Air pollution is linked to divergent cortical thickness patterns in brain regions vulnerable to Alzheimer’s disease

Lauren E Salminen; Xinhui Wang; Andrew J Petkus; Christine Fennema-Notestine; Jeremy A Elman; Donald J Hagler, Jr; Joel D Kaufman; Joshua Millstein; Meredith N Braskie; Joshua Liu; Hyung Jin Cho; Deydeep Kothapalli; Paul M Thompson; Daniel P Beavers; Mark A Espeland; Margaret Gatz; Susan M Resnick; Ira Frahmand; Stephen R Rapp; Caleb E Finch; Jiu-Chiuan Chen; William S Kremen; Carol E Franz

doi:10.1016/j.neuro.2026.103495

. Author manuscript; available in PMC: 2026 Sep 30.

Published in final edited form as: Neurotoxicology. 2026 Jun 19;115:103495. doi: 10.1016/j.neuro.2026.103495

Air pollution is linked to divergent cortical thickness patterns in brain regions vulnerable to Alzheimer’s disease

Lauren E Salminen ^a,^b,^*,¹, Xinhui Wang ^b,^**,¹, Andrew J Petkus ^b, Christine Fennema-Notestine ^c,^d,^e, Jeremy A Elman ^c,^e, Donald J Hagler Jr ^d,^f, Joel D Kaufman ^g, Joshua Millstein ^h, Meredith N Braskie ^a,^b, Joshua Liu ^a,^b, Hyung Jin Cho ^a,^b, Deydeep Kothapalli ^a,^b, Paul M Thompson ^a,^b, Daniel P Beavers ⁱ, Mark A Espeland ^j,^k, Margaret Gatz ^l, Susan M Resnick ^m, Ira Frahmand ⁿ, Stephen R Rapp ^o, Caleb E Finch ^p, Jiu-Chiuan Chen ^b,^h, William S Kremen ^c,^e, Carol E Franz ^c,^e

^a Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, 4676 Admiralty Way, Floor 2, Suite 246, Marina Del Rey, CA 90292, USA

^b Department of Neurology, University of Southern California, 1540 Alcazar Street, Suite 215, Los Angeles, CA 90033, USA

^c Department of Psychiatry, University of California San Diego, 9500 Gilman Drive, La Jolla, CA 92093, USA

^d Department of Radiology, University of California San Diego, La Jolla, 8929 University Center Ln, Ste 101, San Diego, CA 92122, USA

^e Center for Behavior Genetics of Aging, University of California San Diego, 9500 Gilman Drive, Mail Code 0664, La Jolla, CA 92093, USA

^f Center for Multimodal Imaging and Genetics, University of California San Diego, 9500 Gilman Drive, Mail Code 0841, La Jolla, CA 92093, USA

^g Departments of Environmental & Occupational Health Sciences, Medicine, and Epidemiology, University of Washington, 3980 15th Ave NE. Second floor, Seattle, WA 98195, USA

^h Department of Population and Public Health Sciences, University of Southern California, 1845 N Soto St, Los Angeles, CA 90032, USA

ⁱ Departments of Statistical Sciences, Wake Forest University, Winston-Salem, NC 27101, USA

^j Department of Biostatistics and Data Sciences, Wake Forest School of Medicine, 475 Vine Street, Winston-Salem, NC 27101, USA

^k Department of Internal Medicine, Wake Forest School of Medicine, 475 Vine Street, Winston-Salem, NC 27101, USA

^l Center for Economic and Social Research, University of Southern California, 635 Downey Way, VPD, Los Angeles, CA 90089, USA

^m Laboratory of Behavioral Neuroscience, National Institute on Aging, National Institutes of Health, 251 Bayview Boulevard, Baltimore, MD 21224, USA

ⁿ Department of Medicine, University of Wisconsin - Madison, 1685 Highland Ave, Madison, WI 53705, USA

^o Departments of Psychiatry & Behavioral Medicine, Social Sciences & Health Policy, Wake Forest School of Medicine, 791 Jonestown Road, Winston-Salem, NC 27103, USA

^p Andrus Gerontology Center, University of Southern California, 3715 McClintock Ave, Los Angeles, CA 90089, USA

Authors contributed equally to this work

Corresponding author at: Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, 4676 Admiralty Way, Floor 2, Suite 246, Marina Del Rey, CA 90292, USA. salminen@usc.edu (L.E. Salminen)

^**

Correspondence to: Department of Neurology, Keck School of Medicine, University of Southern California, 1540 Alcazar Street, Suite 215, Los Angeles, CA 90033, USA. xinhuiwa@usc.edu (X. Wang).

PMCID: PMC13620171 NIHMSID: NIHMS2198545 PMID: 42314912

Abstract

Introduction:

Ambient air pollution increases Alzheimer’s disease (AD) risk, yet exposure associations with cortical thickness (CTh) in AD-vulnerable brain regions is unclear. Here we examined the associations between PM_2.5 and NO₂ with CTh in an AD meta-region of interest (ROI), and across the cortex, in participants from the Vietnam Era Twin Study of Aging (VETSA) and Women’s Health Initiative Memory Study (WHIMS).

Methods:

We conducted a cross-sectional study using data from 387 VETSA men (M_age=61.9 ±2.6) and 1097 WHIMS women (M_age=77.9 ±3.7) without dementia or stroke prior to MRI. Long-term residential exposures to PM_2.5 and NO₂ were quantified as the 3-year average of monthly estimates prior to MRI that were derived from spatiotemporal models with regionalized universal kriging. Brain MRI scans were processed using FreeSurfer-v.5.3.0 to estimate CTh in 34 bilateral regions parcellated with the Desikan-Killiany atlas. An AD meta-ROI was calculated as the surface-area weighted average of CTh in four bilateral regions (entorhinal, fusiform, inferior temporal, and middle temporal cortices) that are vulnerable to AD. Linear mixed models were conducted separately in each cohort with appropriate covariates.

Results:

In WHIMS, exposures were negatively associated with CTh in the AD meta-ROI (p_PM2.5<0.001; p_NO2=0.018) and diffusely across the cortex. In VETSA, exposures were positively associated with CTh in the AD meta-ROI (p_PM2.5=0.017; p_NO2=0.021) and temporal pole, but effects were age-dependent, becoming negative (though nonsignificant) after age 64.

Discussion:

Positive associations in younger VETSA men, coupled with negative associations in older WHIMS women, may suggest nonmonotonic AD-related neurodegeneration.

Keywords: Air pollution, Structural MRI, Cortical thickness, Brain aging, Alzheimer’s disease

1. Background

Outdoor air pollutants such as fine particulate matter (<2.5μm; PM_2.5) and traffic-related pollution (e.g., nitrogen dioxide, NO₂) are a global health concern due to their ubiquitous presence and detrimental health impact, especially on the cardiovascular and respiratory systems. Beyond these well-established adverse effects, long-term air pollution exposure during late life is now recognized as an important environmental risk factor for Alzheimer’s disease (AD) and related dementias (ADRD)(Livingston et al., 2020). AD is the most common form of dementia, and is pathologically characterized by the abnormal accumulation of amyloid-beta (Aβ) and tau proteins, and progressive neuronal loss that typically begins in medial temporal lobe brain structures and leads to cognitive decline and functional impairment (Jack et al., 2018). Many studies have established associations between late-life air pollution exposure and cognitive difficulties (Chandra et al., 2022; Kulick et al., 2020; Ma et al., 2022), prompting neuroimaging investigations of air pollution exposure in older adults who may be at risk for ADRD. These studies revealed smaller volumes across the whole brain and in medial temporal and frontal lobe structures in individuals exposed to higher air pollution levels (Calderón-Garcidueñas et al., 2026; Casanova et al., 2016; Erickson et al., 2020; Nußbaum et al., 2020; Petkus et al., 2023; Power et al., 2018; Wang et al., 2025; Younan et al., 2020), suggesting that air pollution contributes to neurodegenerative processes observed in AD.

However, most structural MRI studies of air pollution exposure focused on brain volumetrics, which are less mechanistically informative in the early stages of disease. Cortical thickness (CTh) measured from structural MRI is an alternative metric to evaluate gray matter integrity that is more sensitive to subtle brain changes that occur during normal aging and preclinical AD. Although CTh is typically inversely associated with age, increasing evidence shows biphasic cortical trajectories along the continuum from normal cognitive aging to AD, such that CTh increases during early preclinical disease stages and subsequently declines as disease progresses (Montal et al., 2018, 2021; Torso et al., 2022; Williams et al., 2023). To date, four studies have assessed the relationship between air pollution exposures and regional CTh (Cho et al., 2020; Crous-Bou et al., 2020; Ko et al., 2024; Lin et al., 2024), and three focused on AD-vulnerable regions to determine whether exposures contributed to a cortical pattern of AD-related neurodegeneration (Cho et al., 2023; Crous-Bou et al., 2020; Lin et al., 2024). All studies were conducted outside of the United States, in areas with significantly higher air pollution levels. Further, only one of the studies examining exposure effects on AD-related neurodegeneration was conducted using a cross-sectional design (Crous-Bou et al., 2020). The latter study was conducted in a subsample of the Alzheimer’s and Family (ALFA) project and revealed nominal negative associations between NO2 exposures (but not PM_2.5) and an AD meta-ROI consisting of CTh in the entorhinal cortex (ERC), middle temporal gyrus (MTG), inferior temporal gyrus (ITG), fusiform, posterior cingulate (PCC), precuneus, and supramarginal gyrus (SMG). However, participants in the ALFA study were on average less than 50 years of age, and few psychosocial variables were incorporated into their statistical models. As such, the degree to which low-level air pollution exposures confer risk for abnormal CTh in AD-vulnerable regions and across the cortex in cognitively unimpaired older adults remains unclear.

Here we addressed this gap through an NIH-funded collaborative initiative between the Vietnam Era Twin Study of Aging (VETSA) and the Women’s Health Initiative (WHI) Memory Study (WHIMS) cohorts. Our goal was to understand relationships between PM_2.5 and NO₂ exposures with CTh in AD-vulnerable regions, and regionally across the whole brain, separately in each cohort to identify shared MRI signatures of air pollution neurotoxicity across independent samples of community dwelling individuals without dementia. We hypothesized that higher exposure levels would be associated with lower CTh in AD-vulnerable regions, with stronger and more widespread associations in the WHIMS cohort due to the older age of participants.

2. Methods

2.1. Design overview

The collaborative initiative between VETSA and WHIMS was designed to investigate the effects of ambient air pollution exposure on the aging brain in men and women. As described in the methods below, exposure and neuroimaging estimates were conducted harmoniously across cohorts (described in Sections 2.4 through 2.6), but data were analyzed independently within each cohort. Great care was taken into creating and evaluating comparable covariate measures to minimize noise and maximize cross-study interpretations. Procedures were approved by the Institutional Review Boards at the University of Southern California (USC) and the University of California San Diego (UCSD) - the prime analytic sites for WHIMS and VETSA, respectively. Written informed consent was obtained from all participants.

2.2. VETSA overview, participants, and covariates

Participants were randomly recruited from the Vietnam Era Twin Registry (VETR), who participated in the Harvard Twin Study of Substance Abuse (Tsuang et al., 2001). VETR is a large nationally distributed registry of male twin pairs who served in the United States military at some point between 1965 and 1975 (Eisen et al., 1987; Henderson et al., 1990). All VETR members were eligible to participate in the Harvard study and substance use or any related characteristic were not used as inclusion criteria. VETSA inclusion criteria at baseline (VETSA–1; 2002–2008) included being ages 50–59 years at enrollment and both members of a twin pair agreeing to participate (Kremen et al., 2006). VETSA−2 participants included VETSA–1 participants assessed approximately 6 years after VETSA–1; as well as age-matched attrition replacements (ages 56–67, M_age=62) (Kremen et al., 2019). Details of sample ascertainment and data collections are described elsewhere (Kremen et al., 2006, 2019). VETSA participants are a community-dwelling, non-patient sample. Most (80%) did not experience combat and have comparable health, education, and lifestyle characteristics to American men in their age cohort (Schoenborn and Heyman, 2009). In the present study we examined only VETSA–2 participants to reduce the age gap between the VETSA and WHIMS cohorts.

VETSA in-person assessments involved self-report questionnaires, medical history interviews, in-depth neuropsychological testing, and MRI. Except for MRI protocols, assessments using identical protocols occurred at UCSD and Boston University (BU). Neuroimaging for BU participants was conducted at Massachusetts General Hospital (MGH) and almost always administered the day after testing and interviewing.

Sociodemographic variables including region, age, race/ethnicity, lifetime education, household income, and employment status were collected using self-report questionnaires. Lifestyle factors (e.g., smoking status, alcohol consumption, and physical activity) were collected via in-person interviews. As this is a predominantly white population, we combined “Hispanic/Latino”, “American Indian/Alaska Native”, “Asian”, “Black”, or “More than one race” into “Other ethnic or racial background”. Alcohol consumption was coded for consumption of beer, wine, and/or hard liquor in the past two weeks (none; 1–2 drinks per day; 2 + drinks per day). Physical activity was the sum of four items from the self-reported Schaie activity scale (Schaie, 1983), including frequency of participating in sports, physical fitness, walking and hiking, and dancing. The sum was re-coded to reflect activity frequency per week (none; some; 2–4 times; >4 times). Body Mass Index (BMI) was calculated using measured height and weight during the neuropsychological evaluation. Depressive symptoms were assessed using the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977). Health information on physician diagnosis of cardiovascular conditions, stroke, diabetes, and hypertension was self-reported (present/absent) in a medical history interview on the same day as cognitive testing. The Charlson comorbidity score (Charlson et al., 1994) was the count of 15 major chronic conditions: diabetes, emphysema, asthma, cancer, osteoarthritis, rheumatoid arthritis, stroke, heart attack, heart failure, heart surgery, angina, hypertension, peripheral vascular disease, cirrhosis, and AIDS. For VETSA, the Neighborhood Atlas area disadvantage coding system was applied to residential address data to create a measure of neighborhood socioeconomic status (SES) (https://www.neighborhoodatlas.medicine.wisc.edu/) (Kind and Buckingham, 2018). APOE genotype was measured as described previously (Lyons et al., 2013). After accounting for data missingness across all variables of interest (exposures, outcomes, covariates) and excluding prevalent dementia or stroke cases, a total of 387 VETSA participants were included in the final analytic sample.

