A multi-resolution ensemble model of three decision-tree-based algorithms to predict daily NO2 concentration in France 2005-2022.

Understanding and managing the health effects of Nitrogen Dioxide (NO2) requires high resolution spatiotemporal exposure maps. Here, we developed a multi-stage multi-resolution ensemble model that predicts daily NO2 concentration across continental France from 2005 to 2022. Innovations of this work include the computation of daily predictions at a 200 m resolution in large urban areas and the use of a spatio-temporal blocking procedure to avoid data leakage and ensure fair performance estimation. Predictions were obtained after three cascading stages of modeling: (1) predicting NO2 total column density from Ozone Monitoring Instrument satellite; (2) predicting daily NO2 concentrations at a 1 km spatial resolution using a large set of potential predictors such as predictions obtained from stage 1, land-cover and road traffic data; and (3) predicting residuals from stage 2 models at a 200 m resolution in large urban areas. The latter two stages used a generalized additive model to ensemble predictions of three decision-tree algorithms (random forest, extreme gradient boosting and categorical boosting). Cross-validated performances of our ensemble models were overall very good, with a ten-fold cross-validated R2 for the 1 km model of 0.83, and of 0.69 for the 200 m model. All three basis learners participated in the ensemble predictions to various degrees depending on time and space. In sum, our multi-stage approach was able to predict daily NO2 concentrations with a relatively low error. Ensembling the predictions maximizes the chance of obtaining accurate values if one basis learner fails in a specific area or at a particular time, by relying on the other learners. To the best of our knowledge, this is the first study aiming to predict NO2 concentrations in France with such a high spatiotemporal resolution, large spatial extent, and long temporal coverage. Exposure estimates are available to investigate NO2 health effects in epidemiological studies.

Racial/ethnic and neighborhood disparities in metals exposure during pregnancy in the Northeastern United States.

Despite the unequal burden of environmental exposures borne by racially minoritized communities, these groups are often underrepresented in public health research. Here, we examined racial/ethnic disparities in exposure to metals among a multi-ethnic sample of pregnant women. The sample included women enrolled in the PRogramming of Intergenerational Stress Mechanisms (PRISM) pregnancy cohort (N = 382). Urinary metal concentrations (arsenic [As], barium [Ba], cadmium [Cd], cesium [Cs], chromium [Cr], lead [Pb], antimony [Sb]) were measured during mid-pregnancy and information on individual- and neighborhood-level characteristics was ascertained during an in-person interview and from publicly available databases, respectively. Linear regression was used to examine individual and neighborhood characteristics in relation to metal concentrations. Black/Black-Hispanic women had Cd, Cr, Pb, and Sb levels that were 142.0%, 10.9%, 35.0%, and 32.1% higher than White, non-Hispanic women, respectively. Likewise, White-Hispanic women had corresponding levels that were 141.5%, 108.2%, 59.9%, and 38.3% higher. These same metals were also higher among women residing in areas with higher crime, higher diversity, lower educational attainment, lower household income, and higher poverty. Significant disparities in exposure to metals exist and may be driven by neighborhood-level factors. Exposure to metals for pregnant women can be especially harmful. Understanding exposure inequalities and identifying factors that increase risk can help inform targeted public health interventions.

Validating and Comparing Highly Resolved Commercial "Off the Shelf" PM Monitoring Sensors with Satellite Based Hybrid Models, for Improved Environmental Exposure Assessment.

Particulate matter is a common health hazard, and under certain conditions, an ecological threat. While many studies were conducted in regard to air pollution and potential effects, this paper serves as a pilot scale investigation into the spatial and temporal variability of particulate matter (PM) pollution in arid urban environments in general, and Beer-Sheva, Israel as a case study. We explore the use of commercially off the shelf (COTS) sensors, which provide an economical solution for spatio-temporal measurements. We started with a comparison process against an A-grade meteorological station, where it was shown that under specific climatic conditions, a number of COTS sensors were able to produce robust agreement (mean R2=0.93, average SD=17.5). The second stage examined the COTS sensors that were proven accurate in a mobile measurement campaign. Finally, data collected was compared to a validated satellite prediction model. We present how these tests and COTS sensor-kits could then be used to further explain the continuity and dispersion of particulate matter in similar areas.

Spatiotemporal distribution of autism spectrum disorder prevalence among birth cohorts during 2000-2011 in Israel.

