to less organized nerve fibers.
Air Pollution and Emotional Behavior
Over the 1 year of follow-up, measures that describe children’s internalizing problems (i.e., anxiety-depression behaviors and withdrawn-depression behaviors) slightly increased, and measures that describe children’s externalizing problems (i.e., attention, rule-breaking, and aggression) slightly decreased, which was expected for the emotional development in this age group (Commentary Figure 3, Time).
Emotional behavior at ages 9–10 at baseline and 1-year follow-up. Left column (Time): emotional behavior changes at 1-year follow-up. Middle columns (PM2.5 and NO2): Associations between annual PM2.5 and NO2 exposure and emotional behavior at baseline. Right columns (PM2.5×Time and NO2×Time): Associations between exposure and emotional behavior at 1-year follow-up. Beta estimates are reported per 1-unit change in year and exposure. Estimates that are larger than 0 indicate increased emotional behavior problems.
Using two-pollutant models, Herting and colleagues found that annual PM2.5 and NO2 exposures were not associated with emotional behavior outcomes at baseline (Commentary Figure 3, PM2.5 and NO2). In contrast to their hypothesis, annual PM2.5 and NO2 exposures were associated with decreased internalizing problems over the 1 year of follow-up (Commentary Figure 3, PM2.5 × Time and NO2 × Time). PM2.5 was also associated with decreased anxiety-depression and decreased aggression, and NO2 was associated with a decrease in total problems over the 1 year of follow-up. Herting and colleagues noted that the magnitude of the associations was very small and might not be clinically meaningful.
Additional Analyses
When information on road proximity to home, population density, urbanicity, nighttime noise, residential duration, or prenatal PM2.5 and NO2 exposure was included in the analysis, or when participants from rural areas were excluded, the results were similar to the main analysis for both brain connectivity and emotional behavior outcomes. Herting and colleagues found that the associations between PM2.5 and NO2 exposure and mean diffusivity varied by residential duration, with shorter time living at a residence generally associated with better white matter connectivity. However, graphical visualization of the associations for PM2.5 exposure revealed a complex association with mean diffusivity by residential duration that was difficult to interpret (Appendix Figure 9). They also found that the associations between PM2.5 and emotional behavior varied by residential duration, with a shorter time living at a residence associated with fewer internalizing and attention problems and a longer time living at a residence associated with more internalizing and attention problems.
HEI REVIEW COMMITTEE’S EVALUATION
This study leveraged the largest longitudinal study of brain development in the United States to evaluate the effect of air pollution exposure on white matter connectivity and emotional behavior in preadolescence. Herting and colleagues found that annual average PM2.5 and NO2 exposures were associated with altered white matter connectivity in the prefrontal cortex, which often varied by brain hemisphere, and higher NO2 was associated with markers of lower connectivity. However, they did not find that air pollution exposure was associated with decrements in caregiver-reported emotional and behavioral problems. The mean estimated 1-year air pollution exposures in this study population were below the level of the US National Ambient Air Quality Standards for long-term exposure.
In its independent review of the study, the HEI Review Committee concluded that this report presents a thorough investigation into associations between exposure to air pollution and certain measures of brain development. Details on the strengths and limitations of the study are discussed below.
STUDY DESIGN, DATASETS, AND ANALYTICAL APPROACHES
The Committee noted that the study implemented a high-quality design and methods, including the use of a nationally representative cohort of children with detailed information on social and demographic characteristics and rigorous assessment and data processing of brain structure. The Committee thought that it was novel and important to study the preadolescent period, which has received less focus than the prenatal and early life periods in research on environmental exposures and neurodevelopment. Some studies suggest that air pollution exposure might be more harmful postnatally compared to prenatally (Lin et al. 2022), and more research later in childhood is needed.
The Committee appreciated the use of complementary neurodevelopmental outcomes — that is, the use of both brain scans and caregiver-reported mental symptoms. The thorough statistical analysis included a detailed evaluation of nonlinear exposure–response curves and the effect of residential duration; adjustment for potentially important factors that could influence neurodevelopment in children, such as screen time; and a careful assessment of missing data.
The Committee appreciated that the exposure assessment included both prenatal and concurrent air pollution exposures because that approach provided a more comprehensive exposure assessment than prior related research assessing air pollution only once or only over shorter periods of development (Parenteau et al. 2024). Herting and colleagues did not observe meaningful changes in the associations when adjusting for prenatal exposure. However, the exposure assessment was based on residential address only, despite the fact that, on average, children in the United States spend 1,000 hours at school each year (ECS 2023).