2.3. WHIMS overview, participants, and covariates

WHIMS is an ancillary study of WHI – hormone therapy trials that investigate associations between postmenopausal hormone therapy and cognitive function and dementia risk. WHIMS recruited dementia-free community-dwelling women ≥ 65 years of age at the enrollment in 1996–98 (Shumaker et al., 1998). In 2005–06, WHIMS participants from 14 U.S. sites were enrolled in the WHIMS-MRI study and underwent a brain MRI scan (Coker et al., 2009; Jaramillo et al., 2007). The current study was restricted to 1097 women without dementia or history of stroke at the time of MRI, who also had complete data on exposures, MRI measures, and key covariates.

Structured questionnaires were administered at WHI enrollment (1993–1998) to assess age, race/ethnicity, and socioeconomic factors (education, family income, and employment). As this is a predominantly White population, we combined “Hispanic/Latino”, “American Indian/Alaska Native”, “Asian”, “Black”, or “More than one race” into “Other ethnic or racial background”. Depressive symptoms were assessed using the CES-D short form (Andresen et al., 1994) and self-reported use of postmenopausal hormones prior to study participation were also collected at WHI enrollment. Lifestyle factors (smoking, alcohol use, and physical activity) were collected up to MRI scan. BMI was calculated using weight and height measured at the latest visit prior to MRI scan. History of cardiovascular disease (CVD; e.g., heart problems, problems with blood circulation, blood clots) and related risk factors (e.g., hypertension, hypercholesterolemia or diabetes mellitus) were collected up to MRI scan. Good reliability and validity of the physical measures and the self-reported medical histories have been previously documented (Heckbert et al., 2004; Johnson-Kozlow et al., 2007; Margolis et al., 2008). Given the high comorbidity of CVD risk factors and few women endorsing a history of diabetes, we created a binary CVD variable to indicate if a person had none or at least one type of CVD risk factor prior to MRI. Neighborhood SES was characterized using U.S. Census tract-level data (Diez Roux et al., 2001) at the time of MRI, with higher values indicating more socioeconomically-favorable neighborhoods. APOE genotype was measured in a subset of 1047 women (Hayden et al., 2019). Additional details are available in the supplementary materials.

2.4. Neuroimaging acquisition and processing for VETSA and WHIMS cohorts

VETSA MRI scans were collected at MGH on a Siemens Trio 3 T scanner or at UCSD on a GE Medical Systems Discovery 3 T scanner. A T1-weighted Magnetization-Prepared Rapid Acquisition Gradient Echo (MPRAGE) or fast spoiled gradient echo (FSPGR) scan was collected for each participant at MGH and UCSD, respectively. WHIMS MRI scans were collected on 1.5 T scanners across 14 sites using a standardized protocol. A T1-weighted gradient echo scan was collected for each participant.

Regional CTh values were extracted from T1-weighted scans and parcellated into 68 regions of interest (ROI) (34/hemisphere) using FreeSurfer-v5.3.0, which employs the Desikan-Killiany Atlas to define regions. Each MR analyst was trained on a standard dataset to identify acceptable (passed) versus unacceptable (failed) CTh segmentations and were required to pass a reliability test against an established dataset with a Cohen’s kappa ≥ 0.60 before visually inspecting each scan in both datasets. The segmentation passed or failed on a region-by-region basis. The missing rate (% failed) for VETSA participants was < 10% for all regions, and for WHIMS participants was < 10% for all regions except the superior temporal gyri (17.68%), temporal poles (17.14%), and superior frontal gyri (10.03%). Failed regions were imputed for both cohorts using the Iterative Imputer scikit-learn method in Python as described in the supplementary materials. Prior to imputing, we tested whether fails were missing at random using a χ²test between individuals with dementia and cognitively unimpaired participants in WHIMS, adjusting for multiple comparisons across regions using a Bonferroni correction. Only the left pars opercularis, left caudal middle frontal, and left supramarginal gyrus had segmentation fails that were not missing at random by diagnosis, so imputation in this region may limit interpretation. Because all VETSA participants were cognitively unimpaired at the time of scan, no such test was performed. Given that up to 17% of ROIs failed QC in the WHIMS cohort, we conducted supplemental analyses exclusively in participants who passed QC (i.e., no data imputations), as described in the statistical analysis section.

To assess the relationship between exposures and AD-related neurodegeneration, we created an AD meta-ROI based on the surface-area weighted average of imputed mean CTh in four bilateral regions that are vulnerable to AD: (1) entorhinal cortex (ERC), (2) inferior temporal gyrus (ITG), and (3) middle temporal gyrus (MTG), and (4) fusiform gyrus (Jack et al., 2017). CTh in this AD meta-ROI was our primary outcome measure. Secondary analyses focused on regional CTh for each of the 34 bilateral ROIs.

2.5. Geocoding residential addresses

Detailed procedures for geocoding VETSA residential addresses and calculating air pollution assessments are published elsewhere (Franz et al., 2023). Using a combination of 3 sources of residential addresses, we reconstructed and geocoded address histories from 1993 to 2017 for addresses lived at for 6 months or longer. Participants were excluded if more than 20% of their monthly exposure estimates were missing. Residential addresses for WHIMS participants were collected at each visit since its enrollment in 1993, updated at least semiannually during follow-up, and geocoded using standardized procedures (Whitsel et al., 2006).

2.6. Air pollution assessments

Air pollutant levels for PM_2.5 and NO₂ at the anonymized geocoded addresses were calculated by the University of Washington Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air) group using standard procedures for both cohorts. Biweekly mean concentrations of ambient PM_2.5 in μg/m³ and NO₂ in ppb were estimated using validated national spatiotemporal model, an extension of regionalized universal kriging models with partial least squares regression of geographic covariates. The spatiotemporal model used data from intensive regulatory monitors, investigator-deployed non-regulatory monitors, and satellite-derived measurements. Cross-validation R² was 0.89 for PM_2.5 and 0.87 for NO₂ (Kirwa et al., 2021). Additional details are available in the supplementary materials. For statistical analyses, we created personalized PM_2.5 and NO₂ estimates by aggregating these biweekly predictions over the 3 years prior to the MRI date, weighted by the length of stay at each residential location during these 3 years to account for residential mobility.

2.7. Statistical approach for both datasets

All statistical analyses were performed using R version–4.1.2 and SAS version–9.4. We used linear mixed effect models to estimate associations between air pollution exposures and the AD meta-ROI, adjusting for age at MRI and scanner manufacturer (Model A), while modeling random effects for MRI scanning site in WHIMS to control for between-site variability. With only two sites in VETSA, adjusting for manufacturer was equivalent to adjusting for scanner and site. In VETSA analyses we also modeled the twin family as a random effect to account for within-twin pair correlations. Other potential confounders were sequentially evaluated in models B to E including sociodemographic variables (geographic region, race/ethnicity, individual and neighborhood SES), lifestyle factors (smoking status, alcohol consumption, physical activity, and, for WHIMS only, prior postmenopausal hormone use and WHI hormone therapy assignment), and clinical characteristics (BMI, depressive symptoms, CVD risk factors or Charlson comorbidity). To determine the clinical translatability of the association between the exposure and the outcome, we calculated age-equivalent exposure effects by taking the ratio of the association estimate of the exposure variable over the estimate of the age variable from the same model. (Gitsels et al., 2019; Spiegelhalter 2016) We further evaluated if associations were influenced by APOE genotype by covarying for APOE ε4 carrier status (any ε4 allele vs. none) in those with available data. Finally, as VETSA participants were significantly younger than those in WHIMS, we examined whether exposure associations with the AD meta-ROI differed by age by modeling age-by-exposure interaction terms. All covariates were coded similarly in both cohorts to better compare the results. Categorical variables were modeled according to the levels reported except for age, neighborhood SES, and depression severity symptoms, which were analyzed continuously. For inferential analyses, two-tailed tests were applied with statistical significance defined at the 0.05 alpha level.

Analyses of regional CTh followed the same modeling strategy, using the false discovery rate (FDR) approach (Benjamini and Hochberg, 1995) to correct for multiple comparisons across the 34 ROIs. We plotted the estimated beta coefficients from the fully adjusted model (Model E) on cortical brain maps to illustrate the results. To evaluate the impact of imputing data for failed ROIs in WHIMS and VETSA, we repeated the analyses using only the CTh measures of ROIs that passed the image QC. We further conducted 2-pollutants models for PM_2.5 and NO₂ to assess the impact of each exposure conditional on the other.

3. Results

3.1. Descriptive information

Among the 1365 WHIMS women who underwent brain MRI scanning, 1222 passed the image process quality checking. We further excluded 13 women with dementia and 32 women with a history of stroke at the time of MRI. An additional 80 women missing covariates were excluded, resulting in a final analytic sample of 1097 cognitively unimpaired women (Fig. 1a). Similarly, a total of 387 VETSA men were included in the final analytic sample of VETSA cohort (Fig. 1b).

Fig. 1. — Abbreviations: MRI, magnetic resonance imaging; VETSA, Vietnam Era Twin Study of Aging; WHIMS, Women’s Health Initiative Memory Study.

For both cohorts, PM_2.5 exposures were roughly normally distributed in each geographic region, while the distribution of NO₂ exposures were slightly skewed to the right (Figure S1). In WHIMS, women living in the Northeast and Midwest had higher PM_2.5 exposures, while women living in the Northeast and West had higher NO₂ exposures than those from other regions (Table 1). Non-Hispanic White women had significantly lower PM_2.5 and NO₂ exposures than those of other ethnic and racial backgrounds. Compared to their counterparts, women living in the most socioeconomically-unfavorable neighborhoods had the highest PM_2.5 exposures, while those living in the most socioeconomically-favorable neighborhoods had the highest NO₂ exposures (Table 1). Higher NO₂ exposures were also observed among women aged ≥ 80 years, those who completed ≥ 4 years of college, current smokers, those previously or currently drinking alcohol, and women with high BMI (≥35 kg/m²). On average, there was higher CTh in the AD meta-ROI among women who identified as non-Hispanic White, were < 75 years of age, completed high school or GED education, and never smoked. Those with a thinner AD meta-ROI more frequently lived in the South or in the most socioeconomically-favorable neighborhoods (Table 1).

Table 1.

Distribution of air pollution exposures and AD meta-ROI by population characteristics of WHIMS MRI analytic sample.

		PM_2.5 exposure (μg/m³)^*		NO₂ exposure (ppb)^*		AD meta-ROI (mm)^†
Population Characteristics	N (%)	Mean ± SD	P^‡	Mean ± SD	P^‡	Mean ± SD	P^‡

