Poster abstract for HEI Annual Conference 2023
Air Pollution Exposure and Prefrontal Connectivity and Emotional Outcomes in Early Adolescence
Megan M. Herting1, Elisabeth Burnor1, Hedyeh Ahmadi1, Gabrielle Major1, Elizabeth Romero1, Sandrah P. Eckel1, Kiros Berhane2, Rob McConnell1, Joel Schwartz3, Jiu-Chiuan Chen1, William Gauderman, PhD1
1University of Southern California, Los Angeles, CA, USA; 2Columbia University, New York City, NY, USA; 3Harvard University, Cambridge, Massachusetts, USA
Background. We aim to investigate whether ambient PM2.5 and NO2 exposures at 9-10 years are associated with 1) prefrontal connectivity in youth ages 9-10 and 2) subsequent behavioral health problems and risk for neuropsychiatric disorders at ages 10-11 years.
Methods. The current study uses data from the nationwide Adolescent Brain Cognitive Development (ABCD) study℠ (N=11,873), with cross-sectional diffusion tensor imaging (DTI) at 9-10 years and longitudinal data on emotional health outcomes at 9-10 and 11-12 years. Based on residential addresses at ages 9-10, novel hybrid spatiotemporal exposure models were applied to estimate the annual average ambient exposures to PM2.5 and NO2. Prefrontal connectivity was measured using DTI in white matter tracts that innervate the prefrontal cortex. The Child Behavioral Checklist (CBCL) was used to measure internalizing and externalizing emotional problems. We used a directed acyclic graph to a priori identify a minimal sufficient adjustment set of variables, which included child’s age, sex, race/ethnicity, household income, child’s physical activity and screen time use, distance to roadways, perceived neighborhood quality, population density, and urbanicity for both aims, as well as the addition of MRI manufacturer, head motion, and child handedness as precision variables for MRI outcomes. Mixed-effects models were fit with both PM2.5 and NO2 to adjust for study site and socio-demographics. For DTI outcomes, data were stratified by brain hemisphere. For emotional health outcomes, longitudinal models included interaction terms for pollutant-by-time for both pollutants. All results were corrected for multiple comparisons.
Results. The mean annual exposures to PM2.5 and NO2 across 21 study sites were 7.62 ug/m3 [1.72–15.90 ug/m3] and 18.7 ppb [0.73–37.9 ppb], respectively. Significant negative associations were seen between PM2.5 and mean diffusivity (MD) in four tracts in the left hemisphere, one tract in the right hemisphere, and in the corpus callosum (p<0.05). Negative but nonsignificant associations were also seen between NO2 and MD in a total of six white matter tracts. The estimated exposure effects on CBCL outcomes were small, although PM2.5-by-time was statistically significant for anxious/depressed (p<0.001) and aggressive behavior (p=0.05) and internalizing score (p=0.003), whereas NO2-by-time interactions were statistically significant for internalizing problems (p=0.02).
Conclusions. Our analyses indicate that differences in white matter microstructure in adolescents may be associated with PM2.5 in a hemisphere-specific manner, while longitudinal associations with emotional problems with PM2.5 and NO2 are negligible. These findings suggest that white matter microstructure, but not necessarily behavioral problems, may be associated with air pollution concentrations largely below U.S. national ambient air quality standards. The ultimate findings from our study may provide important epidemiologic data on developmental neurotoxicity at low exposure levels to inform the continuing review and risk assessment for the National Ambient Air Quality Standard.