Older adults in more disadvantaged neighborhoods exhibit brain vascular dysfunction Authors: Kevin W Law, Allison C Engstrom, Trevor Lohman, Melanie Quiring, Arunima Kapoor, John Paul M Alitin, Danny JJ Wang, Aimée Gaubert, Kathleen Rodgers, Elizabeth Head, Celina I Valencia, Jillian L Joyce, Farrah Ferrer, Daniel A Nation Published: 2026 ABSTRACT Highlights•This study examines brain microvascular differences in older adults at rest.•Neighborhood deprivation is associated with reduced cerebrovascular reactivity.•Findings are independent of Alzheimer’s disease-related pathology.Keywords: Neighborhood disadvantage, Cerebrovascular reactivity, Brain health, Vascular dementia, Health disparities Introduction While neighborhood disadvantage, a social determinant of health, is implicated as an environmental risk factor for cognitive decline and dementia [1,2], little is known about the biological mechanisms involved. Meanwhile, dementia incidence is increasing, with annual new cases in the U.S. expected to double by 2060 [3]. In addition to Alzheimer’s disease-related (AD) neurodegeneration [4], cerebrovascular health has been suggested as an important mediating pathway between neighborhood disadvantage and cognitive decline [5]. Cerebrovascular reactivity (CVR) measures how well cerebral blood vessels respond to vasodilatory stimuli and regulate cerebral blood flow (CBF). Whole-brain CVR has been associated with cognitive ability in older adults [6]. Growing evidence proposes CVR, as assessed by pseudo-continuous arterial spin labeling (pCASL) magnetic resonance imaging (MRI), as a potentially useful neuroimaging marker of microvascular aging before the onset of cognitive impairment [7,8]. However, it remains unclear how CVR relates to the environment. No studies to date have examined CVR in the context of area-level deprivation, despite the importance of neighborhood socioeconomic factors in brain health. This exploratory study aims to characterize the relationship between neighborhood deprivation and CVR to better understand microvascular differences that may underlie disparities in brain health. Methods Participants Independently living older adults (≥ 55 years) without a history of clinical stroke or dementia were recruited from communities in Southern California as part of the Vascular Senescence and Cognition Study at the University of Southern California (USC) and the University of California, Irvine (UCI). Other exclusion criteria were current major neurological or psychiatric disorders, current use of medications affecting the central nervous system, MRI contraindications, history of moderate-to-severe traumatic brain injury, family history of dominantly inherited neurodegenerative disorders, and uncontrolled systemic illnesses. Demographic information, such as age, sex, education, race, and ethnicity, was self-reported. This study was approved by the Institutional Review Boards at USC and UCI. All participants gave written informed consent prior to participation and were treated in accordance with the Declaration of Helsinki. Data are available upon reasonable request. Neighborhood deprivation Relative neighborhood deprivation was quantified using the Area Deprivation Index (ADI), a composite of 17 neighborhood socioeconomic and housing characteristics from U.S. Census data. ADI data were retrieved from the Neighborhood Atlas website [9]. Twelve-digit census block group geographic identifiers (GEOIDs) were temporarily retrieved from available participant addresses. GEOIDs were then temporally matched to their corresponding California ADI state decile (1–10), with higher values (e.g., 10) indicating greater deprivation. Historical address data were unavailable, so enrollment addresses were treated as proxies for long-term exposure. To ensure physiological measurements reflect prolonged neighborhood exposure, analyses were conducted on data from participants’ most recent visit prior to any change in address. Neuroimaging Neuroimaging data were obtained through brain MRI on a 3T Siemens MAGNETOM Prisma scanner. Participants underwent a 3D T1-weighted magnetization-prepared rapid acquisition gradient-echo (MPRAGE) to examine brain structure and identify abnormalities (repetition time [TR] = 2300 ms; echo time [TE] = 2.98 ms; inversion time [TI] = 900 ms; flip angle = 9 deg; field of view [FOV] = 256 mm; resolution = 1.0 × 1.0 × 1.2 mm3; scan time = 9 min). Resting-state CBF was then obtained using 3D gradient- and spin-echo (GRASE) pCASL (TR = 5000 ms; TE at USC = 36.3 ms; TE at UCI = 37.46 ms; FOV = 240 mm; resolution = 2.5 × 2.5 × 3.4 mm3; slice thickness = 3.42 mm; number of slices = 24; labeling duration = 1517 ms; post-labeling delay = 2000 ms; 32 acquisitions [1 M0 image, 1 dummy image, and 30 alternating tag and control images yielding 15 tag-control pair images]; scan time = 5 min 25 s). Cerebral blood flow pCASL scans were pre-processed using the ASLtbx pipeline implemented in SPM12 within MATLAB, as previously described [10,11]. Pre-processing steps for pCASL scans included motion correction, co-registration to individual participants’ structural T1-weighted image, spatial smoothing, and tag-control subtraction. Absolute CBF values outside the expected physiological range of gray matter (< 10 or > 150 mL/100g tissue/min) were excluded [12,13]. Tag-control pairs were warped to Montreal Neurological Institute space and averaged to create mean CBF maps for each participant. Resulting maps were visually inspected for quality and gross abnormalities (e.g., large signal dropout). Partial volume correction was performed by applying participant-specific gray matter masks derived from the gray matter tissue class segmentation of T1-weighted structural images [14]. Segmented gray matter maps were thresholded at 0.3, binarized, and multiplied by the mean CBF maps to ensure CBF was limited to gray matter. Capnography and cerebrovascular reactivity For simultaneous capnographic monitoring of end-tidal carbon dioxide (ETCO2) during the pCASL scan, participants breathed normally into a nasal cannula connected to an M3015A sidestream CO2 extension module (Philips Medical Systems). The ETCO2 value of each breath was extracted from the raw recording. To correct for sampling tubing latency, ETCO2 time series were shifted by a pre-calibrated duration of time (10 s). Spontaneous CVR in the whole brain was quantified as the percent change in resting-state CBF per unit change in ETCO2 for consistency with our previous studies [8,15]. CVR values were generated using the following formula: APOE genotyping and AD biomarker quantification Fasting blood samples obtained from participants by venipuncture were used to determine ε4 allele carrier status of the apolipoprotein E gene (APOE) and quantify levels of plasma AD biomarker p-tau217. Following plasma separation, DNA was extracted from blood cell pellet fractions. APOE genotyping was conducted on extracted DNA using the TaqMan Single Nucleotide Polymorphism (SNP) Genotyping Assay (Thermo Fisher Scientific) and real-time polymerase chain reaction (PCR), as previously described [16]. APOE ε4 carriers were defined as having at least one APOE ε4 allele. Because APOE ε2 may be protective against AD and APOE ε4 increases risk, participants with the ε2/ε4 genotype were excluded from analyses. Plasma p-tau217 concentrations (pg/mL) were measured by single-molecule array using Simoa Technology (Quanterix). Vascular risk factors Vascular risk factors (VRFs) were assessed through clinical interview and exam and included history of hypertension, hyperlipidemia, diabetes, smoking, cardiovascular disease, atrial fibrillation, and transient ischemic attack. Total VRF count was dichotomized as low versus high burden (0–1 vs. 2+ VRFs) for use in statistical analyses. Two participants with missing hyperlipidemia data were retained for analyses because their VRF burden classification would not have changed regardless of hyperlipidemia status. Statistical analysis All statistical analyses were conducted using IBM SPSS Statistics 29 and R version 4.5.1. Data visualization was performed in GraphPad Prism version 11.0.2. Multiple linear regression tested the environmental effect of neighborhood deprivation, as measured by the ADI, on spontaneous CVR in the whole brain. Model 1 adjusted for covariates age and sex. Model 2 further adjusted for APOE ε4 carrier status and VRF burden. In a sensitivity analysis, the full model additionally adjusted for plasma AD biomarker p-tau217. The regression models were repeated with whole-brain resting-state CBF as the dependent variable to account for effects driven solely by perfusion. CVR outliers were removed if they were greater than ± 3 SD from the mean. After outlier screening, two participants were excluded. Results A total of 132 participants (N = 132) were included in the regression analyses. Participant characteristics are listed in Table 1. Age ranged from 55 years to 89 years. Years of education ranged from 9 to 20. Table 1Participant characteristics.CharacteristicTotal sample (N = 132)Age (years), mean (SD)69.3 (7.6)Sex (female), n (%)86 (65.2)Education (years), mean (SD)16.5 (2.3)Race and ethnicity, n (%) Asian18 (13.6) Black or African American6 (4.5) Hispanic or Latino5 (3.8) Non-Hispanic White103 (78.0)APOE genotype, n (%) ε2/ε32 (1.5) ε3/ε375 (56.8) ε3/ε449 (37.1) ε4/ε46 (4.5) ε4 carrier55 (41.7)Vascular risk factors, n (%) Hypertension49 (37.1) Hyperlipidemiaa67 (50.8) Diabetes19 (14.4) Smoking history45 (34.1) Cardiovascular disease9 (6.8) Atrial fibrillation8 (6.1) Transient ischemic attack4 (3.0) High burden (2+ VRFs)60 (45.5)ADI decile, median (IQR)3 (1–5)CBF (mL/100g tissue/min), mean (SD)48.3 (7.8)CVR (% ∆CBF/∆mmHg ETCO2), mean (SD)19.6 (12.0)P-tau217 (pg/mL),b mean (SD)0.36 (0.24)Abbreviations: ADI = Area Deprivation Index; VRF = vascular risk factor; CBF = cerebral blood flow; CVR = cerebrovascular reactivity.aMissing for 2 participants.bMissing for 19 participants. Participant characteristics. Abbreviations: ADI = Area Deprivation Index; VRF = vascular risk factor; CBF = cerebral blood flow; CVR = cerebrovascular reactivity. Missing for 2 participants. Missing for 19 participants. Higher ADI decile (more deprived) was associated with lower whole-brain CVR, adjusted for age and sex. This finding was unchanged after further adjusting for APOE ε4 carrier status and VRF burden (Table 2). In a sensitivity analysis of the full model, findings remained significant in the subset of participants with valid p-tau217 data, both before (B = −1.33, 95% CI [−2.21, −0.45], p = 0.003, n = 113) and after (B = −1.32, 95% CI [−2.20, −0.44], p = 0.004, n = 113) adjusting for p-tau217. Table 2Association between neighborhood deprivation and cerebrovascular reactivity.Unstandardized coefficients95% CI for BModelVariableBStd. errorβtpLower boundUpper bound1ADI decile−0.940.42−0.18−2.24.027−1.77−0.11Age−0.540.13−0.34−4.11<.001−0.80−0.28Sex0.942.070.040.45.650−3.155.032ADI decile−0.940.42−0.19−2.23.027−1.78−0.11Age−0.550.13−0.34−4.07<.001−0.81−0.28Sex1.232.120.050.58.563−2.965.42APOE ε4 status−1.422.03−0.06−0.70.484−5.432.59VRF burden−1.192.03−0.05−0.59.559−5.212.83Dependent variable: whole-brain cerebrovascular reactivity. Abbreviations: ADI = Area Deprivation Index; VRF = vascular risk factor. N = 132. Association between neighborhood deprivation and cerebrovascular reactivity. Dependent variable: whole-brain cerebrovascular reactivity. Abbreviations: ADI = Area Deprivation Index; VRF = vascular risk factor. N = 132. ADI decile was not associated with whole-brain resting-state CBF in either model 1 (adjusted for age and sex; B = −0.37, 95% CI [−0.91, 0.17], p = 0.174, N = 132) or 2 (further adjusted for APOE ε4 carrier status and VRF burden; B = −0.39, 95% CI [−0.93, 0.14], p = 0.146, N = 132). Plots of the age- and sex-adjusted regressions for whole-brain CVR and CBF are shown in Fig. 1. Mean CVR and CBF values for each ADI decile are available in Supplementary Table S1. Whole-brain regression plots adjusted for age and sex. (A) Relationship between ADI decile and spontaneous CVR. (B) Relationship between ADI decile and resting-state CBF. (C) Representative map of spontaneous CVR overlaid on a template anatomical image. (D) Representative map of resting-state CBF overlaid on a template anatomical image. Abbreviations: ADI = Area Deprivation Index; CVR = cerebrovascular reactivity; CBF = cerebral blood flow. N = 132. Discussion The present findings suggest that older adults living in more deprived environments exhibit weaker dilation of cerebral blood vessels at rest, providing a potential vascular mechanism linking neighborhood disadvantage to increased risk of vascular brain injury. Because our sample consists entirely of participants without clinical stroke or dementia, this relationship may represent an early-stage mechanism that occurs before the development of major clinical symptoms, but longitudinal studies are needed to establish a timeline for these changes. Importantly, the effect of neighborhood deprivation was adjusted for covariates relevant to CVR. It has been previously shown that greater vascular burden, as measured by the Framingham cardiovascular risk profile, is associated with reduced CVR [17]. Additionally, there is emerging evidence that the APOE ε4 allele may be implicated in CVR decline [18]. Our findings remained significant after controlling for VRF burden, APOE ε4 carrier status, and plasma AD biomarker p-tau217, suggesting a microvascular relationship independent of traditional vascular risk and AD-related pathology. Resting-state CBF was not associated with neighborhood deprivation, indicating that the observed effects are not explained by differences in resting perfusion or neuronal metabolism alone but may instead reflect vascular responsiveness. This aligns with previous