Overall	1097 (100)	11.20 ± 2.75		11.40 ± 6.72		2.46 ± 0.23
Age (years)			0.15		< 0.001***		< 0.001***
≥ 71 and < 75	263 (24.0)	11.02 ± 2.57		9.51 ± 5.89		2.51 ± 0.18
≥ 75 and < 80	524 (47.8)	11.14 ± 2.80		11.60 ± 6.75		2.48 ± 0.23
≥ 80 and < 88	310 (28.3)	11.45 ± 2.81		12.65 ± 6.99		2.39 ± 0.26
U.S. Geographic Region			< 0.001***		< 0.001***		< 0.001***
Northeast	254 (23.2)	11.62 ± 2.91		14.09 ± 9.00		2.44 ± 0.28
South	165 (15.0)	10.91 ± 1.38		6.60 ± 2.83		2.33 ± 0.23
Midwest	381 (34.7)	11.41 ± 1.55		9.24 ± 3.43		2.51 ± 0.20
West	297 (27.1)	10.72 ± 4.02		14.52 ± 6.59		2.49 ± 0.20
Ethnicity			< 0.001***		< 0.001***		0.004^**
White (not Hispanic)	1018 (92.8)	11.03 ± 2.63		10.85 ± 6.13		2.47 ± 0.23
Other ethnic or racial background^§	79 (7.2)	13.33 ± 3.34		18.46 ± 9.46		2.39 ± 0.24
Education			0.12		0.004^**		0.02^*
≥ High school or GED	296 (27.0)	11.22 ± 2.42		10.37 ± 5.71		2.49 ± 0.22
> HS/GED but < 4 y of college	445 (40.6)	11.01 ± 2.78		11.52 ± 6.67		2.46 ± 0.23
≥ 4 y of college	356 (32.5)	11.41 ± 2.96		12.09 ± 7.44		2.44 ± 0.25
Employment status			0.96		0.67		0.06
Currently working	142 (12.9)	11.25 ± 2.78		11.31 ± 6.63		2.49 ± 0.23
Not working	94 (8.6)	11.24 ± 2.60		10.83 ± 6.00		2.49 ± 0.20
Retired	861 (78.5)	11.18 ± 2.76		11.47 ± 6.81		2.45 ± 0.24
Family income			0.79		0.40		1.00
< $20,000	239 (21.8)	11.11 ± 2.35		11.39 ± 6.20		2.47 ± 0.22
$20,000 to $34,999	314 (28.6)	11.22 ± 2.77		11.04 ± 6.77		2.46 ± 0.23
$35,000 to $49,999	238 (21.7)	11.03 ± 2.84		11.08 ± 6.96		2.46 ± 0.25
$50,000 to $74,999	160 (14.6)	11.39 ± 3.06		12.34 ± 7.25		2.46 ± 0.22
≥ $75,000	86 (7.8)	11.35 ± 2.78		11.90 ± 6.49		2.46 ± 0.26
Not reported	60 (5.5)	11.36 ± 2.88		11.33 ± 6.32		2.47 ± 0.24
Neighborhood SES^¶			< 0.001***		< 0.001***		0.004^**
Quartile of worst neighborhood	275 (25.1)	11.68 ± 2.59		11.12 ± 7.05		2.46 ± 0.24
Quartile of second worst neighborhood	274 (25.0)	11.11 ± 2.64		10.15 ± 6.64		2.47 ± 0.23
Quartile of third worst neighborhood	273 (24.9)	10.69 ± 2.83		10.78 ± 5.82		2.50 ± 0.22
Quartile of best neighborhood	275 (25.1)	11.30 ± 2.85		13.53 ± 6.84		2.42 ± 0.25
Lifestyle
Smoking status			0.07		0.03^*		0.01^*
Never smoked	624 (56.9)	11.21 ± 2.71		11.00 ± 6.32		2.48 ± 0.23
Past smoker	439 (40.0)	11.10 ± 2.78		11.80 ± 7.06		2.44 ± 0.24
Current Smoker	34 (3.1)	12.22 ± 2.97		13.48 ± 8.60		2.41 ± 0.22
Alcohol use			0.58		< 0.001***		0.98
Non-drinker	137 (12.5)	11.38 ± 2.36		9.25 ± 5.45		2.46 ± 0.25
Past drinker	370 (33.7)	11.21 ± 2.58		11.34 ± 6.42		2.46 ± 0.23
Low consumption (< 1 drink/day)	486 (44.3)	11.20 ± 2.90		12.08 ± 7.01		2.46 ± 0.23
High consumption (≥ 1 drink/day)	104 (9.5)	10.88 ± 3.09		11.21 ± 7.33		2.47 ± 0.23
Moderate or strenuous physical activity ≥ 20 min			0.75		0.27		0.27
No activity	647 (59.0)	11.27 ± 2.71		11.72 ± 6.97		2.45 ± 0.24
1 episode/week	44 (4.0)	11.26 ± 2.65		10.62 ± 6.98		2.45 ± 0.26
2–4 episodes/week	235 (21.4)	11.07 ± 2.75		11.06 ± 6.22		2.46 ± 0.22
> 4 episodes/week	171 (15.6)	11.09 ± 2.92		10.83 ± 6.32		2.49 ± 0.24
Physical Health
Body Mass Index (kg/m²)			0.15		0.01^*		0.30
< 25	334 (30.4)	11.22 ± 2.86		11.85 ± 6.41		2.45 ± 0.24
≥ 25 and < 30	412 (37.6)	11.03 ± 2.70		10.73 ± 6.27		2.47 ± 0.24
≥ 30 and < 35	235 (21.4)	11.21 ± 2.63		11.22 ± 7.01		2.47 ± 0.23
≥ 35	116 (10.6)	11.69 ± 2.80		12.82 ± 8.15		2.43 ± 0.23
CVD risk indicator^#			0.23		0.62		0.06
No	210 (19.1)	11.40 ± 2.74		11.19 ± 5.56		2.49 ± 0.22
Yes	887 (80.9)	11.15 ± 2.75		11.45 ± 6.97		2.46 ± 0.24
Depressive symptoms at WHI enrollment^**			0.61		0.31		0.33
≤ median	685 (62.4)	11.17 ± 2.73		11.24 ± 6.64		2.47 ± 0.22
> median	412 (37.6)	11.25 ± 2.79		11.66 ± 6.85		2.45 ± 0.25
Any prior postmenopausal hormone use before WHI enrollment			0.43		0.20		0.50
No	581 (53.0)	11.26 ± 2.56		11.15 ± 6.33		2.47 ± 0.23
Yes	516 (47.0)	11.13 ± 2.95		11.67 ± 7.13		2.46 ± 0.24
WHI Hormone Therapy Assignment			0.89		0.64		0.57
CEE-alone placebo	200 (18.2)	11.11 ± 2.76		11.01 ± 6.45		2.47 ± 0.23
CEE-alone intervention	208 (19.0)	11.14 ± 2.77		11.58 ± 7.10		2.44 ± 0.24
CEE + MPA placebo	360 (32.8)	11.20 ± 2.66		11.23 ± 6.31		2.46 ± 0.24
CEE + MPA intervention	329 (30.0)	11.28 ± 2.84		11.69 ± 7.07		2.47 ± 0.23
APOE genotype			0.96		1.00		0.49
ε2/2 + ε2/3 + ε3/3	819 (78.2)	11.17 ± 2.74		11.20 ± 6.53		2.46 ± 0.23
ε2/4 + ε3/4 + ε4/4	228 (21.8)	11.16 ± 2.68		11.19 ± 6.58		2.45 ± 0.24
Scanner Manufacturer			< 0.001***		< 0.001***		< 0.001***
General Electric	610 (55.6)	11.61 ± 3.00		11.95 ± 5.68		2.57 ± 0.18
Philips	187 (17.0)	11.41 ± 2.63		15.65 ± 9.80		2.26 ± 0.22
Siemens	183 (16.7)	10.04 ± 2.43		7.46 ± 4.23		2.34 ± 0.22
Toshiba	117 (10.7)	10.51 ± 0.58		7.86 ± 2.66		2.44 ± 0.20

Open in a new tab

Abbreviations: AD, Alzheimer’s Disease; APOE, Apolipoprotein E; CEE, conjugated equine estrogens; CVD, cardiovascular disease; HS/GED, high school or general educational development; mm, millimeter; MPA, medroxyprogesterone acetate; MRI, magnetic resonance imaging; NO₂, nitrogen dioxide; PM_2.5, fine particulate matter; ppb, parts per billion; SD, standard deviation; SES, socioeconomic status; WHI, Women’s Health Initiative; WHIMS, WHI Memory Study.

Air pollution was estimated as the 3-year average exposure prior to the MRI visit.

^†

AD meta-ROI was calculated as the surface-area-weighted thickness in the bilateral entorhinal, fusiform, inferior temporal, and middle temporal cortices. Higher AD meta-ROI values represent thicker cortices in AD vulnerable brain regions.

^‡

p-values were calculated using ANOVA F-tests for mean exposures.

^§

Other ethnic or racial background includes Hispanic women or non-Hispanic women with race in American Indian/Alaska Native, Asian, Black, more than one race, or unknown/not reported.

^¶

Neighborhood SES is the sum of six standardized U.S. Census tract–level variables measuring domains of wealth/income, education, and occupation. A higher neighborhood SES score indicates a more advantageous neighborhood SES. The current categories were classified based on values comparing to the 25, 50 and 75 %iles of the population levels.

CVD and related risk factors were defined by whether a person had reported any history of CVD event (self-reported at WHI enrollment or any incident CVD events occurred before MRI), hypertension, hypercholesterolemia, or diabetes mellitus (none vs. at least one).

^**

Depressive symptoms were assessed using the Center for Epidemiologic Studies Depression Scale short form. The score ranged from 0 to 1 with a higher score indicating a greater likelihood of depression. The current categories were classified based on the median value.

In VETSA, men residing in the South and Midwest had higher PM_2.5 exposures, but lower NO₂ exposures (Table 2), compared to their counterparts. Similar to WHIMS, men living in the most socioeconomically-unfavorable neighborhoods had the highest PM_2.5 exposures, while those living in the most socioeconomically-favorable neighborhoods had the highest NO₂ exposures (Table 2). Lower NO₂ exposures were observed in men who identified as non-Hispanic White and completed high school or GED equivalent. On average, the AD meta-ROI was thicker among men who identified as non-Hispanic White, had never smoked, or had a BMI < 25 kg/m². The meta-ROI was thinner among men living in the most socioeconomically-unfavorable neighborhoods (Table 2).

Table 2.

Distribution of air pollution exposures and AD meta-ROI by population characteristics of VETSA MRI analytic sample.

		PM_2.5 exposure (μg/m³)^*		NO₂ exposure (ppb)^*		AD meta-ROI (mm)^†

Population Characteristics	N (%)	Mean ± SD	P^‡	Mean ± SD	P^‡	Mean ± SD	P^‡

Overall	387 (100)	8.93 ± 2.11		7.16 ± 4.19		2.69 ± 0.11
Age (years)			0.40		0.22		0.31
≥ 55 and < 60	138 (35.7)	9.24 ± 2.12		7.37 ± 4.94		2.68 ± 0.11
≥ 60 and < 64	160 (41.3)	8.78 ± 2.03		6.69 ± 3.12		2.70 ± 0.11
≥ 64 and < 67	89 (23.0)	8.71 ± 2.20		7.68 ± 4.53		2.68 ± 0.10
U.S. Geographic Region			< 0.001***		< 0.001***		0.41
Northeast	41 (10.6)	8.33 ± 1.88		8.18 ± 5.13		2.71 ± 0.09
South	93 (24.0)	9.68 ± 1.25		5.46 ± 2.23		2.69 ± 0.10
Midwest	149 (38.5)	9.80 ± 1.46		6.28 ± 2.73		2.69 ± 0.11
West	104 (26.9)	7.24 ± 2.50		9.56 ± 5.52		2.68 ± 0.10
Ethnicity			0.16		< 0.001***		< 0.001***
White (not Hispanic)	342 (88.4)	8.86 ± 2.08		6.79 ± 3.67		2.70 ± 0.10
Other ethnic or racial background^§	45 (11.6)	9.46 ± 2.30		10.02 ± 6.34		2.63 ± 0.11
Education			0.59		0.005^**		0.52
≤ High school or GED	158 (40.8)	8.87 ± 2.18		6.18 ± 3.59		2.68 ± 0.11
> HS/GED but < 4 y of college	129 (33.3)	9.05 ± 2.00		8.04 ± 4.71		2.69 ± 0.10
≥ 4 y of college	100 (25.8)	8.85 ± 2.15		7.58 ± 4.07		2.70 ± 0.11
Employment status			0.78		0.38		0.85
Currently working	273 (70.5)	8.93 ± 2.07		6.96 ± 3.92		2.69 ± 0.10
Not working	28 (7.2)	9.16 ± 2.45		6.93 ± 4.87		2.68 ± 0.11
Retired	86 (22.2)	8.85 ± 2.14		7.89 ± 4.72		2.69 ± 0.11
Family Income			0.11		0.52		0.85
< $30,000	37 (9.6)	9.04 ± 2.34		6.53 ± 4.53		2.68 ± 0.11
$30,000 to $49,999	59 (15.2)	9.14 ± 1.97		7.56 ± 4.68		2.68 ± 0.13
$50,000 to $79,999	123 (31.8)	8.80 ± 2.13		6.82 ± 4.04		2.69 ± 0.10
$80,000 to $99,999	54 (14.0)	9.41 ± 2.15		7.63 ± 4.52		2.69 ± 0.11
≥ $100,000	75 (19.4)	8.92 ± 1.95		7.52 ± 3.89		2.69 ± 0.10
Not reported	39 (10.1)	8.24 ± 2.19		6.91 ± 3.63		2.72 ± 0.10
National Neighborhood Atlas^¶			0.004^**		< 0.001***		0.03^*
Quartile of worst neighborhood	97 (25.1)	9.52 ± 1.76		6.16 ± 3.06		2.67 ± 0.11
Quartile of second worst neighborhood	98 (25.3)	8.93 ± 1.89		6.12 ± 3.40		2.72 ± 0.10
Quartile of third worst neighborhood	97 (25.1)	8.66 ± 2.03		6.96 ± 3.48		2.68 ± 0.11
Quartile of best neighborhood	95 (24.5)	8.58 ± 2.59		9.48 ± 5.53		2.69 ± 0.10
Lifestyle
Smoking status			0.94		0.9		0.01^*
Never smoked	148 (38.2)	8.93 ± 2.14		7.02 ± 3.92		2.71 ± 0.10
Past smoker	164 (42.4)	8.93 ± 2.07		7.18 ± 4.24		2.69 ± 0.11
Current Smoker	75 (19.4)	8.91 ± 2.19		7.41 ± 4.62		2.67 ± 0.10
Alcohol use			0.73		0.66		0.29
Non-drinker	30 (7.8)	8.74 ± 2.24		6.23 ± 3.07		2.70 ± 0.12
Past drinker	105 (27.1)	9.26 ± 2.01		7.45 ± 4.76		2.69 ± 0.11
Low consumption (1 – 2 drinks/day)	201 (51.9)	8.82 ± 2.16		7.07 ± 3.89		2.69 ± 0.10
High consumption (>2 drinks/ day)	51 (13.2)	8.75 ± 2.04		7.49 ± 4.65		2.68 ± 0.10
Moderate or strenuous physical activity ≥ 20 min			0.95		0.79		0.58
No activity	51 (13.2)	9.07 ± 2.26		6.66 ± 3.79		2.67 ± 0.11
1 episode/week	114 (29.5)	8.88 ± 2.19		7.05 ± 4.33		2.69 ± 0.11
2–4 episodes/week	77 (19.9)	8.71 ± 2.01		7.48 ± 4.47		2.71 ± 0.12
> 4 episodes/week	145 (37.5)	9.03 ± 2.06		7.26 ± 4.08		2.69 ± 0.10
Physical Health
Body Mass Index (kg/m²)			0.26		0.78		0.02^*
< 25	93 (24.0)	8.79 ± 2.16		7.19 ± 4.25		2.74 ± 0.10
≥ 25 and < 30	178 (46.0)	8.85 ± 2.19		7.00 ± 3.94		2.69 ± 0.10
≥ 30 and < 35	97 (25.1)	8.95 ± 1.94		7.10 ± 4.37		2.65 ± 0.11
≥ 35	19 (4.9)	10.16 ± 1.69		8.82 ± 5.11		2.69 ± 0.11
Charlson comorbidity score^#			0.33		0.48		0.52
0	101 (26.1)	8.82 ± 2.16		6.71 ± 3.93		2.70 ± 0.11
≥1	286 (73.9)	8.96 ± 2.10		7.32 ± 4.27		2.69 ± 0.11
Depressive symptoms^**			0.77		0.74		0.33
≤ median	225 (58.1)	8.95 ± 2.10		7.22 ± 4.16		2.70 ± 0.11
> median	162 (41.9)	8.89 ± 2.14		7.08 ± 4.24		2.68 ± 0.11
APOE genotype			0.95		0.44		0.20
ε2/2 + ε2/3 + ε3/3	295 (76.2)	8.86 ± 2.08		7.12 ± 3.98		2.69 ± 0.11
ε2/4 + ε3/4 + ε4/4	92 (23.8)	9.14 ± 2.21		7.29 ± 4.83		2.70 ± 0.10
Scanner Manufacturer			0.10		0.83		0.76
General Electric	257 (66.4)	9.09 ± 2.12		7.16 ± 3.99		2.69 ± 0.11
Siemens	130 (33.6)	8.60 ± 2.07		7.18 ± 4.57		2.69 ± 0.09