Previous studies have demonstrated high correlations between estimated residential- and school-based exposure to PM2.5 and NO2, likely because they are generally geographically close; however, residential exposures cannot be assumed to reflect school exposures more broadly because in certain communities the exposures are meaningfully different (Hoek et al. 2024). Although it was beyond the scope of the current study, the Committee recommended that future studies in this cohort should leverage the school-based sampling scheme of the ABCD Study to assess ambient air pollution exposure at school and then incorporate this information into a time–activity-based exposure assessment.
FINDINGS AND INTERPRETATION
Herting and colleagues detected associations between air pollution and white matter microstructure that varied by the specific pollutant, brain region, and measure of white matter connectivity. Their mixed associations are consistent with prior studies using diffusion tensor MRI that report that associations can also vary by exposure timing (Parenteau et al. 2024). For example, a study of preadolescent children in the Netherlands observed that early childhood NO2 exposure was also associated with lower fractional anisotropy, but in contrast to Herting and colleagues, they also observed similar changes with prenatal and early childhood PM2.5 exposure (Guxens et al. 2022; Lubczynska et al. 2020). Thus, emerging evidence suggests that air pollution can alter white matter microstructure, but the specific effects and long-term implications are unclear.
Although brain imaging provides an objective biomarker of exposure-related changes, a disadvantage is that it is not linked to disease states or clinical outcomes. Indeed, the observed changes in brain microstructure did not correlate with emotional behavior problems in this study. The Committee also noted that the changes in white matter structure were subtle and only assessed cross-sectionally, and that detection of changes in emotional behavior might require follow-up beyond 1 year.
In contrast to their hypothesis, Herting and colleagues observed a small but apparent protective association between air pollution and emotional behavior. Although other studies have observed sporadic protective associations, the Committee noted that based on toxicological and epidemiological evidence as a whole, it is biologically implausible that air pollution exposure would elicit a protective effect (Lin et al. 2022; Parenteau et al. 2024; Rodulfo-Cardenas et al. 2023) and observed associations might be due to chance or residual bias.
An additional limitation was that caregivers only evaluated the emotional-behavioral assessment. Children ages 9–11 are capable of self-reporting emotional behavior, and they can reliably self-report mental symptoms, particularly internalizing problems such as feelings of depression and anxiety (Achenbach et al. 1987; Riley 2004). For example, a recent study showed that children ages 8–13 self-reported more emotional problems than their caregivers (Caqueo-Urízar et al. 2022). Thus, the emotional behaviors in the current study might have been underestimated.
FUTURE DIRECTIONS
The Committee remarked that this study added to existing literature on air pollution and neurodevelopment, but that there were numerous opportunities for future research. In addition to adding a school-based air pollution exposure assessment and longer follow-up, potentially to capture ongoing subtle changes in neurodevelopment, they underscored the importance of integrating complementary outcomes. Further research on brain connectivity should combine evidence from structural and functional MRI that includes evaluation of brain region sizes and activity, self-reported mental states, and neuropsychological assessments. Such outcome integration might elucidate specific links between objective brain biomarkers and mental or developmental disorders.
CONCLUSIONS
In summary, Herting and colleagues examined whether concurrent and prenatal exposure to outdoor PM2.5 and NO2 was associated with changes in white matter connectivity in the prefrontal cortex and with measures of emotional behavior over a 1-year period in children transitioning to adolescence. They observed alterations in white matter connectivity that varied by pollutant and region of the prefrontal cortex but did not find that emotional behavior was negatively affected. This study adds to the existing body of literature demonstrating that air pollution can alter neurodevelopment, even at levels below current regulatory standards for annual exposure. Future research is needed to integrate complementary outcomes to help link objective brain biomarkers with clinical disorders.
ACKNOWLEDGMENTS
The HEI Review Committee thanks the ad hoc reviewers for their help in evaluating the scientific merit of the Investigators’ Report. The Committee is also grateful to Hanna Boogaard for her oversight of the study, to Eva Tanner for assistance with reviewing the report and preparing its Commentary, to Mary Brennan for editing the report and its Commentary, and to Kristin Eckles and Hope Green for their roles in preparing this Research Report for publication.
Footnotes
* A list of abbreviations and other terms appears at the end of this volume.
REFERENCES
Achenbach TM, McConaughy SH, Howell CT. 1987. Child/adolescent behavioral and emotional problems: Implications of cross-informant correlations for situational specificity. Psychol Bull
101:213–232. https://doi.org/10.1037/0033-2909.101.2.213.