PURPOSE: Studies indicate an apparent sharp increasing trend in autism spectrum disorder (ASD) incidence and prevalence worldwide. This nationwide study aims at depicting ASD prevalence distribution in Israel in both space and time. METHODS: Based on data from Israel National Insurance Institute, the study population included all children born in Israel 2000-2011 (n = 1,786,194), of whom 11,699 (0.655%) were subsequently diagnosed with ASD (until December 31, 2016). Prevalence was calculated and mapped by dividing the number of ASD cases within each year of birth by the number of births during that year, for each spatial unit, and similarly for several spatiotemporal levels of aggregation. RESULTS: ASD prevalence varies substantially across different geographic areas in Israel, with considerably higher prevalence concentrated in central Israel. Strong associations were found between locality-level socioeconomic index, ethnicity, and peripherality and ASD prevalence, and even after adjustment for them, excess prevalence for ASD still persisted in certain localities. No spatial dependence of prevalence, with and without adjustment for the locality-level variables, was found (Moran's I = -0.000546, -0.00335, respectively). CONCLUSIONS: Our findings provide important insights regarding health disparities affecting ASD diagnosis, directing further health policy intervention and further research.

Ensemble averaging based assessment of spatiotemporal variations in ambient PM2.5 concentrations over Delhi, India, during 2010-2016.

Elevated levels of ambient air pollution has been implicated as a major risk factor for morbidities and premature mortality in India, with particularly high concentrations of particulate matter in the Indo-Gangetic plain. High resolution spatiotemporal estimates of such exposures are critical to assess health effects at an individual level. This article retrospectively assesses daily average PM2.5 exposure at 1 km × 1 km grids in Delhi, India from 2010-2016, using multiple data sources and ensemble averaging approaches. We used a multi-stage modeling exercise involving satellite data, land use variables, reanalysis based meteorological variables and population density. A calibration regression was used to model PM2.5: PM10 to counter the sparsity of ground monitoring data. The relationship between PM2.5 and its spatiotemporal predictors was modeled using six learners; generalized additive models, elastic net, support vector regressions, random forests, neural networks and extreme gradient boosting. Subsequently, these predictions were combined under a generalized additive model framework using a tensor product based spatial smoothing. Overall cross-validated prediction accuracy of the model was 80% over the study period with high spatial model accuracy and predicted annual average concentrations ranging from 87 to 138 µg/m3. Annual average root mean squared errors for the ensemble averaged predictions were in the range 39.7-62.7 µg/m3 with prediction bias ranging between 4.6-11.2 µg/m3. In addition, tree based learners such as random forests and extreme gradient boosting outperformed other algorithms. Our findings indicate important seasonal and geographical differences in particulate matter concentrations within Delhi over a significant period of time, with meteorological and land use features that discriminate most and least polluted regions. This exposure assessment can be used to estimate dose response relationships more accurately over a wide range of particulate matter concentrations.

Racial Disparities in Associations between Neighborhood Demographic Polarization and Birth Weight.

Neighborhood demographic polarization, or the extent to which a privileged population group outnumbers a deprived group, can affect health by influencing social dynamics. While using birth records from 2001 to 2013 in Massachusetts (n = 629,675), we estimated the effect of two demographic indices, racial residential polarization (RRP) and economic residential polarization (ERP), on birth weight outcomes, which are established predictors of the newborn's future morbidity and mortality risk. Higher RRP and ERP was each associated with higher continuous birth weight and lower odds for low birth weight and small for gestational age, with evidence for effect modification by maternal race. On average, per interquartile range increase in RRP, the birth weight was 10.0 g (95% confidence interval: 8.0, 12.0) higher among babies born to white mothers versus 6.9 g (95% CI: 4.8, 9.0) higher among those born to black mothers. For ERP, it was 18.6 g (95% CI: 15.7, 21.5) higher among those that were born to white mothers versus 1.8 g (95% CI: -4.2, 7.8) higher among those born to black mothers. Racial and economic polarization towards more privileged groups was associated with healthier birth weight outcomes, with greater estimated effects in babies that were born to white mothers than those born to black mothers.

Effects of Maternal Homelessness, Supplemental Nutrition Programs, and Prenatal PM2.5 on Birthweight.