Open in a new tab

Abbreviations: AD, Alzheimer’s Disease; APOE, Apolipoprotein E; HS/GED, high school or general educational development; mm, millimeter; MRI, magnetic resonance imaging; NO₂, nitrogen dioxide; PM_2.5, fine particulate matter; ppb, parts per billion; SD, standard deviation; VETSA, Vietnam Era Twin Study of Aging.

Air pollution was estimated as the 3-year average exposure prior to the MRI visit.

^†

^‡

p-values were calculated using ANOVA F-tests for mean exposures.

^§

Other ethnic or racial background includes Hispanic man or non-Hispanic men with race in American Indian/Alaska Native, Asian, Black, or More than one race.

^¶

National Neighborhood Atlas is the national percentile rankings at the block group level from 1 to 100. A ranking of 1 indicates the lowest level of “disadvantage” within the nation and a ranking of 100 indicates the highest level of “disadvantage”. The current categories were classified based on values comparing to the 25, 50 and 75 %iles of the population levels.

Charlson comorbidity score is calculated as the count of 15 major chronic conditions: diabetes, emphysema, asthma, cancer, osteoarthritis, rheumatoid arthritis, stroke, heart attack, heart failure, heart surgery, angina, hypertension, peripheral vascular disease, cirrhosis, and AIDS. Value 0 represents no chronic conditions and ≥ 1 represents at least one of the above 15 chronic conditions.

^**

Depressive symptoms at VETSA were assessed using the Center for Epidemiologic Studies Depression Scale standard form. The score ranged from 0 to 60 with a higher score indicating more depressive symptoms. The current categories were classified based on the median value.

Descriptive comparisons between participants excluded and included in the final analytic samples of both cohorts are presented in Tables S1 and S2. Comparison of the analytic samples of the two cohorts is included in Table S3. All these results were summarized in the supplementary materials.

3.2. Associations between air pollution exposures and CTh in the AD meta-ROI

Table 3 shows the associations between exposures and the standardized AD meta-ROI CTh with adjustment by different sets of covariates among the WHIMS or VETSA samples, respectively. We observed little differences in the associations between air pollution exposures and the AD meta-ROI comparing the fully adjusted model E with other partially adjusted models (Models B-D) except the minimally adjusted model A. We reported the results from Model E because it controlled for all potential confounders that might affect the associations.

Table 3.

Associations between air pollution exposures and standardized AD meta-ROI.

a). WHIMS MRI analytic sample

	PM_2.5 exposure^*				NO₂ exposure^*
Models^‡	β^†	95% Confidence Interval	t	p	β^†	95% Confidence Interval	t	p

Model A	−0.042	−0.068, −0.016	−3.18	0.001	−0.012	−0.022, −0.002	−2.41	0.016
Model B	−0.050	−0.076, −0.024	−3.77	< 0.001	−0.013	−0.023, −0.003	−2.44	0.015
Model C	−0.049	−0.075, −0.023	−3.70	< 0.001	−0.013	−0.023, −0.002	−2.41	0.016
Model D	−0.049	−0.076, −0.023	−3.70	< 0.001	−0.013	−0.023, −0.002	−2.39	0.017
Model E	−0.050	−0.076, −0.024	−3.75	< 0.001	−0.013	−0.023, −0.002	−2.38	0.018
b). VETSA MRI analytic sample
	PM_2.5 exposure^*				NO₂ exposure^*
Models^§	β^†	95% Confidence Interval	t	p	β^†	95% Confidence Interval	t	p
Model A	0.051	0.003, 0.099	2.09	0.038	0.012	−0.012, 0.036	0.98	0.330
Model B	0.071	0.016, 0.127	2.50	0.013	0.028	0.001, 0.054	2.05	0.042
Model C	0.067	0.011, 0.122	2.37	0.019	0.031	0.005, 0.057	2.31	0.022
Model D	0.067	0.013, 0.121	2.42	0.017	0.031	0.005, 0.056	2.35	0.020
Model E	0.067	0.013, 0.121	2.42	0.017	0.031	0.005, 0.056	2.34	0.021

Open in a new tab

Abbreviations: APOE, Apolipoprotein E; MRI, magnetic resonance imaging; NO₂, nitrogen dioxide; PM_2.5, fine particulate matter; ppb, parts per billion; SES, socioeconomic status; VETSA, Vietnam Era Twin Study of Aging; WHI, Women’s Health Initiative; WHIMS, WHI Memory Study.

Air pollution was estimated as the 3-year average exposure prior to the MRI visit.

^†

β represents the average change in standardized AD meta-ROI value for each unit increase of the exposure. A negative β means higher air pollution exposure was associated with a smaller AD meta-ROI value representing a thinner cortex in AD vulnerable brain regions.

^‡

Models for the WHIMS cohort:

Model A: adjusted for age, scanner manufacturer, and random effect of clinic sites.

Model B: adjusted for Model A covariates + geographic region, race/ethnicity, education, income, employment status, and neighborhood SES

Model C: adjusted for Model B covariates + lifestyle factors (smoking; alcohol use; physical activities), prior postmenopausal hormone use, and WHI hormone therapy assignment

Model D: adjusted for Model C covariates + body mass index and depressive symptoms (logit transformation of the raw score)

Model E: adjusted for Model D covariates + cardiovascular disease and related risk factors

^§

Models for the VETSA cohort:

Model A: adjusted for age, scanner manufacturer, and random effect of twin family indicator.

Model B: adjusted for Model A covariates + geographic region, race/ethnicity, education, income, employment status, and national neighborhood atlas

Model C: adjusted for Model B covariates + lifestyle factors (smoking; alcohol use; physical activities)

Model D: adjusted for Model C covariates + body mass index and depressive symptoms Model E: adjusted for Model D covariates + Charlson comorbidity

In WHIMS, women exposed to higher air pollution levels had lower CTh in the AD meta-ROI (β_PM2.5=− 0.050 per 1 μg/m³, 95% CI=[−0.076, −0.024], p < 0.001; β_NO2=−0.013 per 1 ppb, 95% CI=[−0.023, −0.002], p = 0.018) after adjusting for scanner manufacturer, sociodemographics, lifestyle factors, clinical covariates, and a random effect of clinic sites (Model E, Table 3a). Using the estimated effect of 1-year increase in age from the same model for comparison, these associations were equivalent to the effects of 12.9 months of aging for PM_2.5 and 3.3 months of aging for NO₂, respectively.

In contrast to WHIMS, VETSA men who were exposed to higher air pollution levels had higher CTh in the AD meta-ROI (β_PM2.5=0.067 per 1 μg/m³, 95% CI=[0.013, 0.121], p = 0.017; β_NO2=0.031 per 1 ppb, 95% CI=[0.005, 0.056], p = 0.021) after adjusting for scanner manufacturer, sociodemographics, lifestyle factors, clinical covariates, and a random effect for twin family indicator (Model E, Table 3b).

Among participants with APOE genotype data (WHIMS, n = 1047; VETSA, n = 387), associations remained similar with further adjustment for APOE genotype (Model F vs. Model E, Table S4).

3.2.1. Age-stratified analyses and age-by-exposure interactions on the AD meta-ROI

Fully adjusted associations between PM_2.5 and the AD meta-ROI differed significantly by age in VETSA (p interaction=0.01, Table S5). The positive associations between PM_2.5 exposure and the AD meta-ROI reduced gradually between ages 55–64 years, and were negative, though nonsignificant, after age 65 (Fig. 2a). However, in WHIMS women, the associations between PM_2.5 exposure and the AD meta-ROI were negative and the effects did not change much from age 71 to 88 years (Fig. 2a). There were no significant interactions between NO₂ exposure and age on the AD meta-ROI for either cohort (Fig. 2b, Table S5).

Fig. 2. — Associations between (a) PM_2.5 or (b) NO₂ exposures and the standardized AD meta-ROI at each age within the cohort-specific age ranges. Results were based on models adjusted for the same covariates in Model E of Table 3, plus an interaction term of age and air pollution exposure. Numeric results of the interaction tests are presented in Table S5. The solid dots represent beta estimates per 1 μg/m³ PM_2.5 (a) or 1ppb NO₂ (b) from multivariate adjusted association between air pollution exposure and standardized AD meta-ROI value at different ages. Gray areas indicate 95% confidence intervals. Negative beta values indicate higher air pollution exposure is associated with a smaller AD meta-ROI value representing thinner cortex in AD vulnerable brain regions. *Abbreviations:* NO₂, nitrogen dioxide; PM_2.5, fine particulate matter; ppb, parts per billion; ROI, region of interest; VETSA, Vietnam Era Twin Study of Aging; WHIMS, Women’s Health Initiative Memory Study.

3.3. Associations between air pollution exposures and regional CTh

In WHIMS, higher PM_2.5 exposures were associated with significantly lower CTh in 23 of the 34 ROIs after adjusting for covariates and multiple comparisons (Table S6a). Effect estimates were generally small, ranging from −0.03 to −0.06 for significant ROIs. The negative associations spanned all four lobes of the brain, with the strongest effects in the MTG, ERC, banks of superior temporal sulcus (bSTS), pars triangularis, and frontal pole (Table S6a, Fig. 3a). Higher NO₂ exposures were associated with lower CTh in the MTG after FDR correction (p_FDR=0.0429; Table S6a). Repeating the analyses using raw (unimputed) ROI data in WHIMS with missing values excluded yielded similar results (Table S8 and Figure S2a).

Fig. 3. — Beta estimates of the associations between air pollution and standardized CTh in 34 regions of interest for (a) WHIMS and (b) VETSA. Results were based on models adjusted for the same covariates in Model E of Table 3. Numeric results are presented in Table S6 (WHIMS) and Table S7 (VETSA). The color bar represents beta estimates (per 1 μg/m³ PM_2.5 or 1ppb NO₂) from multivariate adjusted association between air pollution exposure and standardized CTh of each brain region. All brain regions were plotted regardless of statistical significance of the associations. Negative beta values indicate higher air pollution exposure is associated with a thinner cortex. *Abbreviations*: CTh, cortical thickness; FDR, false discovery rate; MRI, magnetic resonance imaging; NO₂, nitrogen dioxide; PM_2.5, fine particulate matter; ppb, parts per billion; VETSA, Vietnam Era Twin Study of Aging; WHIMS, Women’s Health Initiative Memory Study.

In VETSA, higher PM_2.5 exposures were associated with higher CTh in the temporal pole after FDR correction (p_FDR=0.0376), and trends for greater CTh in the MTG, bSTS, and insula (Table S7a, Fig. 3b). There were no significant associations between NO₂ exposures and CTh in any brain region (Table S7b). Repeated analyses using raw (unimputed) ROI data in VETSA with missing values excluded showed similar results (Table S9, Figure S2b).

Similar results were seen in the sensitivity analyses of 2-pollutants models on regional CTh for both cohorts (Tables S10 and S11, Figure S3).

4. Discussion

This study examined relationships between air pollution exposure and CTh in AD-vulnerable regions, and regionally across the whole brain, in two cohorts of older adults. Higher exposures were associated with lower CTh in the AD meta-ROI in older women from the WHIMS cohort, equivalent to the effects of approximately one year of aging for PM_2.5 and three months of aging for NO₂, respectively. There were also widespread inverse associations between exposures and CTh across most of the temporal lobe and discrete regions in the frontal, parietal, and occipital lobes in older women. Conversely, higher exposures in men from the VETSA cohort were associated with greater CTh in the AD meta-ROI, and in the temporal pole. A significant age-by-exposure interaction was observed for AD meta-ROI CTh in VETSA only, such that higher exposures were associated with greater CTh until approximately age 64 when the association became negative.