findings linking deficits in CVR, but not CBF, to social isolation, another environmental risk factor for vascular brain injury [19]. Neighborhood socioeconomic status can influence characteristics of the local social and physical environment, collectively shaping health outcomes [20]. While the ADI, being an aggregate disadvantage measure, is a proxy for the combined effect of many environmental factors, it nonetheless reflects neighborhood-level burdens that may contribute to early cerebrovascular dysfunction. For example, neighborhood deprivation is generally associated with food desert prevalence [21] and greater air pollution [22], among other burdens. Thus, in addition to stress related to socioeconomic pressures, possible pathways explaining these CVR deficits include lifestyle and dietary constraints, as well as reduced air quality. Mental stress, high dietary sodium, and ambient particulate matter exposure have been linked to endothelial dysfunction and impaired vasodilatory signaling [[23], [24], [25]], which affect CVR [26]. Regardless of the specific mechanism(s) driving the relationship between neighborhood-level disadvantage and spontaneous CVR, these findings highlight the importance of considering the exposome, which comprises lifetime environmental exposures [27], when evaluating cerebrovascular health. These findings should be considered in light of several limitations. Many participants resided in relatively affluent areas, and there were fewer participants from highly deprived areas. Moreover, the sample was primarily composed of non-Hispanic White participants, which limits generalizability of the results. As previously noted, the ADI cannot isolate specific environmental factors that mechanistically explain the present findings [20], nor does it fully reflect variability in the neighborhood environment [28]. Furthermore, the ADI measures neighborhood socioeconomic conditions at a single time point and may not capture cumulative exposures or changes in the environment that influence physiological outcomes across the life course. Despite these limitations, this is, to our knowledge, the first study to analyze CVR in the context of socioeconomic environment. Longitudinal studies using a life-course approach and more diverse samples are warranted to better understand the relationship between the exposome and cerebral hemodynamics. Building on the present findings, these future studies will ultimately inform neighborhood-level interventions addressing environmental deprivation in communities that experience a high burden of cerebrovascular disease. CRediT authorship contribution statement Kevin W. Law: Writing – original draft, Visualization, Investigation, Formal analysis, Data curation, Conceptualization. Allison C. Engstrom: Writing – review & editing, Investigation, Data curation. Trevor Lohman: Writing – review & editing, Investigation, Data curation. Melanie Quiring: Writing – review & editing, Data curation. Arunima Kapoor: Writing – review & editing, Investigation, Data curation. John Paul M. Alitin: Writing – review & editing, Investigation, Data curation. Danny J.J. Wang: Writing – review & editing, Resources, Methodology. Aimée Gaubert: Writing – review & editing, Project administration, Investigation, Data curation. Kathleen Rodgers: Writing – review & editing, Resources, Methodology. Elizabeth Head: Writing – review & editing, Resources, Methodology. Celina I. Valencia: Writing – review & editing. Jillian L. Joyce: Writing – review & editing, Investigation, Data curation. Farrah Ferrer: Writing – review & editing, Investigation, Data curation. Daniel A. Nation: Writing – review & editing, Supervision, Methodology, Investigation, Funding acquisition, Conceptualization. Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Kathleen Rodgers reports a relationship with NeuTherapeutics that includes: consulting or advisory. Given their role as an editorial board member, Elizabeth Head had no involvement in the peer review of this article and had no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to another journal editor. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements This research was supported by National Institutes of Health grants (DAN: R01AG082073, R01AG064228), (AT, DAN: P01AG052350), (USC ADRC: P30AG066530). The authors graciously thank the participants and lab members of the Vascular Senescence and Cognition study for their contributions to this research. Footnotes Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cccb.2026.100562. Supplementary materials References Associated Data Supplementary Materials