Few studies examined the impact of maternal socioeconomic status and of its combined effects with environmental exposures on birthweight. Our goal was to examine the impact of maternal homelessness (mothers ever homeless or who lived in shelters during pregnancy) and participation in the Special Supplemental Nutrition Program for Women, Infants and Children (WIC) during pregnancy in conjunction with air pollution exposure on birthweight in the Boston-based Children's HealthWatch cohort from 2007 through 2015 (n = 3366). Birthweight was obtained from electronic health records. Information on maternal homelessness and WIC participation during pregnancy were provided via a questionnaire. Prenatal fine particulate matter (PM2.5) exposures, estimated at the subject's residential address, were calculated for each trimester. We fit linear regression models adjusting for maternal and child characteristics, seasonality, and block-group-level median household income and examined the interactions between PM2.5 and each covariate. Prenatal maternal homelessness was associated with reduced birthweight (-55.7 g, 95% CI: -97.8 g, -13.7 g), while participating in WIC was marginally associated with increased birthweight (36.1 g, 95% CI: -7.3 g, 79.4 g). Only average PM2.5 during the second trimester was marginally associated with reduced birthweight (-8.5 g, 95% CI: -19.3, 2.3) for a 1 µg/m3 increase in PM2.5. The association of PM2.5 during the second trimester with reduced birthweight was stronger among non-Hispanic Black mothers and trended toward significance among immigrants and single mothers. Our study emphasizes the independent and synergistic effects of social and environmental stressors on birthweight, particularly the potentially protective effect of participating in WIC for vulnerable populations.

Fine Particulate Air Pollution and Birthweight: Differences in Associations Along the Birthweight Distribution.

BACKGROUND: Maternal exposure to fine particulate air pollution (PM2.5) during pregnancy is associated with lower newborn birthweight, which is a risk factor for chronic disease. Existing studies typically report the average association related with PM2.5 increase, which does not offer information about potentially varying associations at different points of the birthweight distribution. METHODS: We retrieved all birth records in Massachusetts between 2001 and 2013 then restricted our analysis to full-term live singletons (n = 775,768). Using the birthdate, gestational age, and residential address reported at time of birth, we estimated the average maternal PM2.5 exposure during pregnancy of each birth. PM2.5 predictions came from a model that incorporates satellite, land use, and meteorologic data. We applied quantile regression to quantify the association between PM2.5 and birthweight at each decile of birthweight, adjusted for individual and neighborhood covariates. We considered effect modification by indicators of individual and neighborhood socioeconomic status (SES). RESULTS: PM2.5 was negatively associated with birthweight. An interquartile range increase in PM2.5 was associated with a 16 g [95% confidence interval (CI) = 13, 19] lower birthweight on average, 19 g (95% CI = 15, 23) lower birthweight at the lowest decile of birthweight, and 14 g (95% CI = 9, 19) lower birthweight at the highest decile. In general, the magnitudes of negative associations were larger at lower deciles. We did not find evidence of effect modification by individual or neighborhood SES. CONCLUSIONS: In full-term live births, PM2.5 and birthweight were negatively associated with more severe associations at lower quantiles of birthweight.

County-level radon exposure and all-cause mortality risk among Medicare beneficiaries.

BACKGROUND: Radon is an inert gas formed from the decay of naturally-occurring materials in the earth's crust. It infiltrates into homes from soil, water, and construction materials. Its decay products are radionuclides, which attach to ambient particles. Residential radon is one of the leading risk factors for lung cancer. The scarce evidence for associations with other mortality causes originates mostly from occupational studies. METHODS: In a cohort study with 14 years of follow-up (2000-2013), we evaluated the association between chronic radon exposure and all-cause mortality, and explored whether there are subpopulations who are more vulnerable to radon effects. We included 87,296,195 person-years of follow-up from all Medicare beneficiaries in the Mid-Atlantic and Northeastern U.S. states. We examined the association between the logarithm of county-averaged radon (ln(Rn)) and mortality and assessed effect modification by chronic conditions. RESULTS: An interquartile range increase in the ln(Rn) was associated with a 2·62% increase (95% CI 2·52%; 2·73%) in mortality, independent of PM2.5 exposure. Larger mortality risks were observed among individuals with respiratory, cardiovascular and metabolic diseases, with the highest associations observed among those with diabetes (4·98% increase), heart failure (4·58% increase), and chronic obstructive pulmonary disease (4·49% increase). CONCLUSION: We found an increased risk for all-cause mortality associated with increased radon exposure. The risk was enhanced among susceptible individuals with chronic conditions. We believe this is the first cohort study to identify populations at higher risk for non-malignant health consequences of radon exposure. Due to the limitations in exposure assessment and availability of individual confounders, these findings should be interpreted with caution.

Neighborhood Greenness Attenuates the Adverse Effect of PM2.5 on Cardiovascular Mortality in Neighborhoods of Lower Socioeconomic Status.