Previous studies reported adverse effects of air pollution on total gray matter volume (Erickson et al., 2020), regional cortical gray matter volumes from voxel-based morphometry (Casanova et al., 2016), and a measure of “AD similarity” based on a voxelwise probability map of cortical and deep gray matter (Younan et al., 2020). Two cross-sectional studies examined exposure effects on FreeSurfer-derived CTh metrics in older adults on structural MRI. In the ALFA project, Crous-Bou et al. (2020) examined exposure associations with CTh in an AD meta-ROI consisting of the ERC, MTG, ITG, fusiform, PCC, precuneus, and SMG among cognitively unimpaired adults genetically enriched for AD (M_age=57.7 ± 7.6). Consistent with our findings in WHIMS women, higher NO₂ exposures were associated with marginally lower CTh in the meta-ROI after adjusting for age, sex, education, and APOE genotype. However, PM_2.5 exposures were not associated with CTh, and age-by-exposure interactions were nonsignificant. The lack of PM_2.5 effect in ALFA may be due to nonsignificant effects in the PCC, SMG, and precuneus, which were not included in our AD meta-ROI, were not associated with PM_2.5 in VETSA participants, and were not associated with NO₂ in either the VETSA or WHIMS cohorts. Various AD meta-ROIs have been analyzed in prior work, each consisting of a unique set of regions that define AD “vulnerability” (Dickerson et al., 2009; Clifford R. Jack Jr et al., 2017; McEvoy et al., 2009; Pettigrew et al., 2016; Schwarz et al., 2016). Although there is no consensus in the literature as to which meta-ROI is superior, our AD meta-ROI was selected based on evidence that a composite measure of surface area-weighted CTh in the ITG, MTG, ERC, and fusiform distinguished cognitively impaired adults from age-matched controls with ≥ 90% sensitivity and specificity (Jack et al., 2017). ALFA participants also were exposed to significantly higher pollution levels (NO₂: 57.5 ± 12.9 ppb; PM_2.5: 17 ± 2.3 μg/m³) than those in VETSA (NO₂: 7.2 ± 4.2 ppb; PM_2.5: 8.9 ± 6.7 μg/m³) or WHIMS (NO₂: 11.4 ± 6.7 ppb; PM_2.5: 11.2 ± 2.8 μg/m³), yet this could only explain the lack of regional exposure associations with the PCC, SMG, and precuneus in our study. Thus, we postulate that nonsignificant PM_2.5 effects in ALFA participants were driven by age, as the ALFA cohort was younger on average (M_age=57.7 ± 7.6) than VETSA (M_age=61.9 ± 2.6), and approximately 20 years younger on average than WHIMS (M_age=77.9 ± 3.7). This is noteworthy because in contrast to non-significant age-by-exposure effects reported in the ALFA study, we did detect age-by-exposure effects on the AD meta-ROI in VETSA participants until approximately age 64, with stronger positive associations at younger ages. Interactions between age and sex may explain this discrepancy across studies, as sex-stratified analyses in the ALFA study revealed exposure effects in females only, yet VETSA consists of only males. Finally, our study adjusted for many more demographic and psychosocial variables that have been shown to influence brain integrity and strengthened the effect sizes of both pollutants on AD meta-ROI CTh in both cohorts as shown in Table 3.

Our regional CTh results can be compared to those from the Environmental Pollution-Induced Neurological EFfects (EPINEF) study in South Korea. Cho et al. (2020) used surface-based morphometry to examine air pollution associations with regional CTh among 957 non-demented older adults (M_age=67.3 ± 6.4). Higher PM_2.5 and NO₂ exposures were associated with lower CTh in the lateral temporal, inferior parietal, prefrontal, parahippocampal, insular, and fusiform gyri, and higher CTh in the lateral occipital cortex and postcentral gyri after adjusting for sociodemographic, lifestyle, and health-related risk factors. Higher NO₂ also was associated with lower PCC CTh. Authors also reported sex-by-exposure interactions on temporal lobe CTh, with lower CTh in males than females at a similar level of PM_2.5 exposure, but not NO₂ (Cho et al., 2020). These findings partially align with our results showing that higher PM_2.5 exposure was associated with lower CTh in the inferior parietal and insular cortex, and a trend for lower CTh in the parahippocampal gyrus of WHIMS women. We cannot make head-to-head comparisons to EPINEF’s purported results in “prefrontal” and “lateral temporal” cortices as these are not directly comparable to Desikan-Killiany atlas labels that may describe several ROIs. We also cannot speak to the purported sex effect in the EPINEF study given the bifurcation of males and females in the VETSA and WHIMS cohorts, and because WHIMS women were on average 10 years older than the average age of EPINEF participants. NO₂ exposures were not associated with CTh in any of the aforementioned cortical regions in either VETSA or WHIMS, possibly due to significantly lower NO₂ exposure levels than in EPINEF (M_NO2=29.0 ± 9.8 ppb) (Cho et al., 2020).

One important consistency between our study and Cho et al. (2020) is the observed positive association between exposures and temporal lobe CTh in VETSA men. Increasing evidence suggests that greater CTh does not always indicate better brain health. Several neuroimaging studies now support a biphasic pattern of CTh along the AD continuum (Fortea et al., 2011; Harrison et al., 2021; Montal et al., 2018, 2021; Pegueroles et al., 2017; Williams et al., 2023), characterized by increased CTh in early disease stages, followed by cortical thinning as disease progresses (Williams et al., 2023). Greater CTh in early disease has been attributed to neuroinflammation, neuronal hypertrophy, and initial Aβ accumulation prior to detectable tauopathy (Fortea et al., 2014; Heinzinger et al., 2023; Ismail et al., 2020). This nonlinear CTh pattern has been shown in individuals who meet criteria for clinically significant amyloid burden without significant tauopathy in the lateral occipital cortex and temporal-parietal areas such as the MTG, temporal pole, and precuneus (Fortea et al., 2014; Harrison et al., 2021; Pegueroles et al., 2017; Spotorno et al., 2023). In our study, CTh in these regions was positively associated with PM_2.5 in VETSA men, prior to multiple test correction. Although biomarker data were unavailable in our study, at average age 62, very few VETSA participants are likely to have been amyloid-positive (Fleisher et al., 2013). Furthermore, tauopathy is strongly positively correlated with age, typically beginning in medial temporal areas (Schöll et al., 2016). This may explain why positive associations between exposures and CTh decreased with age in our study, and became non-significant after age 64. Thus, the observed positive associations between exposures and CTh in younger VETSA men may reflect an early stage of neuropathology. As WHIMS participants were on average > 15 years older than those in VETSA, we suspect that the diffuse inverse associations between exposures and CTh in WHIMS represents later stage exposure-related neuropathology. Additional studies are needed to test this hypothesis in relation to amyloid and tau biomarkers.

Limitations of this study include our lack of amyloid and tau biomarkers and our cross-sectional design, which hinder our ability to fully examine the biphasic model. Although comparing effects in two independent samples provided insights into possible biphasic patterns, the observed age-by-exposure interaction effects cannot be interpreted independent of sex and we cannot draw any conclusions about causality. Future studies should examine three-way associations between age, sex, and air pollution exposures on CTh to better characterize these relationships. Furthermore, we could not fully control inherent differences in sample composition and design procedures that reflect the methodologies of the overarching studies. While these differences should be considered in relation to our interpretations, they did not affect our results since the analyses were conducted independently. Finally, restriction to cognitively healthy participants with good imaging data might introduce selection bias due to depletion of susceptible individuals (survivor bias) if those who experienced the greatest air pollution-related cortical thinning were not recruited due to dementia or removed from the study by the imaging analyst. Consequently, estimated associations between air pollution exposure and cortical thickness may be biased toward the null.

5. Conclusion

In summary, our study examined associations between air pollution exposures and CTh in AD vulnerable regions, and secondarily, individual regions across the cortex in two independent samples. Higher exposures were associated with lower CTh in AD-vulnerable regions and across most of the brain in women from the WHIMS cohort, whereas higher exposures were associated with higher CTh in these same AD-vulnerable regions in men from the VETSA cohort. Future longitudinal work should examine whether these divergent findings reflect the previously reported biphasic pattern of AD-related neurodegeneration, whereby increased CTh occurs in the earliest disease stages followed by cortical thinning as disease advances. Longitudinal studies are required to validate these interpretations, ideally in a large sample that includes men and women recruited from the same study and with neuroinflammation and AD biomarkers available for analysis.

Supplementary Material

supplementary results

NIHMS2198545-supplement-supplementary_results.docx^{(27.8KB, docx)}

supplementary tables and figures

NIHMS2198545-supplement-supplementary_tables_and_figures.docx^{(2.9MB, docx)}

Acknowledgements

The air pollution models were developed under a STAR research assistance agreement, No. RD831697 (MESA Air) and RD-83830001 (MESA Air Next Stage), awarded by the US Environmental Protection Agency (EPA). This study is supported by P01AG055367, R01AG033078 (PI: Dr. Chen), RF1AG054068 (PI: Dr. Chen), R01ES025888 (PI: Drs. Chen & Kaufman), the Southern California Environmental Health Sciences Center (5P30ES007048), and the Alzheimer’s Disease Research Center at USC (NIA; P50AG005142 and P30AG066530). Dr. Salminen is supported by grant 1R01ES033961-01, which is co-funded by the National Institute of Environmental Health Sciences (NIEHS) and the National Institute on Aging (NIA). Dr. Espeland receives funding from the Wake Forest Alzheimer’s Disease Core Center (P30AG049638-01A1). Dr. Resnick is supported by the Intramural Research Program, National Institute on Aging, NIH. The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts HHSN268201600018C, HHSN268201600001C, HHSN268201600002C, HHSN268201600003C, and HHSN268201600004C. A list of contributors to WHI is available at https://www-whi-org.s3.us-west-2.amazonaws.com/wp-content/uploads/WHI-Investigator-Long-List.pdf.

Funding

This research was supported in part by the Intramural Research Program of the National Institutes of Health (NIH). The contributions of the NIH author are considered Works of the United States Government. The findings and conclusions presented in this paper are those of the author and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services. The funders/sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

GLOSSARY

AD: Alzheimer’s disease
ADRD: Alzheimer’s disease and related dementias
APOE: Apolipoprotein E
CTh: cortical thickness
FDR: false discovery rate
MRI: magnetic resonance imaging
NO₂: nitrogen dioxide
PM_2.5: fine particulate matter
ppb: parts per billion
ROI: region of interest
SES: socioeconomic status
VETSA: Vietnam Era Twin Study of Aging
WHIMS: Women’s Health Initiative Memory Study

Appendix A. Supporting information

Supplementary data associated with this article can be found in the online version at doi:10.1016/j.neuro.2026.103495.

Footnotes

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

CRediT authorship contribution statement

Beavers Daniel P: Writing – review & editing. Petkus Andrew J: Writing – review & editing, Methodology. Thompson Paul M: Writing – review & editing. Xinhui Wang: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Formal analysis, Data curation, Conceptualization. Margaret Gatz: Writing – review & editing. Elman Jeremy A: Writing – review & editing, Methodology, Data curation. Espeland Mark A: Writing – review & editing. Fennema-Notestin Christine: Writing – review & editing, Writing – original draft, Methodology, Data curation. Ira Frahmand: Writing – review & editing. Resnick Susan M: Writing – review & editing. Hagler Jr Donald J: Writing – review & editing, Methodology, Data curation. Finch Caleb E: Writing – review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. Rapp Stephen R: Writing – review & editing. Joshua Millstein: Writing – review & editing, Methodology. Kaufman Joel D: Writing – review & editing, Methodology, Data curation. Joshua Liu: Data curation. Braskie Meredith N: Writing – review & editing, Writing – original draft, Methodology, Data curation. Deydeep Kothapalli: Data curation. Salminen Lauren E: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Conceptualization. Hyung Jin Cho: Data curation. Kremen William S: Writing – review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. Jiu-Chiuan Chen: Writing – review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. Franz Carol E: Writing – review & editing, Supervision, Methodology, Funding acquisition, Conceptualization.

Data availability

VETSA data are available for researchers through a data request form available on the VETSA website: (https://psychiatry.ucsd.edu/research/programs-centers/vetsa/index.html). Data are not currently downloadable from a public website due to privacy and ethical restrictions established by the Vietnam Era Twin Registry (VETR) at the Seattle VA. WHIMS data, codebook, and analytic code used in this paper are held by the NIH-funded Coordinating Center of the Women’s Health Initiative at the Fred Hutchinson Cancer Research Center and may be accessed as described on the Women’s Health Initiative website: https://www.whi.org/page/working-with-whi-data.