[PubMed] [Google Scholar]
Di Q, Amini H, Shi L, Kloog I, Silvern R, et al. 2020. Assessing NO(2) concentration and model uncertainty with high spatiotemporal resolution across the contiguous United States using ensemble model averaging. Environ Sci Technol
54:1372–1384; https://doi.org/10.1021/acs.est.9b03358.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Fan CC, Marshall A, Smolker H, Gonzalez MR, Tapert SF, et al. 2021. Adolescent Brain Cognitive Development (ABCD) study Linked External Data (LED): Protocol and practices for geocoding and assignment of environmental data. Dev Cogn Neurosci
52:101030; https://doi.org/10.1016/j.dcn.2021.101030.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Guxens M, Lubczyńska MJ, Pérez-Crespo L, Muetzel RL, El Marroun H, et al. 2022. Associations of Air Pollution on the Brain in Children: A Brain Imaging Study. Research Report 209. Boston, MA: Health Effects Institute. [PMC free article] [PubMed] [Google Scholar]
Lin L, Zhan X, Jin C, Liang J, Jing J, Dong G. 2022. The epidemiological evidence linking exposure to ambient particulate matter with neurodevelopmental disorders: A systematic review and meta-analysis. Environ Res
209:112876; https://doi.org/10.1016/j.envres.2022.112876.
[DOI] [PubMed] [Google Scholar]
Lubczynska MJ, Muetzel RL, El Marroun H, Basagana X, Strak M, et al. 2020. Exposure to air pollution during pregnancy and childhood, and white matter microstructure in preadolescents. Environ Health Perspect
128:027005; https://doi.org/10.1289/EHP4709.
[DOI] [PMC free article] [PubMed] [Google Scholar]
Peterson BS, Bansal R, Sawardekar S, Nati C, Elgabalawy ER, et al. 2022. Prenatal exposure to air pollution is associated with altered brain structure, function, and metabolism in childhood. J Child Psychol Psychiatry
63:1316–1331; https://doi.org/10.1111/jcpp.13578.
[DOI] [PubMed] [Google Scholar]
Ribeiro M, Yordanova YK, Noblet V, Herbet G, Ricard D. 2024. White matter tracts and executive functions: A review of causal and correlation evidence. Brain
147:352–371; https://doi.org/10.1093/brain/awad308.
[DOI] [PubMed] [Google Scholar]
Rodulfo-Cardenas R, Ruiz-Sobremazas D, Biosca-Brull J, Cabré M, Blanco J, et al. 2023. The influence of environmental particulate matter exposure during late gestation and early life on the risk of neurodevelopmental disorders: A systematic review of experimental evidences. Environ Res
236(Pt 1):116792; https://doi.org/10.1016/j.envres.2023.116792.
[DOI] [PubMed] [Google Scholar]
United States Environmental Protection Agency (US EPA). 2016. Integrated Science Assessment (ISA) for Oxides of Nitrogen — Health Criteria. EPA/600/R-15/068, 2016. Washington, DC: US EPA. [Google Scholar]
United States Environmental Protection Agency (US EPA). 2019. Integrated Science Assessment for Particulate Matter. Washington, DC: US EPA. [PubMed] [Google Scholar]
United States Environmental Protection Agency (US EPA). 2024b. Integrated Science Assessment (ISA) for Lead (Final Report). EPA/600/R-23/375, 2024. Washington, DC: US EPA. [Google Scholar]
Zablotsky B, Ng AE, Black LI, Blumberg SJ. 2023. Diagnosed Developmental Disabilities in Children Aged 3–17 Years: United States, 2019–2021. NCHS Data Brief, no 473. Hyattsville, MD: National Center for Health Statistics. 2023. Available: https://dx.doi.org/10.15620/cdc:129520. [PubMed] [Google Scholar]
Data used in the preparation of this article were obtained from the Adolescent Brain Cognitive Development (ABCD) Study (https://abcdstudy.org), held in the National Institute of Mental Health (NIMH) Data Archive (NDA). This is a multisite, longitudinal study designed to recruit more than 10,000 children aged 9–10 and follow them over 10 years into early adulthood. The ABCD Study is supported by the National Institutes of Health and additional federal partners under award numbers U01DA041048, U01DA050989, U01DA051016, U01DA041022, U01DA051018, U01DA051037, U01DA050987, U01DA041174, U01DA041106, U01DA041117, U01DA041028, U01DA041134, U01DA050988, U01DA051039, U01DA041156, U01DA041025, U01DA041120, U01DA051038, U01DA041148, U01DA041093, U01DA041089, U24DA041123, U24DA041147. A full list of supporters is available at https://abcdstudy.org/federal-partners.html. A listing of participating sites and a complete listing of the study investigators can be found at https://abcdstudy.org/consortium_members/. ABCD consortium investigators designed and implemented the study and/or provided data, but did not necessarily participate in the analysis or writing of this report. This manuscript reflects the views of the authors and may not reflect the opinions or views of the National Institutes of Health or ABCD consortium investigators.
The ABCD data repository grows and changes over time. The ABCD data used in this report came from NIMH Data Archive Digital Object Identifiers 10.15154/1523041 and 10.15154/8873-zj65.