Features of the environment may modify the effect of particulate matter ≤2.5 µm in aerodynamic diameter (PM2.5) on health. Therefore, we investigated how neighborhood sociodemographic and land-use characteristics may modify the association between PM2.5 and cardiovascular mortality. We obtained residence-level geocoded cardiovascular mortality cases from the Massachusetts Department of Public Health (n = 179,986), and PM2.5 predictions from a satellite-based model (2001⁻2011). We appended census block group-level information on sociodemographic factors and walkability, and calculated neighborhood greenness within a 250 m buffer surrounding each residence. We found a 2.54% (1.34%; 3.74%) increase in cardiovascular mortality associated with a 10 µg/m³ increase in two-day average PM2.5. Walkability or greenness did not modify the association. However, when stratifying by neighborhood sociodemographic characteristics, smaller PM2.5 effects were observed in greener areas only among cases who resided in neighborhoods with a higher population density and lower percentages of white residents or residents with a high school diploma. In conclusion, the PM2.5 effects on cardiovascular mortality were attenuated by higher greenness only in areas with sociodemographic features that are highly correlated with lower socioeconomic status. Previous evidence suggests health benefits linked to neighborhood greenness may be stronger among lower socioeconomic groups. Attenuation of the PM2.5⁻mortality relationship due to greenness may explain some of this evidence.

Association between prenatal particulate air pollution exposure and telomere length in cord blood: Effect modification by fetal sex.

INTRODUCTION: In utero particulate matter exposure produces oxidative stress that impacts cellular processes that include telomere biology. Newborn telomere length is likely critical to an individual's telomere biology; reduction in this initial telomere setting may signal increased susceptibility to adverse outcomes later in life. We examined associations between prenatal particulate matter with diameter ≤2.5 µm (PM2.5) and relative leukocyte telomere length (LTL) measured in cord blood using a data-driven approach to characterize sensitive windows of prenatal PM2.5 effects and explore sex differences. METHODS: Women who were residents of Mexico City and affiliated with the Mexican Social Security System were recruited during pregnancy (n = 423 for analyses). Mothers' prenatal exposure to PM2.5 was estimated based on residence during pregnancy using a validated satellite-based spatio-temporally resolved prediction model. Leukocyte DNA was extracted from cord blood obtained at delivery. Duplex quantitative polymerase chain reaction was used to compare the relative amplification of the telomere repeat copy number to single gene (albumin) copy number. A distributed lag model incorporating weekly averages for PM2.5 over gestation was used in order to explore sensitive windows. Sex-specific associations were examined using Bayesian distributed lag interaction models. RESULTS: In models that included child's sex, mother's age at delivery, prenatal environmental tobacco smoke exposure, pre-pregnancy BMI, gestational age, birth season and assay batch, we found significant associations between higher PM2.5 exposure during early pregnancy (4-9 weeks) and shorter LTL in cord blood. We also identified two more windows at 14-19 and 34-36 weeks in which increased PM2.5 exposure was associated with longer LTL. In stratified analyses, the mean and cumulative associations between PM2.5 and shortened LTL were stronger in girls when compared to boys. CONCLUSIONS: Increased PM2.5 during specific prenatal windows was associated with shorter LTL and longer LTL. PM2.5 was more strongly associated with shortened LTL in girls when compared to boys. Understanding sex and temporal differences in response to air pollution may provide unique insight into mechanisms.

Use of earth observations for temperature exposure assessment in epidemiological studies.

PURPOSE OF REVIEW: Low, high, extreme, and variable temperatures have been linked to multiple adverse health outcomes, particularly among the elderly and children. Recent models incorporating satellite remote sensing data have mitigated several limitations of previous studies, improving exposure assessment. This review focuses on these new temperature exposure models and their application in epidemiological studies. RECENT FINDINGS: Satellite observations of land surface temperature have been used to model air temperature across large spatial areas at high spatiotemporal resolutions. These models enable exposure assessment of entire populations and have been shown to reduce error in exposure estimates, thus mitigating downward bias in health effect estimates. SUMMARY: Satellite-based models improve our understanding of spatiotemporal variation in temperature and the associated health effects. Further research should focus on improving the resolution of these models, especially in urban areas, and increasing their use in epidemiological studies of direct temperature exposure and vector-borne diseases.

Using Open Street Map Data in Environmental Exposure Assessment Studies: Eastern Massachusetts, Bern Region, and South Israel as a Case Study.

There is an increase in the awareness of the importance of spatial data in epidemiology and exposure assessment (EA) studies. Most studies use governmental and ordnance surveys, which are often expensive and sparsely updated, while in most developing countries, there are often no official geo-spatial data sources. OpenStreetMap (OSM) is an open source Volunteered Geographic Information (VGI) mapping project. Yet very few environmental epidemiological and EA studies have used OSM as a source for road data. Since VGI data is either noncommercial or governmental, the validity of OSM is often questioned. We investigate the robustness and validity of OSM data for use in epidemiological and EA studies. We compared OSM and Governmental Major Road Data (GRD) in three different regions: Massachusetts, USA; Bern, Switzerland; and Beer-Sheva, South Israel. The comparison was done by calculating data completeness, positional accuracy, and EA using traditional exposure methods. We found that OSM data is fairly complete and accurate in all regions. The results in all regions were robust, with Massachusetts showing the best fits (R² 0.93). Results in Bern (R² 0.78) and Beer-Sheva (R² 0.77) were only slightly lower. We conclude by suggesting that OSM data can be used reliably in environmental assessment studies.

Association between particulate air pollution exposure during pregnancy and postpartum maternal psychological functioning.

METHODS: We studied associations between prenatal exposure to particulate matter with diameter ≤ 2.5 µm (PM2.5) and postpartum psychological functioning in a lower income, ethnically mixed sample of urban US women enrolled in a pregnancy cohort study. Analyses included 557 mothers who delivered at ≥37 weeks gestation. Daily estimates of residential PM2.5 over gestation were derived using a satellite-based spatio-temporally resolved model. Outcomes included the Edinburgh Postnatal Depression Scale (EPDS) score from 6 or 12 months postpartum and subscale scores for anhedonia, depressive and anxiety symptoms. Associations were also examined within racial/ethnic groups. Distributed lag models (DLMs) were implemented to identify windows of vulnerability during pregnancy. RESULTS: Most mothers had less than a high school education (64%) and were primarily Hispanic (55%) and Black (29%). In the overall sample, a DLM adjusted for age, race, education, prenatal smoking, and season of delivery, we found significant associations between higher PM2.5 exposure in the second trimester and increased anhedonia subscale scores postpartum. In race stratified analyses, mid-pregnancy PM2.5 exposure was significantly associated with increased total EPDS scores as well as higher anhedonia and depressive symptom subscale scores among Black women. CONCLUSIONS: Increased PM2.5 exposure in mid-pregnancy was associated with increased depressive and anhedonia symptoms, particularly in Black women.

Estimating the Effects of PM2.5 on Life Expectancy Using Causal Modeling Methods.

BACKGROUND: Many cohort studies have reported associations between PM2.5 and the hazard of dying, but few have used formal causal modeling methods, estimated marginal effects, or directly modeled the loss of life expectancy. OBJECTIVE: Our goal was to directly estimate the effect of PM2.5 on the distribution of life span using causal modeling techniques. METHODS: We derived nonparametric estimates of the distribution of life expectancy as a function of PM2.5 using data from 16,965,154 Medicare beneficiaries in the Northeastern and mid-Atlantic region states (129,341,959 person-years of follow-up and 6,334,905 deaths). We fit separate inverse probability-weighted logistic regressions for each year of age to estimate the risk of dying at that age given the average PM2.5 concentration at each subject's residence ZIP code in the same year, and we used Monte Carlo simulations to estimate confidence intervals. RESULTS: The estimated mean age at death for a population with an annual average PM2.5 exposure of 12 µg/m3 (the 2012 National Ambient Air Quality Standard) was 0.89 y less (95% CI: 0.88, 0.91) than estimated for a counterfactual PM2.5 exposure of 7.5 µg/m3. In comparison, life expectancy at 65 y of age increased by 0.9 y between 2004 and 2013 in the United States. We estimated that 23.5% of the Medicare population would die before 76 y of age if exposed to PM2.5 at 12 µg/m3 compared with 20.1% if exposed to an annual average of 7.5 µg/m3. CONCLUSIONS: We believe that this is the first study to directly estimate the effect of PM2.5 on the distribution of age at death using causal modeling techniques to control for confounding. We find that reducing PM2.5 concentrations below the 2012 U.S. annual standard would substantially increase life expectancy in the Medicare population. https://doi.org/10.1289/EHP3130.

Assessment of indoor and outdoor particulate air pollution at an urban background site in Iran.

The relationship between indoor and outdoor particulate air pollution was investigated at an urban background site on the Payambar Azam Campus of Mazandaran University of Medical Sciences in Sari, Northern Iran. The concentration of particulate matter sized with a diameter less than 1 µm (PM1.0), 2.5 µm (PM2.5), and 10 µm (PM10) was evaluated at 5 outdoor and 12 indoor locations. Indoor sites included classrooms, corridors, and office sites in four university buildings. Outdoor PM concentrations were characterized at five locations around the university campus. Indoor and outdoor PM measurements (1-min resolution) were conducted in parallel during weekday mornings and afternoons. No difference found between indoor PM10 (50.1 ± 32.1 µg/m3) and outdoor PM10 concentrations (46.5 ± 26.0 µg/m3), indoor PM2.5 (22.6 ± 17.4 µg/m3) and outdoor PM2.5 concentration (22.2 ± 15.4 µg/m3), or indoor PM1.0 (14.5 ± 13.4 µg/m3) and outdoor mean PM1.0 concentrations (14.2 ± 12.3 µg/m3). Despite these similar concentrations, no correlations were found between outdoor and indoor PM levels. The present findings are not only of importance for the potential health effects of particulate air pollution on people who spend their daytime over a period of several hours in closed and confined spaces located at a university campus but also can inform regulatory about the improvement of indoor air quality, especially in developing countries.

Identifying sensitive windows for prenatal particulate air pollution exposure and mitochondrial DNA content in cord blood.

INTRODUCTION: Changes in mitochondrial DNA (mtDNA) can serve as a marker of cumulative oxidative stress (OS) due to the mitochondria's unique genome and relative lack of repair systems. In utero particulate matter ≤2.5µm (PM2.5) exposure can enhance oxidative stress. Our objective was to identify sensitive windows to predict mtDNA damage experienced in the prenatal period due to PM2.5 exposure using mtDNA content measured in cord blood. MATERIAL AND METHODS: Women affiliated with the Mexican social security system were recruited during pregnancy in the Programming Research in Obesity, Growth, Environment and Social Stressors (PROGRESS) study. Mothers with cord blood collected at delivery and complete covariate data were included (n=456). Mothers' prenatal daily exposure to PM2.5 was estimated using a satellite-based spatio-temporally resolved prediction model and place of residence during pregnancy. DNA was extracted from umbilical cord leukocytes. Quantitative real-time polymerase chain reaction (qPCR) was used to determine mtDNA content. A distributive lag regression model (DLM) incorporating weekly averages of daily PM2.5 predictions was constructed to plot the association between exposure and OS over the length of pregnancy. RESULTS: In models that included child's sex, mother's age at delivery, prenatal environmental tobacco smoke exposure, birth year, maternal education, and assay batch, we found significant associations between higher PM2.5 exposure during late pregnancy (35-40weeks) and lower mtDNA content in cord blood. CONCLUSIONS: Increased PM2.5 during a specific prenatal window in the third trimester was associated with decreased mtDNA content suggesting heightened sensitivity to PM-induced OS during this life stage.

Long-term exposure to fine particulate matter, residential proximity to major roads and measures of brain structure.

BACKGROUND AND PURPOSE: Long-term exposure to ambient air pollution is associated with cerebrovascular disease and cognitive impairment, but whether it is related to structural changes in the brain is not clear. We examined the associations between residential long-term exposure to ambient air pollution and markers of brain aging using magnetic resonance imaging. METHODS: Framingham Offspring Study participants who attended the seventh examination were at least 60 years old and free of dementia and stroke were included. We evaluated associations between exposures (fine particulate matter [PM2.5] and residential proximity to major roadways) and measures of total cerebral brain volume, hippocampal volume, white matter hyperintensity volume (log-transformed and extensive white matter hyperintensity volume for age), and covert brain infarcts. Models were adjusted for age, clinical covariates, indicators of socioeconomic position, and temporal trends. RESULTS: A 2-µg/m(3) increase in PM2.5 was associated with -0.32% (95% confidence interval, -0.59 to -0.05) smaller total cerebral brain volume and 1.46 (95% confidence interval, 1.10 to 1.94) higher odds of covert brain infarcts. Living further away from a major roadway was associated with 0.10 (95% confidence interval, 0.01 to 0.19) greater log-transformed white matter hyperintensity volume for an interquartile range difference in distance, but no clear pattern of association was observed for extensive white matter. CONCLUSIONS: Exposure to elevated levels of PM2.5 was associated with smaller total cerebral brain volume, a marker of age-associated brain atrophy, and with higher odds of covert brain infarcts. These findings suggest that air pollution is associated with insidious effects on structural brain aging even in dementia- and stroke-free persons.