References

Andresen EM, Malmgren JA, Carter WB, Patrick DL, 1994. Screening for Depression in Well Older Adults: Evaluation of a Short Form of the CES-D. Am. J. Prev. Med. 10 (2), 77–84. 10.1016/S0749-3797(18)30622-6. [DOI] [PubMed] [Google Scholar]
Benjamini Y, Hochberg Y, 1995. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J. R. Stat. Soc. Ser. B (Methodol.). 10.1111/j.2517-6161.1995.tb02031.x. [DOI] [Google Scholar]
Calderón-Garcidueñas L, Hernández-Luna J, Galaz-Montoya CI, Clouston SAP, Aiello-Mora MV, Amaro de Gante J, et al. 2026. Cortical, subcortical, and cerebellar atrophy and cognition deficits in Metropolitan Mexico City teens and young adults exposed to fine particulate matter (PM_2.5) - neurodegeneration is in progress. Front. Neurol. 17, 1772916. 10.3389/fneur.2026.1772916. [DOI] [PMC free article] [PubMed] [Google Scholar]
Casanova R, Wang X, Reyes J, Akita Y, Serre ML, Vizuete W, et al. 2016. A Voxel-Based Morphometry Study Reveals Local Brain Structural Alterations Associated with Ambient Fine Particles in Older Women. Front. Human. Neurosci. 10, 495. 10.3389/fnhum.2016.00495. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chandra M, Rai CB, Kumari N, Sandhu VK, Chandra K, Krishna M, et al. 2022. Air Pollution and Cognitive Impairment across the Life Course in Humans: A Systematic Review with Specific Focus on Income Level of Study Area. Int. J. Environ. Res. Public. Health 19 (3), 1405. 10.3390/ijerph19031405. [DOI] [PMC free article] [PubMed] [Google Scholar]
Charlson M, Szatrowski TP, Peterson J, Gold J, 1994. Validation of a combined comorbidity index. J. Clin. Epidemiol. 47 (11), 1245–1251. 10.1016/0895-4356(94)90129-5. [DOI] [PubMed] [Google Scholar]
Cho J, Jang H, Park H, Noh Y, Sohn J, Koh S-B, et al. 2023. Alzheimer’s disease- like cortical atrophy mediates the effect of air pollution on global cognitive function. Environ. Int. 171, 107703. 10.1016/j.envint.2022.107703. [DOI] [PubMed] [Google Scholar]
Cho J, Noh Y, Kim SY, Sohn J, Noh J, Kim W, et al. 2020. Long-Term Ambient Air Pollution Exposures and Brain Imaging Markers in Korean Adults: The Environmental Pollution-Induced Neurological EFfects (EPINEF) Study. Environ. Health Perspect. 128 (11), 117006. 10.1289/EHP7133. [DOI] [PMC free article] [PubMed] [Google Scholar]
Coker LH, Hogan PE, Bryan NR, Kuller LH, Margolis KL, Bettermann K, et al. 2009. Postmenopausal hormone therapy and subclinical cerebrovascular disease: the WHIMS-MRI Study. Neurology 72 (2), 125–134. 10.1212/01.wnl.0000339036.88842.9e. [DOI] [PMC free article] [PubMed] [Google Scholar]
Crous-Bou M, Gascon M, Gispert JD, Cirach M, Sánchez-Benavides G, Falcon C, et al. 2020. Impact of urban environmental exposures on cognitive performance and brain structure of healthy individuals at risk for Alzheimer’s dementia. Environ. Int. 138, 105546. 10.1016/j.envint.2020.105546. [DOI] [PubMed] [Google Scholar]
Dickerson BC, Bakkour A, Salat DH, Feczko E, Pacheco J, Greve DN, et al. 2009. The cortical signature of Alzheimer’s disease: regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals. Cereb. cortex 19 (3), 497–510. 10.1093/cercor/bhn113. [DOI] [PMC free article] [PubMed] [Google Scholar]
Diez Roux AV, Merkin SS, Arnett D, Chambless L, Massing M, Nieto FJ, et al. 2001. Neighborhood of residence and incidence of coronary heart disease. New. Engl. J. Med. 345 (2), 99–106. 10.1056/NEJM200107123450205. [DOI] [PubMed] [Google Scholar]
Eisen S, True W, Goldberg J, Henderson W, Robinette CD, 1987. The Vietnam Era Twin (VET) Registry: method of construction. Acta Genet. Med. Et. Gemellol. 36 (1), 61–66. 10.1017/s0001566000004591. [DOI] [PubMed] [Google Scholar]
Erickson LD, Gale SD, Anderson JE, Brown BL, Hedges DW, 2020. Association between Exposure to Air Pollution and Total Gray Matter and Total White Matter Volumes in Adults: A Cross-Sectional Study. Brain Sci. 10 (3). 10.3390/brainsci10030164. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fleisher AS, Chen K, Liu X, Ayutyanont N, Roontiva A, Thiyyagura P, et al. 2013. Apolipoprotein E ε4 and age effects on florbetapir positron emission tomography in healthy aging and Alzheimer disease. Neurobiol. Aging 34 (1), 1–12. 10.1016/j.neurobiolaging.2012.04.017. [DOI] [PubMed] [Google Scholar]
Fortea J, Sala-Llonch R, Bartrés-Faz D, Lladó A, Solé-Padullés C, Bosch B, et al. 2011. Cognitively preserved subjects with transitional cerebrospinal fluid ß-amyloid 1–42 values have thicker cortex in Alzheimer’s disease vulnerable areas. Biol. Psychiatry 70 (2), 183–190. 10.1016/j.biopsych.2011.02.017. [DOI] [PubMed] [Google Scholar]
Fortea J, Vilaplana E, Alcolea D, Carmona-Iragui M, Sánchez-Saudinos M-B, Sala I, et al. 2014. Cerebrospinal fluid β-amyloid and phospho-tau biomarker interactions affecting brain structure in preclinical Alzheimer disease. Ann. Neurol. 76 (2), 223–230. 10.1002/ana.24186. [DOI] [PubMed] [Google Scholar]
Franz CE, Gustavson DE, Elman JA, Fennema-Notestine C, Hagler DJ Jr, Baraff A, et al. 2023. Associations Between Ambient Air Pollution and Cognitive Abilities from Midlife to Early Old Age: Modification by APOE Genotype. J. Alzheimer’S. Dis. JAD 93 (1), 193–209. 10.3233/JAD-221054. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gitsels LA, Kulinskaya E, Wright NR, 2019. How medical advances and health interventions will shape future longevity. Br. Actuar. J. 24, e10. 10.1017/S1357321719000059. [DOI] [Google Scholar]
Harrison TM, Du R, Klencklen G, Baker SL, Jagust WJ, 2021. Distinct effects of beta-amyloid and tau on cortical thickness in cognitively healthy older adults. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 17 (7), 1085–1096. 10.1002/alz.12249. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hayden KM, Gaussoin SA, Hunter JC, Manson JE, Sachs BC, Shadyab AH, et al. 2019. Cognitive resilience among APOE ε4 carriers in the oldest old. Int. J. Geriatr. Psychiatry 34 (12), 1833–1844. 10.1002/gps.5199. [DOI] [PMC free article] [PubMed] [Google Scholar]
Heckbert SR, Kooperberg C, Safford MM, Psaty BM, Hsia J, McTiernan A, et al. 2004. Comparison of self-report, hospital discharge codes, and adjudication of cardiovascular events in the Women’s Health Initiative. Am. J. Epidemiol. 160 (12), 1152–1158. 10.1093/aje/kwh314. [DOI] [PubMed] [Google Scholar]
Heinzinger N, Maass A, Berron D, Yakupov R, Peters O, Fiebach J, et al. 2023. Exploring the ATN classification system using brain morphology. Alzheimer’S. Res. Ther. 15 (1), 50. 10.1186/s13195-023-01185-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
Henderson WG, Eisen S, Goldberg J, True WR, Barnes JE, Vitek ME, 1990. The Vietnam Era Twin Registry: a resource for medical research. Public. Health Rep. 105 (4), 368–373. 〈https://www.ncbi.nlm.nih.gov/pubmed/2116638〉. [PMC free article] [PubMed] [Google Scholar]
Ismail R, Parbo P, Madsen LS, Hansen AK, Hansen KV, Schaldemose JL, et al. 2020. The relationships between neuroinflammation, beta-amyloid and tau deposition in Alzheimer’s disease: a longitudinal PET study. J. Neuroinflamm. 17 (1), 151. 10.1186/s12974-020-01820-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jack CR Jr, Bennett DA, Blennow K, 2018. NIA-AA research framework: toward a biological definition of Alzheimer’s disease. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 〈https://alz-journals.onlinelibrary.wiley.com/doi/abs/10.1016/j.jalz.2018.02.018〉. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jack CR Jr, Wiste HJ, Weigand SD, Therneau TM, Lowe VJ, Knopman DS, et al. 2017. Defining imaging biomarker cut points for brain aging and Alzheimer’s disease. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 13 (3), 205–216. 10.1016/j.jalz.2016.08.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jaramillo SA, Felton D, Andrews L, Desiderio L, Hallarn RK, Jackson SD, et al. 2007. Enrollment in a brain magnetic resonance study: results from the Women’s Health Initiative Memory Study Magnetic Resonance Imaging Study (WHIMS-MRI). Acad. Radiol. 14 (5), 603–612. 10.1016/j.acra.2007.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
Johnson-Kozlow M, Rock CL, Gilpin EA, Hollenbach KA, Pierce JP, 2007. Validation of the WHI brief physical activity questionnaire among women diagnosed with breast cancer. Am. J. Health Behav. 31 (2), 193–202. 10.5555/ajhb.2007.31.2.193. [DOI] [PubMed] [Google Scholar]
Kind AJH, Buckingham WR, 2018. Making Neighborhood-Disadvantage Metrics Accessible - The Neighborhood Atlas. New. Engl. J. Med. 378 (26), 2456–2458. 10.1056/NEJMp1802313. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kirwa K, Szpiro AA, Sheppard L, Sampson PD, Wang M, Keller JP, et al. 2021. Fine-Scale Air Pollution Models for Epidemiologic Research: Insights From Approaches Developed in the Multi-ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Curr. Environ. Health Rep. 8 (2), 113–126. 10.1007/s40572-021-00310-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ko J, Sohn J, Noh Y, Koh S-B, Lee S-K, Kim S-Y, et al. 2024. Effects of ambient air pollution on brain cortical thickness and subcortical volume: A longitudinal neuroimaging study. Neuroepidemiology 1–11. 10.1159/000539467. [DOI] [PubMed] [Google Scholar]
Kremen WS, Franz CE, Lyons MJ, 2019. Current status of the Vietnam era twin study of aging (VETSA). Twin research human genetics official journal International Society Twin Studies 22 (6), 783–787. 10.1017/thg.2019.125. [DOI] [PubMed] [Google Scholar]
Kremen WS, Thompson-Brenner H, Leung Y-MJ, Grant MD, Franz CE, Eisen SA, et al. 2006. Genes, environment, and time: the Vietnam Era Twin Study of Aging (VETSA). Twin res. human genet. offic. j. Int. Soc. Twin Studies 9 (6), 1009–1022. 10.1375/183242706779462750. [DOI] [PubMed] [Google Scholar]
Kulick ER, Wellenius GA, Boehme AK, Joyce NR, Schupf N, Kaufman JD, et al. 2020. Long-term exposure to air pollution and trajectories of cognitive decline among older adults. Neurology 94 (17), e1782–e1792. 10.1212/WNL.0000000000009314. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lin Y-C, Fan K-C, Wu C-D, Pan W-C, Chen J-C, Chao Y-P, et al. 2024. Yearly change in air pollution and brain aging among older adults: a community-based study in Taiwan. Environ. Int. 190, 108876. 10.1016/j. envint.2024.108876. [DOI] [PubMed] [Google Scholar]
Livingston G, Huntley J, Sommerlad A, Ames D, Ballard C, Banerjee S, et al. 2020. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 396 (10248), 413–446. 10.1016/S0140-6736(20)30367-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lyons MJ, Genderson M, Grant MD, Logue M, Zink T, McKenzie R, et al. 2013. Gene-environment interaction of ApoE genotype and combat exposure on PTSD. Am. Journal Med. Genet. Part B Neuropsychiatric Genet. Offi. Publication Int. Soc. Psychiatric Genet. 162B (7), 762–769. 10.1002/ajmg.b.32154. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ma Y-H, Chen H-S, Liu C, Feng Q-S, Feng L, Zhang Y-R, et al. 2022. Association of Long-term Exposure to Ambient Air Pollution With Cognitive Decline and Alzheimer’s Disease-Related Amyloidosis. Biol. Psychiatry. 10.1016/j.biopsych.2022.05.017. [DOI] [PubMed] [Google Scholar]
Margolis, Lihong Qi KL, Brzyski R, Bonds DE, Howard BV, Kempainen S, et al. 2008. Validity of diabetes self-reports in the Women’s Health Initiative: comparison with medication inventories and fasting glucose measurements. Clin. Trials 5 (3), 240–247. 10.1177/1740774508091749. [DOI] [PMC free article] [PubMed] [Google Scholar]
McEvoy LK, Fennema-Notestine C, Roddey JC, Hagler DJ, Jr, Holland, D., Karow, D.S., et al. 2009. Alzheimer disease: quantitative structural neuroimaging for detection and prediction of clinical and structural changes in mild cognitive impairment. Radiology 251 (1), 195–205. 10.1148/radiol.2511080924. [DOI] [PMC free article] [PubMed] [Google Scholar]
Montal V, Vilaplana E, Alcolea D, Pegueroles J, Pasternak O, González-Ortiz S, et al. 2018. Cortical microstructural changes along the Alzheimer’s disease continuum. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 14 (3), 340–351. 10.1016/j.jalz.2017.09.013. [DOI] [PubMed] [Google Scholar]
Montal V, Vilaplana E, Pegueroles J, Bejanin A, Alcolea D, Carmona-Iragui M, et al. 2021. Biphasic cortical macro- and microstructural changes in autosomal dominant Alzheimer’s disease. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 17 (4), 618–628. 10.1002/alz.12224. [DOI] [PMC free article] [PubMed] [Google Scholar]
Nußbaum R, Lucht S, Jockwitz C, Moebus S, Engel M, Jöckel K-H, et al. 2020. Associations of Air Pollution and Noise with Local Brain Structure in a Cohort of Older Adults. Environ. Health Perspect. 128 (6), 67012. 10.1289/EHP5859. [DOI] [PMC free article] [PubMed] [Google Scholar]
Pegueroles J, Vilaplana E, Montal V, Sampedro F, Alcolea D, Carmona-Iragui M, et al. 2017. Longitudinal brain structural changes in preclinical Alzheimer’s disease. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 13 (5), 499–509. 10.1016/j.jalz.2016.08.010. [DOI] [PubMed] [Google Scholar]
Petkus AJ, Salminen LE, Wang X, Driscoll I, Millstein J, Beavers DP, et al. 2023. Alzheimer’s Related Neurodegeneration Mediates Air Pollution Effects on Medial Temporal Lobe Atrophy. medRxiv Prepr. Serv. Health Sci. 10.1101/2023.11.29.23299144. [DOI] [Google Scholar]
Pettigrew C, Soldan A, Zhu Y, Wang M-C, Moghekar A, Brown T, et al. 2016. Cortical thickness in relation to clinical symptom onset in preclinical AD. NeuroImage. Clin. 12, 116–122. 10.1016/j.nicl.2016.06.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
Power MC, Lamichhane AP, Liao D, Xu X, Jack CR, Gottesman RF, et al. 2018. The Association of Long-Term Exposure to Particulate Matter Air Pollution with Brain MRI Findings: The ARIC Study. Environ. Health Perspect. 126 (2), 027009. 10.1289/EHP2152. [DOI] [PMC free article] [PubMed] [Google Scholar]
Radloff LS, 1977. The CES-D Scale: A Self-Report Depression Scale for Research in the General Population. Appl. Psychol. Meas. 1 (3), 385–401. 10.1177/014662167700100306. [DOI] [Google Scholar]
Schaie K (1983). Longitudinal studies of adult psychological development. 10.1093/geronj/39.6.765a. [DOI] [Google Scholar]
Schoenborn CA, Heyman KM, 2009. Health characteristics of adults aged 55 years and over: United States, 2004–2007. Natl. Health Stat. Rep. (16), 1–31. 〈https://www.cdc.gov/nchs/data/nhsr/nhsr016.pdf〉. [PubMed]
Schöll M, Lockhart SN, Schonhaut DR, O’Neil JP, Janabi M, Ossenkoppele R, et al. 2016. PET Imaging of Tau Deposition in the Aging Human Brain. Neuron 89 (5), 971–982. 10.1016/j.neuron.2016.01.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
Schwarz CG, Gunter JL, Wiste HJ, Przybelski SA, Weigand SD, Ward CP, et al. 2016. A large-scale comparison of cortical thickness and volume methods for measuring Alzheimer’s disease severity. NeuroImage. Clin. 11, 802–812. 10.1016/j.nicl.2016.05.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shumaker SA, Reboussin BA, Espeland MA, Rapp SR, McBee WL, Dailey M, et al. 1998. The Women’s Health Initiative Memory Study (WHIMS): a trial of the effect of estrogen therapy in preventing and slowing the progression of dementia. Control. Clin. Trials 19 (6), 604–621. 10.1016/s0197-2456(98)00038-5. [DOI] [PubMed] [Google Scholar]
Spiegelhalter D, 2016. How old are you, really? Communicating chronic risk through 'effective age' of your body and organs. BMC Med Inf. Decis Mak, 16 (104). 10.1186/s12911-016-0342-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
Spotorno N, Strandberg O, Vis G, Stomrud E, Nilsson M, Hansson O, 2023. Measures of cortical microstructure are linked to amyloid pathology in Alzheimer’s disease. Brain a J. Neurol. 146 (4), 1602–1614. 10.1093/brain/awac343. [DOI] [PMC free article] [PubMed] [Google Scholar]
Torso M, Ridgway GR, Hardingham I, Schwarz AJ, Chance SA, Alzheimer’s Disease Neuroimaging Initiative, 2022. Vivo Detection of Changes Related to Cortical Columnar Organization and Neuroinflammation Across the AD Continuum. J. Prev. Alzheimer’S. Dis. 9 (4), 769–779. 10.14283/jpad.2022.59. [DOI] [PubMed] [Google Scholar]
Tsuang MT, Bar JL, Harley RM, Lyons MJ, 2001. The Harvard Twin Study of Substance Abuse: what we have learned. Harv. Rev. Psychiatry 9 (6), 267–279. 〈https://www.ncbi.nlm.nih.gov/pubmed/11600486〉. [PubMed] [Google Scholar]
Wang X, Salminen LE, Petkus AJ, Frahmand I, Millstein J, Beavers DP, et al. 2025. Association between late-life air pollution exposure and medial temporal lobe atrophy in older women. Neurotoxicology 112, 103378. 10.1016/j.neuro.2025.103378. [DOI] [PMC free article] [PubMed] [Google Scholar]
Whitsel EA, Quibrera PM, Smith RL, Catellier DJ, Liao D, Henley AC, Heiss G, 2006. Accuracy of commercial geocoding: assessment and implications. Epidemiol. Perspect. & Innov. EP+I 3, 8. 10.1186/1742-5573-3-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Williams ME, Elman JA, Bell TR, Dale AM, Eyler LT, Fennema-Notestine C, et al. 2023. Higher cortical thickness/volume in Alzheimer’s-related regions: protective factor or risk factor? Neurobiol. Aging 129, 185–194. 10.1016/j.neurobiolaging.2023.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
Younan D, Wang X, Casanova R, Barnard R, Gaussoin SA, Saldana S, et al. 2020. PM2.5 associated with gray matter atrophy reflecting increased Alzheimers risk in older women. Neurology. 10.1212/WNL.0000000000011149. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplementary results

NIHMS2198545-supplement-supplementary_results.docx^{(27.8KB, docx)}

supplementary tables and figures

NIHMS2198545-supplement-supplementary_tables_and_figures.docx^{(2.9MB, docx)}

Data Availability Statement

[R1] Andresen EM, Malmgren JA, Carter WB, Patrick DL, 1994. Screening for Depression in Well Older Adults: Evaluation of a Short Form of the CES-D. Am. J. Prev. Med. 10 (2), 77–84. 10.1016/S0749-3797(18)30622-6. [DOI] [PubMed] [Google Scholar]

[R2] Benjamini Y, Hochberg Y, 1995. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J. R. Stat. Soc. Ser. B (Methodol.). 10.1111/j.2517-6161.1995.tb02031.x. [DOI] [Google Scholar]

[R3] Calderón-Garcidueñas L, Hernández-Luna J, Galaz-Montoya CI, Clouston SAP, Aiello-Mora MV, Amaro de Gante J, et al. 2026. Cortical, subcortical, and cerebellar atrophy and cognition deficits in Metropolitan Mexico City teens and young adults exposed to fine particulate matter (PM_2.5) - neurodegeneration is in progress. Front. Neurol. 17, 1772916. 10.3389/fneur.2026.1772916. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R4] Casanova R, Wang X, Reyes J, Akita Y, Serre ML, Vizuete W, et al. 2016. A Voxel-Based Morphometry Study Reveals Local Brain Structural Alterations Associated with Ambient Fine Particles in Older Women. Front. Human. Neurosci. 10, 495. 10.3389/fnhum.2016.00495. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R5] Chandra M, Rai CB, Kumari N, Sandhu VK, Chandra K, Krishna M, et al. 2022. Air Pollution and Cognitive Impairment across the Life Course in Humans: A Systematic Review with Specific Focus on Income Level of Study Area. Int. J. Environ. Res. Public. Health 19 (3), 1405. 10.3390/ijerph19031405. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] Charlson M, Szatrowski TP, Peterson J, Gold J, 1994. Validation of a combined comorbidity index. J. Clin. Epidemiol. 47 (11), 1245–1251. 10.1016/0895-4356(94)90129-5. [DOI] [PubMed] [Google Scholar]

[R7] Cho J, Jang H, Park H, Noh Y, Sohn J, Koh S-B, et al. 2023. Alzheimer’s disease- like cortical atrophy mediates the effect of air pollution on global cognitive function. Environ. Int. 171, 107703. 10.1016/j.envint.2022.107703. [DOI] [PubMed] [Google Scholar]

[R8] Cho J, Noh Y, Kim SY, Sohn J, Noh J, Kim W, et al. 2020. Long-Term Ambient Air Pollution Exposures and Brain Imaging Markers in Korean Adults: The Environmental Pollution-Induced Neurological EFfects (EPINEF) Study. Environ. Health Perspect. 128 (11), 117006. 10.1289/EHP7133. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] Coker LH, Hogan PE, Bryan NR, Kuller LH, Margolis KL, Bettermann K, et al. 2009. Postmenopausal hormone therapy and subclinical cerebrovascular disease: the WHIMS-MRI Study. Neurology 72 (2), 125–134. 10.1212/01.wnl.0000339036.88842.9e. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] Crous-Bou M, Gascon M, Gispert JD, Cirach M, Sánchez-Benavides G, Falcon C, et al. 2020. Impact of urban environmental exposures on cognitive performance and brain structure of healthy individuals at risk for Alzheimer’s dementia. Environ. Int. 138, 105546. 10.1016/j.envint.2020.105546. [DOI] [PubMed] [Google Scholar]

[R11] Dickerson BC, Bakkour A, Salat DH, Feczko E, Pacheco J, Greve DN, et al. 2009. The cortical signature of Alzheimer’s disease: regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals. Cereb. cortex 19 (3), 497–510. 10.1093/cercor/bhn113. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R12] Diez Roux AV, Merkin SS, Arnett D, Chambless L, Massing M, Nieto FJ, et al. 2001. Neighborhood of residence and incidence of coronary heart disease. New. Engl. J. Med. 345 (2), 99–106. 10.1056/NEJM200107123450205. [DOI] [PubMed] [Google Scholar]

[R13] Eisen S, True W, Goldberg J, Henderson W, Robinette CD, 1987. The Vietnam Era Twin (VET) Registry: method of construction. Acta Genet. Med. Et. Gemellol. 36 (1), 61–66. 10.1017/s0001566000004591. [DOI] [PubMed] [Google Scholar]

[R14] Erickson LD, Gale SD, Anderson JE, Brown BL, Hedges DW, 2020. Association between Exposure to Air Pollution and Total Gray Matter and Total White Matter Volumes in Adults: A Cross-Sectional Study. Brain Sci. 10 (3). 10.3390/brainsci10030164. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R15] Fleisher AS, Chen K, Liu X, Ayutyanont N, Roontiva A, Thiyyagura P, et al. 2013. Apolipoprotein E ε4 and age effects on florbetapir positron emission tomography in healthy aging and Alzheimer disease. Neurobiol. Aging 34 (1), 1–12. 10.1016/j.neurobiolaging.2012.04.017. [DOI] [PubMed] [Google Scholar]

[R16] Fortea J, Sala-Llonch R, Bartrés-Faz D, Lladó A, Solé-Padullés C, Bosch B, et al. 2011. Cognitively preserved subjects with transitional cerebrospinal fluid ß-amyloid 1–42 values have thicker cortex in Alzheimer’s disease vulnerable areas. Biol. Psychiatry 70 (2), 183–190. 10.1016/j.biopsych.2011.02.017. [DOI] [PubMed] [Google Scholar]

[R17] Fortea J, Vilaplana E, Alcolea D, Carmona-Iragui M, Sánchez-Saudinos M-B, Sala I, et al. 2014. Cerebrospinal fluid β-amyloid and phospho-tau biomarker interactions affecting brain structure in preclinical Alzheimer disease. Ann. Neurol. 76 (2), 223–230. 10.1002/ana.24186. [DOI] [PubMed] [Google Scholar]

[R18] Franz CE, Gustavson DE, Elman JA, Fennema-Notestine C, Hagler DJ Jr, Baraff A, et al. 2023. Associations Between Ambient Air Pollution and Cognitive Abilities from Midlife to Early Old Age: Modification by APOE Genotype. J. Alzheimer’S. Dis. JAD 93 (1), 193–209. 10.3233/JAD-221054. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] Gitsels LA, Kulinskaya E, Wright NR, 2019. How medical advances and health interventions will shape future longevity. Br. Actuar. J. 24, e10. 10.1017/S1357321719000059. [DOI] [Google Scholar]

[R20] Harrison TM, Du R, Klencklen G, Baker SL, Jagust WJ, 2021. Distinct effects of beta-amyloid and tau on cortical thickness in cognitively healthy older adults. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 17 (7), 1085–1096. 10.1002/alz.12249. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] Hayden KM, Gaussoin SA, Hunter JC, Manson JE, Sachs BC, Shadyab AH, et al. 2019. Cognitive resilience among APOE ε4 carriers in the oldest old. Int. J. Geriatr. Psychiatry 34 (12), 1833–1844. 10.1002/gps.5199. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] Heckbert SR, Kooperberg C, Safford MM, Psaty BM, Hsia J, McTiernan A, et al. 2004. Comparison of self-report, hospital discharge codes, and adjudication of cardiovascular events in the Women’s Health Initiative. Am. J. Epidemiol. 160 (12), 1152–1158. 10.1093/aje/kwh314. [DOI] [PubMed] [Google Scholar]

[R23] Heinzinger N, Maass A, Berron D, Yakupov R, Peters O, Fiebach J, et al. 2023. Exploring the ATN classification system using brain morphology. Alzheimer’S. Res. Ther. 15 (1), 50. 10.1186/s13195-023-01185-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] Henderson WG, Eisen S, Goldberg J, True WR, Barnes JE, Vitek ME, 1990. The Vietnam Era Twin Registry: a resource for medical research. Public. Health Rep. 105 (4), 368–373. 〈https://www.ncbi.nlm.nih.gov/pubmed/2116638〉. [PMC free article] [PubMed] [Google Scholar]

[R25] Ismail R, Parbo P, Madsen LS, Hansen AK, Hansen KV, Schaldemose JL, et al. 2020. The relationships between neuroinflammation, beta-amyloid and tau deposition in Alzheimer’s disease: a longitudinal PET study. J. Neuroinflamm. 17 (1), 151. 10.1186/s12974-020-01820-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] Jack CR Jr, Bennett DA, Blennow K, 2018. NIA-AA research framework: toward a biological definition of Alzheimer’s disease. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 〈https://alz-journals.onlinelibrary.wiley.com/doi/abs/10.1016/j.jalz.2018.02.018〉. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R27] Jack CR Jr, Wiste HJ, Weigand SD, Therneau TM, Lowe VJ, Knopman DS, et al. 2017. Defining imaging biomarker cut points for brain aging and Alzheimer’s disease. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 13 (3), 205–216. 10.1016/j.jalz.2016.08.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R28] Jaramillo SA, Felton D, Andrews L, Desiderio L, Hallarn RK, Jackson SD, et al. 2007. Enrollment in a brain magnetic resonance study: results from the Women’s Health Initiative Memory Study Magnetic Resonance Imaging Study (WHIMS-MRI). Acad. Radiol. 14 (5), 603–612. 10.1016/j.acra.2007.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] Johnson-Kozlow M, Rock CL, Gilpin EA, Hollenbach KA, Pierce JP, 2007. Validation of the WHI brief physical activity questionnaire among women diagnosed with breast cancer. Am. J. Health Behav. 31 (2), 193–202. 10.5555/ajhb.2007.31.2.193. [DOI] [PubMed] [Google Scholar]

[R30] Kind AJH, Buckingham WR, 2018. Making Neighborhood-Disadvantage Metrics Accessible - The Neighborhood Atlas. New. Engl. J. Med. 378 (26), 2456–2458. 10.1056/NEJMp1802313. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] Kirwa K, Szpiro AA, Sheppard L, Sampson PD, Wang M, Keller JP, et al. 2021. Fine-Scale Air Pollution Models for Epidemiologic Research: Insights From Approaches Developed in the Multi-ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Curr. Environ. Health Rep. 8 (2), 113–126. 10.1007/s40572-021-00310-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] Ko J, Sohn J, Noh Y, Koh S-B, Lee S-K, Kim S-Y, et al. 2024. Effects of ambient air pollution on brain cortical thickness and subcortical volume: A longitudinal neuroimaging study. Neuroepidemiology 1–11. 10.1159/000539467. [DOI] [PubMed] [Google Scholar]

[R33] Kremen WS, Franz CE, Lyons MJ, 2019. Current status of the Vietnam era twin study of aging (VETSA). Twin research human genetics official journal International Society Twin Studies 22 (6), 783–787. 10.1017/thg.2019.125. [DOI] [PubMed] [Google Scholar]

[R34] Kremen WS, Thompson-Brenner H, Leung Y-MJ, Grant MD, Franz CE, Eisen SA, et al. 2006. Genes, environment, and time: the Vietnam Era Twin Study of Aging (VETSA). Twin res. human genet. offic. j. Int. Soc. Twin Studies 9 (6), 1009–1022. 10.1375/183242706779462750. [DOI] [PubMed] [Google Scholar]

[R35] Kulick ER, Wellenius GA, Boehme AK, Joyce NR, Schupf N, Kaufman JD, et al. 2020. Long-term exposure to air pollution and trajectories of cognitive decline among older adults. Neurology 94 (17), e1782–e1792. 10.1212/WNL.0000000000009314. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] Lin Y-C, Fan K-C, Wu C-D, Pan W-C, Chen J-C, Chao Y-P, et al. 2024. Yearly change in air pollution and brain aging among older adults: a community-based study in Taiwan. Environ. Int. 190, 108876. 10.1016/j. envint.2024.108876. [DOI] [PubMed] [Google Scholar]

[R37] Livingston G, Huntley J, Sommerlad A, Ames D, Ballard C, Banerjee S, et al. 2020. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 396 (10248), 413–446. 10.1016/S0140-6736(20)30367-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] Lyons MJ, Genderson M, Grant MD, Logue M, Zink T, McKenzie R, et al. 2013. Gene-environment interaction of ApoE genotype and combat exposure on PTSD. Am. Journal Med. Genet. Part B Neuropsychiatric Genet. Offi. Publication Int. Soc. Psychiatric Genet. 162B (7), 762–769. 10.1002/ajmg.b.32154. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R39] Ma Y-H, Chen H-S, Liu C, Feng Q-S, Feng L, Zhang Y-R, et al. 2022. Association of Long-term Exposure to Ambient Air Pollution With Cognitive Decline and Alzheimer’s Disease-Related Amyloidosis. Biol. Psychiatry. 10.1016/j.biopsych.2022.05.017. [DOI] [PubMed] [Google Scholar]

[R40] Margolis, Lihong Qi KL, Brzyski R, Bonds DE, Howard BV, Kempainen S, et al. 2008. Validity of diabetes self-reports in the Women’s Health Initiative: comparison with medication inventories and fasting glucose measurements. Clin. Trials 5 (3), 240–247. 10.1177/1740774508091749. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R41] McEvoy LK, Fennema-Notestine C, Roddey JC, Hagler DJ, Jr, Holland, D., Karow, D.S., et al. 2009. Alzheimer disease: quantitative structural neuroimaging for detection and prediction of clinical and structural changes in mild cognitive impairment. Radiology 251 (1), 195–205. 10.1148/radiol.2511080924. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R42] Montal V, Vilaplana E, Alcolea D, Pegueroles J, Pasternak O, González-Ortiz S, et al. 2018. Cortical microstructural changes along the Alzheimer’s disease continuum. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 14 (3), 340–351. 10.1016/j.jalz.2017.09.013. [DOI] [PubMed] [Google Scholar]

[R43] Montal V, Vilaplana E, Pegueroles J, Bejanin A, Alcolea D, Carmona-Iragui M, et al. 2021. Biphasic cortical macro- and microstructural changes in autosomal dominant Alzheimer’s disease. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 17 (4), 618–628. 10.1002/alz.12224. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R44] Nußbaum R, Lucht S, Jockwitz C, Moebus S, Engel M, Jöckel K-H, et al. 2020. Associations of Air Pollution and Noise with Local Brain Structure in a Cohort of Older Adults. Environ. Health Perspect. 128 (6), 67012. 10.1289/EHP5859. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R45] Pegueroles J, Vilaplana E, Montal V, Sampedro F, Alcolea D, Carmona-Iragui M, et al. 2017. Longitudinal brain structural changes in preclinical Alzheimer’s disease. Alzheimer’S. Dement. J. Alzheimer's. Assoc. 13 (5), 499–509. 10.1016/j.jalz.2016.08.010. [DOI] [PubMed] [Google Scholar]

[R46] Petkus AJ, Salminen LE, Wang X, Driscoll I, Millstein J, Beavers DP, et al. 2023. Alzheimer’s Related Neurodegeneration Mediates Air Pollution Effects on Medial Temporal Lobe Atrophy. medRxiv Prepr. Serv. Health Sci. 10.1101/2023.11.29.23299144. [DOI] [Google Scholar]

[R47] Pettigrew C, Soldan A, Zhu Y, Wang M-C, Moghekar A, Brown T, et al. 2016. Cortical thickness in relation to clinical symptom onset in preclinical AD. NeuroImage. Clin. 12, 116–122. 10.1016/j.nicl.2016.06.010. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R48] Power MC, Lamichhane AP, Liao D, Xu X, Jack CR, Gottesman RF, et al. 2018. The Association of Long-Term Exposure to Particulate Matter Air Pollution with Brain MRI Findings: The ARIC Study. Environ. Health Perspect. 126 (2), 027009. 10.1289/EHP2152. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] Radloff LS, 1977. The CES-D Scale: A Self-Report Depression Scale for Research in the General Population. Appl. Psychol. Meas. 1 (3), 385–401. 10.1177/014662167700100306. [DOI] [Google Scholar]

[R50] Schaie K (1983). Longitudinal studies of adult psychological development. 10.1093/geronj/39.6.765a. [DOI] [Google Scholar]

[R51] Schoenborn CA, Heyman KM, 2009. Health characteristics of adults aged 55 years and over: United States, 2004–2007. Natl. Health Stat. Rep. (16), 1–31. 〈https://www.cdc.gov/nchs/data/nhsr/nhsr016.pdf〉. [PubMed]

[R52] Schöll M, Lockhart SN, Schonhaut DR, O’Neil JP, Janabi M, Ossenkoppele R, et al. 2016. PET Imaging of Tau Deposition in the Aging Human Brain. Neuron 89 (5), 971–982. 10.1016/j.neuron.2016.01.028. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R53] Schwarz CG, Gunter JL, Wiste HJ, Przybelski SA, Weigand SD, Ward CP, et al. 2016. A large-scale comparison of cortical thickness and volume methods for measuring Alzheimer’s disease severity. NeuroImage. Clin. 11, 802–812. 10.1016/j.nicl.2016.05.017. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] Shumaker SA, Reboussin BA, Espeland MA, Rapp SR, McBee WL, Dailey M, et al. 1998. The Women’s Health Initiative Memory Study (WHIMS): a trial of the effect of estrogen therapy in preventing and slowing the progression of dementia. Control. Clin. Trials 19 (6), 604–621. 10.1016/s0197-2456(98)00038-5. [DOI] [PubMed] [Google Scholar]

[R55] Spiegelhalter D, 2016. How old are you, really? Communicating chronic risk through 'effective age' of your body and organs. BMC Med Inf. Decis Mak, 16 (104). 10.1186/s12911-016-0342-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] Spotorno N, Strandberg O, Vis G, Stomrud E, Nilsson M, Hansson O, 2023. Measures of cortical microstructure are linked to amyloid pathology in Alzheimer’s disease. Brain a J. Neurol. 146 (4), 1602–1614. 10.1093/brain/awac343. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R57] Torso M, Ridgway GR, Hardingham I, Schwarz AJ, Chance SA, Alzheimer’s Disease Neuroimaging Initiative, 2022. Vivo Detection of Changes Related to Cortical Columnar Organization and Neuroinflammation Across the AD Continuum. J. Prev. Alzheimer’S. Dis. 9 (4), 769–779. 10.14283/jpad.2022.59. [DOI] [PubMed] [Google Scholar]

[R58] Tsuang MT, Bar JL, Harley RM, Lyons MJ, 2001. The Harvard Twin Study of Substance Abuse: what we have learned. Harv. Rev. Psychiatry 9 (6), 267–279. 〈https://www.ncbi.nlm.nih.gov/pubmed/11600486〉. [PubMed] [Google Scholar]

[R59] Wang X, Salminen LE, Petkus AJ, Frahmand I, Millstein J, Beavers DP, et al. 2025. Association between late-life air pollution exposure and medial temporal lobe atrophy in older women. Neurotoxicology 112, 103378. 10.1016/j.neuro.2025.103378. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R60] Whitsel EA, Quibrera PM, Smith RL, Catellier DJ, Liao D, Henley AC, Heiss G, 2006. Accuracy of commercial geocoding: assessment and implications. Epidemiol. Perspect. & Innov. EP+I 3, 8. 10.1186/1742-5573-3-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R61] Williams ME, Elman JA, Bell TR, Dale AM, Eyler LT, Fennema-Notestine C, et al. 2023. Higher cortical thickness/volume in Alzheimer’s-related regions: protective factor or risk factor? Neurobiol. Aging 129, 185–194. 10.1016/j.neurobiolaging.2023.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R62] Younan D, Wang X, Casanova R, Barnard R, Gaussoin SA, Saldana S, et al. 2020. PM2.5 associated with gray matter atrophy reflecting increased Alzheimers risk in older women. Neurology. 10.1212/WNL.0000000000011149. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Air pollution is linked to divergent cortical thickness patterns in brain regions vulnerable to Alzheimer’s disease

Lauren E Salminen

Xinhui Wang

Andrew J Petkus

Christine Fennema-Notestine

Jeremy A Elman

Donald J Hagler Jr

Joel D Kaufman

Joshua Millstein

Meredith N Braskie

Joshua Liu

Hyung Jin Cho

Deydeep Kothapalli

Paul M Thompson

Daniel P Beavers

Mark A Espeland

Margaret Gatz

Susan M Resnick

Ira Frahmand

Stephen R Rapp

Caleb E Finch

Jiu-Chiuan Chen

William S Kremen

Carol E Franz

Abstract

Introduction:

Methods:

Results:

Discussion:

1. Background

2. Methods

2.1. Design overview

2.2. VETSA overview, participants, and covariates

2.3. WHIMS overview, participants, and covariates

2.4. Neuroimaging acquisition and processing for VETSA and WHIMS cohorts

2.5. Geocoding residential addresses

2.6. Air pollution assessments

2.7. Statistical approach for both datasets

3. Results

3.1. Descriptive information

Fig. 1. Flowchart of the two study cohorts.

Table 1.

Table 2.

3.2. Associations between air pollution exposures and CTh in the AD meta-ROI

Table 3.

3.2.1. Age-stratified analyses and age-by-exposure interactions on the AD meta-ROI

Fig. 2. Associations between air pollution exposures and AD meta-ROI across cohort-specific age ranges.

3.3. Associations between air pollution exposures and regional CTh

Fig. 3. Associations between air pollution exposures and regional cortical thickness.

4. Discussion

5. Conclusion

Supplementary Material

Acknowledgements

Funding

GLOSSARY

Appendix A. Supporting information

Footnotes

Data availability

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases