Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Feb 19.
Published in final edited form as: J Gerontol B Psychol Sci Soc Sci. 2026 Feb 5;81(3):gbaf270. doi: 10.1093/geronb/gbaf270

Perceived Control, Brain Health, and Cognitive Reserve: Longitudinal Resilience Mechanisms in Black, White, and Hispanic Older Adults

Monica E Walters 1,*, Kiana A Scambray 1, Emily P Morris 1, Jordan D Palms 1, Robrielle M Pierce 1, Ketlyne Sol 2,3, Nicole Schupf 4, Adam M Brickman 4, Laura B Zahodne 1,3
PMCID: PMC12914702  NIHMSID: NIHMS2146645  PMID: 41416891

Abstract

OBJECTIVES:

Perceived control is a psychosocial factor consisting of internal (mastery) and external (constraints) control. We assessed whether mastery and constraints related to two longitudinal resilience mechanisms, brain maintenance and cognitive reserve.

METHODS:

Participants included White (n=199), Black (n=262), and Hispanic (n=319) older adults (Mage=74.48, SD=6.04; 63% women) from the Washington Heights-Inwood Columbia Aging Project. Brain health measures included cortical thickness in Alzheimer’s disease signature regions, total hippocampal volume, total gray matter volume, white matter hyperintensities, and their composite. Global cognition was a composite of four cognitive domains. Mastery and constraints were assessed separately in relation to longitudinal brain health and cognitive reserve (residual and moderation approaches). Univariate and bivariate latent growth curve models assessed patterns overall, and multiple-group models assessed differences by race and ethnicity. Covariates included age, sex/gender, years of education, intracranial volume, and race and ethnicity.

RESULTS:

Greater mastery related to greater baseline cortical thickness for Hispanic participants. Greater constraints related to greater baseline hippocampal volume for Black participants, but faster hippocampal volume decline for White participants. Controlling for brain health, greater constraints related to worse cognition initially and over time for the entire group. Greater brain health decline more strongly related to greater cognitive decline at higher levels of constraints among Hispanic and Black individuals.

DISCUSSION:

We found more evidence for cognitive reserve, rather than brain maintenance, as a resilience mechanism linking perceived control to cognitive health. Minoritized older adults may be particularly vulnerable to the negative impact of environmental constraints on cognitive reserve.

Keywords: cognition, health disparities, latent growth curve models, multiple-group models, locus of control

Various experiential factors may affect dementia risk through resilience mechanisms (Stern et al., 2023)–either by preserving brain health (brain maintenance; Nyberg et al., 2012) or by reducing the impact of worse brain health on cognition (cognitive reserve; Stern, 2002; Stern et al., 2020, 2023). Brain maintenance is a theoretical concept that suggests individuals maintain cognitive health through intact brain health; that is, age-/disease-related neuropathology does not develop due to genetic or experiential factors (Nyberg et al., 2012; Stern et al., 2020). Cognitive reserve is a theoretical concept that suggests experiential factors contribute to better than expected cognition given neuropathology (Stern, 2002; Stern et al., 2020). Racial and ethnic dementia inequalities may reflect not only differential exposure to experiential factors (Zahodne et al., 2021), but also the differential impact of these factors on brain health and/or cognitive reserve (Zahodne, 2021). Indeed, the most commonly studied dementia protective factor, education level, contributes to cognitive reserve more strongly among non-Hispanic White (henceforth White) older adults than non-Hispanic Black (henceforth Black) or Hispanic older adults (Avila et al., 2021). Therefore, additional studies are needed to evaluate factors beyond education that contribute to resilience among minoritized groups to identify intervention targets. One such experiential factor is perceived control.

Perceived control reflects one’s perception of control over their life circumstances (Lachman & Weaver, 1998) and includes two subdomains. Mastery (internal control) reflects the perception that life circumstances are under the individuals’ control, and Constraints (external control) reflects the perception that life circumstances are under external forces’ control (Ross & Mirowsky, 2013). Mastery and constraints are recognized as unique dimensions (Lachman & Weaver, 1998) that are shaped by different levels of environmental influence and may show divergent associations with brain health and cognitive reserve across racial and ethnic groups. Research suggests that social conditions may influence the development of control beliefs (Ross & Mirowsky, 2013). For example, someone may perceive constraints due to their experiences with structural inconsistency, or the inability to achieve societally legitimate goals (e.g., occupational advancement) due to inequitable allocation of resources (Ross & Mirowsky, 2013) (e.g., structural racism; Bailey et al., 2017). Thus, minoritized racial and ethnic groups who have disproportionate exposure to structural barriers may, in turn, perceive higher constraints and be more strongly impacted by this exposure when considering brain and cognitive health outcomes (Kessler, 1979).

Higher mastery and lower constraints relate to better cognition (Anderson et al., 2018; Du et al., 2024; Sutin et al., 2018; Wight et al., 2003; Windsor & Anstey, 2008). However, when considering race and ethnicity, some differences emerge. First, cognitive disparities may relate to level differences (Ward et al., 2019) in perceived control. Indeed, some suggest Black and Hispanic individuals have lower perceived control compared to White individuals (Lachman et al., 2011; Ross & Mirowsky, 2013), which may reflect both structural barriers and cultural differences. In two studies using a nationally representative sample, a cognitive disparity existed among participants with high mastery, such that Black participants had increased cognitive impairment and dementia risk compared to White participants with equivalent mastery (Nkwata et al., 2021, 2022). These studies suggest that the protective effects of mastery on cognition may be limited to White older adults. Others suggest that constraints may be particularly important to racial and ethnic differences in cognition. For example, studies in regional (Zahodne et al., 2021) and national (Zahodne et al., 2019) samples, including Black, White, and Hispanic older adults, and a national sample including Black and White adults (Zahodne et al., 2017), suggest that constraints partially mediated racial and ethnic disparities in multiple cognitive outcomes. Additionally, in a racially and geographically diverse cognitive training trial, constraints helped explain some of the race-related differences in training gains from cognitive interventions (Zahodne, Meyer, et al., 2015). Considering the separate effects of perceived control on cognitive health by race and ethnicity will help clarify associations.

Less is known about whether perceived control relates to cognition through resilience mechanisms, including brain maintenance and cognitive reserve. Some studies suggest that perceived control relates directly to brain health (Hashimoto et al., 2015; Pruessner et al., 2005). Specifically, among Japanese younger adults, higher overall perceived control (i.e., a measure that combines mastery and constraints) correlated positively with gray matter volume in the anterior cingulate cortex, striatum, ventrolateral prefrontal cortices, and right anterior insular cortex (Hashimoto et al., 2015). Additionally, among Canadian adults, higher mastery correlated positively with hippocampal volume (Pruessner et al., 2005). Although providing initial support for perceived control operating through the brain maintenance hypothesis, these studies have limitations, including (1) not specifically examining constraints, (2) limited generalizability to older adults due to the small sample and convenience sampling methodology (Pruessner et al., 2005), and (3) cross-sectional measurement of brain health. Further, these studies did not report on the racial and ethnic makeup of their samples, preventing assessment of differences across racial or ethnic identity.

Regarding the cognitive reserve hypothesis, one study found that higher overall perceived control was protective of episodic memory among participants with low hippocampal volume (Zahodne et al., 2018), with effects more pronounced for Hispanic and Black older adults than for White older adults. Therefore, perceived control may be a particularly important contributor to cognitive reserve among minoritized older adults. However, there is still a need for longitudinal studies of cognitive reserve with diverse older adults. Longitudinal designs are optimal to assess resilience mechanisms in cognitive aging (Stern et al., 2020, 2023), since changes in brain health and cognition are needed to isolate aging processes from long-standing individual differences (e.g., due to genetics and/or early life factors). Although evidence is limited, it is possible that perceived control operates through both brain and cognitive resilience mechanisms.

The goal of the current study was to clarify resilience mechanisms of perceived control across racial and ethnic groups. First, we assessed whether perceived control related to longitudinal brain health (Figure 1). Our first hypothesis was that greater perceived control (i.e., higher mastery and lower constraints) would relate to slower decline in brain volume and cortical thickness, and slower accumulation of white matter hyperintensities (WMH), providing evidence for brain maintenance. Second, we assessed whether perceived control operated through longitudinal cognitive reserve (Figure 2); that is, does perceived control predict better cognition than expected given brain health? We operationalized cognitive reserve through “residual” and “moderation” approaches, as current guidelines recommend both approaches as effective demonstrations of cognitive reserve (Stern et al., 2020, 2023), but little work has directly compared them. Specifically, we considered the effect of perceived control on cognitive trajectories after controlling for brain health trajectories (residual approach) and whether perceived control moderated the association between brain health trajectories and cognitive trajectories (moderation approach). We hypothesized that greater perceived control would relate to better cognitive trajectories independent of brain health (Hypothesis 2a; residual approach) and would attenuate associations between brain health and cognition (Hypothesis 2b; moderation approach). Third, we assessed whether patterns of associations between perceived control, brain health, and cognitive reserve differed across racial and ethnic groups. By using multiple-group models to assess effect modification, we hypothesized that associations between (1) perceived control and brain health (Hypothesis 3a) and (2) perceived control and cognitive reserve (Hypothesis 3b) would be stronger among Hispanic and Black individuals compared to White individuals.

Figure 1. Latent growth curve model illustrating perceived control in relation to brain health.

Figure 1.

Open in a new tab

Note. The model estimates a latent intercept and a latent slope of brain health across 3 MRI scans, with slope parameters set to the average years since baseline. Brain health indicators of cortical thickness in Alzheimer’s disease signature regions, total gray matter volume, total hippocampal volume, and white matter hyperintensities were assessed in separate models. Mastery and constraints were assessed in separate models.

Figure 2. Bivariate latent growth curve model illustrating perceived control in relation to cognitive reserve.

Figure 2.

Open in a new tab

Note. The model estimates associations between brain health and cognition, with slope parameters set to the average years since baseline. Brain health was a z-scored composite of the four brain health indicators (cortical thickness in Alzheimer’s disease signature regions, total gray matter volume, total hippocampal volume, white matter hyperintensities). Cognitive reserve was assessed in two ways. First, the associations between perceived control (i.e., mastery, constraints) and cognition intercept and slope were assessed when controlling for brain health intercept and slope (residual approach; solid, thick blue lines between perceived control and global cognition residual). Second, perceived control (i.e., mastery, constraints) was assessed as a moderator of the associations between brain health and cognition (moderation approach; dotted, thick blue lines between perceived control and brain health-global cognition associations). Covariates in both approaches included age, sex/gender, race and ethnicity, and years of education.

Methods

Participants

Participants were drawn from the Washington Heights-Inwood Columbia Aging Project (WHICAP; Manly et al., 2005; Tang et al., 2001). WHICAP is approved by the Institutional Review Board of Columbia University, and participants provide written informed consent. WHICAP is an ongoing, community-based longitudinal cohort study based in Manhattan, New York. WHICAP had three recruitment waves, 1992, 1999, and 2009-present, with follow-up every 18–24 months. Demographic information came from participants’ initial study assessment, which occurred between 1994 and 2022 for the current sample. An MRI sub-study began in 2004 (Brickman et al., 2008), and a psychosocial module was administered beginning in 2017. Prior work found that those who participated in the MRI sub-study were younger, less likely to be women, and more likely to be Black than those who refused, but there were similar education levels between groups (Brickman et al., 2008). Comparisons with those who were not scanned yielded similar results (Brickman et al., 2008). For the current study, participants were initially included if they completed the psychosocial module and had at least one MRI scan matched with a cognitive assessment (N=874). “Baseline” is defined as the first time a participant had a cognitive assessment and an MRI scan within the same year (Mdifference=60 days) and occurred between 2010/2011 and 2022/2023 for the current sample. Data for perceived control came from the psychosocial module and occurred within two years, on average, from participants’ baseline assessment. Participants were excluded if they had dementia at baseline or the wave in which perceived control was assessed (n=63); were missing data on sociodemographic covariates (n=6); or did not identify as Black, White, or Hispanic (n=25), leaving a final analytic sample of 780. Participants were followed for up to 11 years (M=4.21, SD=2.75), with up to five cognitive (M=3.17, SD=1.35) and three MRI (M=1.64, SD=0.72) assessments included in the current study.

Measures

MRI

MRI scans were completed on a 3T scanner at Columbia University Medical Center as previously described (Turney et al., 2023). Cortical thickness and volumetric measures based on T1-weighted images were derived from FreeSurfer version 6.0 (http://surfer.nmr.mgh.harvard.edu/). White matter hyperintensities were derived from T2-weighted FLAIR images as previously described (Brickman et al., 2009, 2011, 2012).

We included four brain health indicators: average cortical thickness across nine Alzheimer’s disease (AD) signature regions (mm) (Dickerson et al., 2009), total gray matter volume (mm3), total hippocampal volume (bilateral volume summed; mm3), and log-transformed total WMH volume. Baseline estimated total intracranial volume (ICV; mm3) was a covariate in volumetric analyses assessing brain maintenance and was used to adjust volumetric brain health indicators prior to including them in a brain health composite for cognitive reserve analyses. To counter excessive variance in analyses, volumetric measures were converted to cm3, then total gray matter volume was additionally divided by 10, whereas ICV was additionally divided by 100.

Perceived Control

Participants were administered a modified 12-item version of the Perceived Control Scale (Lachman & Weaver, 1998), with four items measuring mastery (e.g., “I can do just about anything I really set my mind to”) and eight items measuring constraints (e.g., “What happens in my life is often beyond my control”). Participants indicated whether they agreed or not (yes=1, no=0) with items in each subdomain, with separate scores for mastery and constraints summed. Analyses were conducted separately for mastery (α=.69) and constraints (α=.76), since they reflect two independent subdomains (Lachman & Weaver, 1998), and prior work using WHICAP data suggests that patterns found for overall perceived control seem to be driven by constraints in this sample (Zahodne et al., 2018).

Global Cognition

WHICAP uses a comprehensive neuropsychological battery to assess cognition. A previous factor analysis identified four cognitive domains that were invariant across language groups (i.e., measuring the underlying cognitive construct similarly; Siedlecki et al., 2010): processing speed, language, memory, and visuospatial functioning. The specific neuropsychological tests that make up each domain have been previously described (Siedlecki et al., 2010) and can be found in Supplementary Table 1. Z-scored composites of these domains were averaged to calculate a global cognition score, with up to two missing domain scores allowed for inclusion in the composite.

Covariates

Covariates included baseline age (range: 62–93), sex/gender (male=0; female=1), years of education (range: 0–20), race and ethnicity (two dummy-coded variables for Black and Hispanic used for analyses in the full sample only), and ICV.

Statistical Analyses

Data cleaning and descriptive statistics were conducted in SAS v 9.4. Descriptive statistics characterized the sample overall and by race and ethnicity. Differences between racial and ethnic groups were assessed with Analyses of Variance (ANOVAs); follow-up Tukey HSD tests described pairwise differences. Unconditional latent growth curve models (LGCMs) characterized longitudinal changes in brain health and cognition. Conditional LGCMs assessed adjusted associations between perceived control and global cognition. LGCMs, bivariate LGCMs, and multiple-group models (described in more detail below) were conducted in Mplus v. 8.11 (Muthén & Muthén, 2017) and use full information maximum likelihood estimation to address missing data in the outcomes. Significance was assessed at p<.05 (two-tailed).

Perceived Control and Longitudinal Brain Health

LGCMs assessed how mastery and/or constraints related to changes in brain health across three waves. Analyses controlled for race and ethnicity, age, sex/gender, ICV (for volumetric analyses), and years of education. A latent intercept and slope were derived separately for each of the four brain health indicators (Figure 1). Growth parameters for the slope factor corresponded to the average years since baseline (0, 3.99, 6.81). To better understand potential mechanisms, we considered each of the brain health indicators separately.

Perceived Control and Longitudinal Cognitive Reserve

Bivariate LGCMs assessed how perceived control related to trajectories of cognitive reserve (i.e., including trajectories for both brain health and cognition). Specifically, we used two common approaches to examine whether mastery and/or constraints contributed to cognitive reserve (Figure 2): “residual” and “moderation”. For both approaches, we operationalized brain health using a composite of the four brain health indicators (i.e., total gray matter volume, total hippocampal volume, cortical thickness of AD signature regions, WMH) rather than the individual brain health indicators because (1) cognitive reserve analyses focus on cognitive variance that is unexplained by overall brain health rather than any individual measure (Stern et al., 2023) and (2) to reduce the number of models to run for parsimony, as we did not have hypotheses regarding whether perceived control would uniquely influence cognitive variance unexplained by only one particular brain health indicator (residual) or uniquely moderate the relationship between one particular brain health indicator and cognition (moderation). To create the composite, first, total hippocampal volume and total gray matter volume at each wave were adjusted for ICV by regressing these measures on ICV and saving the residual (Jack et al., 1998; Voevodskaya et al., 2014). Next, z-scores based on means and standard deviations from baseline were computed for each of the four brain health indicators at each of the three waves. WMH values were reversed to be on the same scale as the other brain health indicators (i.e., higher values mean better brain health). A brain health composite at each of the three waves was calculated as the average of the four z-scored brain health indicators for that wave.

In the “residual” approach, we assessed associations between perceived control and cognitive trajectories when accounting for brain health trajectories (Zahodne, Manly, et al., 2015). A LGCM estimated an intercept and slope of brain health based on the brain health composite scores across three waves, with the growth parameters set to the average years since baseline (i.e., 0, 3.99, 6.81). A separate LGCM estimated an intercept and slope of cognition based on the cognitive composite scores across five waves. Growth parameters for the cognition slope corresponded to the average years since baseline (0, 2.06, 4.11, 5.79, 7.46). The brain health growth curve was used to control the cognition growth curve for brain health trajectories in the estimation of the cognitive reserve residual (i.e., difference between observed and expected cognitive trajectories given brain health).

In the “moderation” approach (Stern et al., 2020), we assessed whether mastery and/or constraints moderated associations between brain health trajectories and cognitive trajectories. Latent interactions between perceived control and the latent brain health intercept and slope were estimated with the XWITH command in Mplus. Separate models created interaction terms between perceived control and latent brain health intercept (in relation to cognition intercept and cognition slope) and between perceived control and latent brain health slope (in relation to cognition slope).

Racial and Ethnic Differences in Associations

Models for the full sample were repeated in a multiple-group framework where race and ethnicity served as the grouping variable, with pairwise comparisons made between each group. All covariate effects were fixed to equivalence between groups. Fixed models set all parameters of interest (e.g., associations involving perceived control) to equivalence between groups. A series of free models systematically freed the association between one parameter of interest (e.g., the association between mastery and hippocampal volume intercept) between groups. A chi-square test determined whether freeing each individual parameter significantly improved model fit.

Results

Table 1 provides descriptive statistics for the full sample and by race and ethnicity. Hispanic and Black participants reported greater mastery than White participants. Hispanic participants reported greater constraints than Black and White participants. Cortical thickness in the AD signature regions, total gray matter volume, and total hippocampal volume were greater among White than Black and Hispanic participants, whose brain health indicators were similar to each other. Black participants had more WMH than White and Hispanic participants. Regarding cognition, Hispanic participants had the lowest scores, followed by Black, then White participants.

Table 1.

Descriptive Characteristics at Baseline of Sample Overall and by Race and Ethnicity.

Variable Full Sample (n=780) Non-Hispanic Black (n=262) Hispanic (n=319) Non-Hispanic White (n=199) Overall ANOVA Post-Hoc Difference
M ± SD or N (%) M ± SD or N (%) M ± SD or N (%) M ± SD or N (%)
Age 74.48 ± 6.04 73.96 ± 6.16 75.29 ± 6.29 73.86 ± 5.32 F(2,777)=4.91, p=.008 B = W < H
Sex/Gender F(2,777)=5.32, p=.005 B = H ≠ W
Male 286 (36.67%) 87 (33.21%) 107 (33.54%) 92 (46.23%)
Female 494 (63.33%) 175 (66.79%) 212 (66.46%) 107 (53.77%)
Years of Education 12.48 ± 4.57 13.80 ± 2.84 9.09 ± 4.17 16.18 ± 2.98 F(2,777)=282.95, p<.001 H < B < W
Brain Health
 Cortical Thickness (mm) 2.61 ± 0.12 2.59 ± 0.11 2.61 ± 0.11 2.65 ± 0.12 F(2,777)=11.85, p<.001 B = H < W
 Total Gray Matter Volume (mm3) 547,644.68 ± 53,630.75 536,879.08 ± 51,719.60 538,174. 51 ± 48,442.72 576,999.29 ± 53,453.94 F(2,777)=44.54, p<.001 B = H < W
 Total Hippocampal Volume (mm3) 7,155.98 ± 845.67 7,073.35 ± 889.23 7,092.12 ± 816.12 7,367.13 ± 800.77 F(2,777)=8.52, p<.001 B = H < W
 White Matter Hyperintensities (log mm) 7.92 ± 0.84 8.09 ± 0.88 7.84 ± 0.78 7.81 ± 0.84 F(2,777)=8.97, p<.001 H = W < B
Perceived Control
 Mastery 3.64 ± 0.82 3.79 ± 0.62 3.74 ± 0.66 3.27 ± 1.13 F(2,742)=26.68, p<.001 B = H > W
 Constraints 1.52 ± 1.87 1.20 ± 1.58 2.10 ± 2.13 0.97 ± 1.45 F(2,725)=27.91, p<.001 B = W < H
 Global Cognition 0.58 ± 0.55 0.66 ± 0.44 0.22 ± 0.50 1.04 ± 0.35 F(2,776)=217.10, p<.001 H < B < W
Open in a new tab

Note. M = mean; SD = standard deviation; B = non-Hispanic Black; W = non-Hispanic White; H = Hispanic; ANOVA = Analysis of Variance. Differences between racial and ethnic groups were assessed with Analyses of Variance with post-hoc Tukey HSD tests. Significant post-hoc differences were assessed at p<.05.

Unconditional LGCMs described changes in brain health and cognition over time for all participants (Supplementary Table 2). Each brain health indicator worsened linearly over time (i.e., cortical thickness, gray matter, and hippocampal volume decline; WMH accumulation). There was significant interindividual variability in the intercept for all brain health indicators, and significant interindividual variability in the slope for hippocampal volume and WMH, but not for total gray matter volume or cortical thickness. Cognition declined linearly over time and exhibited interindividual variability in the intercept and slope.

Supplementary Table 3 presents associations between perceived control and cognition. Mastery was unrelated to cognition intercept and slope. More constraints related to lower cognition intercept (β=−0.177, 95% CI:[−0.231, −0.124]) and faster cognitive decline (β=−0.177, 95% CI:[−0.315, −0.040]). Supplementary Table 4 presents the summary of chi-square differences from the multiple-group models assessing whether associations between perceived control and cognition differed across White, Black, and Hispanic participants. No differences in associations between perceived control and cognition intercept or slope were found for any comparison.

Perceived Control and Longitudinal Brain Health

LGCMs assessed whether mastery and/or constraints related to the latent intercept or slope of the four brain health indicators. In the entire sample, neither mastery nor constraints related to the intercepts or slopes of cortical thickness, total gray matter volume, total hippocampal volume, or WMH (Table 2).

Table 2.

Standardized Associations Between Perceived Control and Brain Health in the Overall Sample.

Variable Model Parameters Model Fit
β 95% CI RMSEA CFI SRMR
LL UL Value 95% CI p
Mastery (n=745)
Cortical Thickness .037 .003, .065 .747 .990 .110
 Mastery → Intercept 0.039 −0.031 0.110
 Mastery → Slope 0.011 −0.145 0.167
Total Gray Matter Volume .024 .000, .053 .929 .998 .059
 Mastery → Intercept 0.008 −0.039 0.054
 Mastery → Slope 0.034 −0.079 0.148
Total Hippocampal Volume .028 .000, .056 .891 .996 .062
 Mastery → Intercept 0.038 −0.025 0.101
 Mastery → Slope 0.005 −0.104 0.113
White Matter Hyperintensities .038 .011, .062 .776 .990 .112
 Mastery → Intercept −0.050 −0.121 0.022
 Mastery → Slope −0.034 −0.161 0.093
Constraints (n=728)
Cortical Thickness .027 .000, .057 .884 .995 .057
 Constraints → Intercept −0.014 −0.087 0.058
 Constraints → Slope 0.133 −0.069 0.334
Total Gray Matter Volume .021 .000, .051 .944 .998 .052
 Constraints → Intercept −0.029 −0.078 0.020
 Constraints → Slope 0.035 −0.105 0.174
Total Hippocampal Volume .025 .000, .054 .914 .997 .064
 Constraints → Intercept −0.012 −0.078 0.054
 Constraints → Slope −0.016 −0.155 0.122
White Matter Hyperintensities .037 .009, .062 .781 .990 .112
 Constraints → Intercept 0.071 −0.003 0.145
 Constraints → Slope 0.103 −0.054 0.261
Open in a new tab

Note. β=Standardized Parameter Estimate; CI=Confidence Interval; LL=Lower Limit; UL=Upper Limit; RMSEA=Root Mean Square Error of Approximation; CFI=Comparative Fit Index; SRMR=Standardized Root Mean Square Residual. Models control for age, sex/gender, years of education, and race and ethnicity. Intracranial volume was also included as a covariate in total gray matter volume and total hippocampal volume analyses.

Perceived Control and Longitudinal Cognitive Reserve

Next, we assessed whether mastery and/or constraints contributed to cognitive reserve using complementary “residual” and “moderation” approaches in the full sample (Table 3).

Table 3.

Standardized Associations Between Perceived Control and Longitudinal Cognitive Reserve.

Variable Model Parameters Model Fit
β 95% CI RMSEA CFI SRMR Log-likelihood AIC BIC
LL UL Value 95% CI p
Residual Approach
Mastery .051 .043, .059 .427 .965 .069
 Mastery → Cognition Intercept 0.014 −0.039 0.068
 Mastery → Cognition Slope 0.066 −0.055 0.187
Constraints .051 .043, .060 .379 .965 .073
 Constraints → Cognition Intercept −0.175 −0.229 −0.121
 Constraints → Cognition Slope −0.175 −0.310 −0.039
Moderation Approach a
Mastery
 Brain Intercept Model −1,822.103 3,714.207 3,875.675
  Mastery × Brain Intercept → Cognition Intercept −0.028 −0.076 0.021
  Mastery × Brain Intercept → Cognition Slope −0.044 −0.178 0.090
 Brain Slope Model −1,822.634 3,713.268 3,870.123
  Mastery × Brain Slope → Cognition Slope 0.069 −0.104 0.242
Constraints
 Brain Intercept Model −1,767.134 3,604.268 3,764.929
  Constraints × Brain Intercept → Cognition Intercept 0.006 −0.055 0.067
  Constraints × Brain Intercept → Cognition Slope 0.107 −0.092 0.306
 Brain Slope Model −1,762.244 3,592.489 3,748.559
  Constraints × Brain Slope → Cognition Slope 0.291 0.031 0.550
Open in a new tab

Note. β=Standardized Parameter Estimate; CI=Confidence Interval; LL=Lower Limit; UL=Upper Limit; RMSEA=Root Mean Square Error of Approximation; CFI=Comparative Fit Index; SRMR=Standardized Root Mean Square Residual; AIC=Akaike Information Criterion; BIC=Bayesian Information Criterion. Mastery: n=745; Constraints: n=728. Models control for age, sex/gender, years of education, and race and ethnicity. Bolded values indicate significant effects at p<.05.

a

Models with latent interactions were run with Mplus “ANALYSIS” parameters “TYPE=RANDOM” and “ALGORITHM=INTEGRATION” which do not provide the same fit statistics.

Residual Approach to Cognitive Reserve

In models controlling for brain health trajectories, there was no evidence that mastery related to cognition intercept or slope beyond the effect of brain health. In contrast, more constraints related to worse cognitive trajectories after accounting for brain health trajectories. Specifically, more constraints relate to lower initial cognitive reserve and faster decline in cognitive reserve (i.e., worse cognitive trajectories than expected given brain health).

Moderation Approach to Cognitive Reserve

Next, we assessed whether mastery and/or constraints moderated associations between brain health and cognition trajectories. We found no evidence of interactions between mastery or constraints and the brain health intercept in predicting cognition intercept or slope. We also found no evidence of an interaction between mastery and the brain health slope in predicting cognition slope. However, an interaction between constraints and the brain health slope in predicting cognition slope indicated a stronger association between brain health slope and cognition slope among those with higher compared to lower constraints, suggesting that higher constraints are associated with less cognitive reserve over time.

Overall, both “residual” and “moderation” approaches showed evidence for a negative relationship between constraints (but not mastery) and longitudinal cognitive reserve. Only the “residual” approach showed evidence for an additional cross-sectional relationship between greater constraints and lower initial cognitive reserve.

Racial and Ethnic Differences in Associations

Perceived Control and Longitudinal Brain Health

Multiple-group models tested for differences in associations between perceived control and brain health intercepts or slopes across groups (Supplementary Table 5). There were no group differences in how mastery related to trajectories of total gray matter volume, total hippocampal volume, or WMH. Specifically, mastery was not associated with the intercept or slope of these brain outcomes for any group. Similarly, there were no group differences in associations between constraints and trajectories of cortical thickness, total gray matter volume, or WMH. Constraints were not related to either the intercept or the slope of these brain outcomes for any group.

In contrast, the relationship between mastery and initial cortical thickness in AD signature regions differed between White and Hispanic participants. Specifically, greater mastery related to greater initial cortical thickness for Hispanic (β=0.142, 95% CI:[0.040, 0.244]), but not White (β=−0.048, 95% CI:[−0.196, 0.100]), participants. There were no differences in this relationship for the remaining group comparisons. Like the other brain health slope effects, mastery was not related to cortical thickness slope, and there were no differences in this association between groups.

There was also a difference in how constraints related to initial hippocampal volume when comparing Black participants to White and Hispanic participants. Specifically, more constraints related to greater initial hippocampal volume for Black participants (vs. White: β=0.118, 95% CI:[0.004, 0.232]; vs. Hispanic: β=0.116, 95% CI:[0.005, 0.227]), but not White (β=−0.111, 95% CI:[−0.241, 0.019]) or Hispanic (β=−0.079, 95% CI:[−0.177, 0.019]) participants. To explore whether this finding may reflect selective survival, we reran the analysis with Black adults stratified by the average life expectancy of Black adults in 2022 (73 years; Ndugga et al., 2024). Although neither association was significant in these relatively small, stratified samples, the association was 22% larger for the oldest (n=130; β=0.128, 95% CI:[−0.037, 0.294]), compared with the youngest (n=114; β=0.105, 95% CI:[−0.072, 0.283]), Black older adults, suggesting that this effect may be driven by selective survival.

Finally, there was a difference in how constraints related to hippocampal volume slope when comparing White participants to Black and Hispanic participants. Specifically, more constraints related to faster hippocampal volume decline for White older adults when compared to both Black (White estimate: β=−0.338, 95% CI:[−0.558, −0.118]) and Hispanic (White estimate: β=−0.256, 95% CI:[−0.476, −0.036]) older adults. However, constraints were not related to hippocampal volume decline for either Black (β=0.082, 95% CI:[−0.158, 0.322]) or Hispanic (β=0.098, 95% CI:[−0.176, 0.372]) older adults.

Perceived Control and Longitudinal Cognitive Reserve

Multiple-group models also tested for differences in associations between perceived control and cognitive reserve across groups (Supplementary Table 5). In models using the “residual” approach, associations between mastery and cognitive reserve trajectories were similarly null across racial and ethnic groups. Consistent with the model for the full sample, constraints were similarly associated with lower initial cognitive reserve and faster decline in cognitive reserve across all groups.

In models using the “moderation” approach, associations were similar across groups with two exceptions. First, model fit improved when freeing the interaction between mastery and the brain health slope in predicting cognition slope for White versus Black participants. While this interaction was more than three times larger among Black (β=0.089, 95% CI:[−0.043, 0.220]) than White (β=0.028, 95% CI:[−0.336, 0.391]) participants, it was not statistically significant in either group.

Second, constraints moderated the association between brain health slope and cognition slope among Black (β=0.416, 95% CI:[0.164, 0.667]) and Hispanic (β=0.377, 95% CI:[0.093, 0.661]) participants, but not White participants (vs. Black: β=0.099, 95% CI:[−0.184, 0.382]; vs. Hispanic: β=0.097, 95% CI:[−0.264, 0.459]). Both Black (Figure 3 Panel A) and Hispanic (Figure 3 Panel B) participants with high constraints exhibited a stronger association between brain health decline and cognitive decline compared to those with low constraints.

Figure 3. Associations Between Brain Health Slope and Cognition Slope Stratified by Race and Ethnicity and Constraints.

Figure 3.

Open in a new tab

Note. To visualize the effect of constraints, participants were grouped into high and low constraints groups based on a median split. Panel A shows the comparison between White and Black participants based on model-estimated person-specific slopes from the multiple-group model including these two groups. Panel B shows the comparison between White and Hispanic participants based on model-estimated person-specific slopes from the multiple-group model including these two groups. White participants had similar associations between brain health slope and cognition slope regardless of constraints. Black and Hispanic participants with high constraints had a stronger association between brain health slope and cognition slope than those with low constraints.

Discussion

We characterized resilience mechanisms underlying longitudinal associations between perceived control and later-life cognitive health across race and ethnicity. In the full sample, we found evidence that more constraints may lead to worse cognitive trajectories not through brain health mechanisms (i.e., brain maintenance) but rather through cognitive reserve mechanisms. Associations appeared strongest among older adults from minoritized groups (i.e., Hispanic, Black). Together, these results suggest that reducing environmental constraints (e.g., structural racism) that limit individuals’ control over life outcomes may enhance cognitive reserve and reduce racial and ethnic cognitive inequalities.

Perceived Control and Brain Health Across Race and Ethnicity

We did not find support for our first hypothesis that perceived control would relate to brain health trajectories within the full sample. Specifically, neither mastery nor constraints were associated with the intercept or slope of the four brain health outcomes. We found both partial support and opposite findings for our third hypothesis, examining differences by race and ethnicity. The lack of findings among the full sample, in conjunction with certain findings in the multiple-group models, suggests that considering associations between perceived control and brain health in heterogenous samples may mask group-specific effects and potential differential mechanisms between groups. Additionally, the specific brain measures that showed associations with perceived control may help guide future research. The only association found for mastery was in relation to initial cortical thickness in AD signature regions among Hispanic older adults. Greater constraints were selectively related to hippocampal volume among White and Black older adults, though in conflicting directions.

We extended previous research by addressing limitations. Specifically, we included (1) mastery and constraints as separate indicators to pinpoint which aspects of perceived control most strongly related to brain health, (2) a larger sample of older adults, (3) longitudinal outcomes, and (4) a racially and ethnically diverse sample, allowing us to highlight potential racial and ethnic differences in links between perceived control and brain health. Overall, we found limited evidence of an association between perceived control and brain health. Notably, the specific components of perceived control seem to relate to brain health differently across racial and ethnic groups. Our findings for Hispanic older adults are consistent with previous cross-sectional studies (Hashimoto et al., 2015; Pruessner et al., 2005), which may all be subject to reverse causality.

Since neither previous study explicitly examined constraints, it is unclear whether the effect of constraints on hippocampal volume loss among White older adults is unique to our sample. However, this effect aligns with a stress mechanism. Prior research suggests that constraints relate to perceived stress, more so for White than Black adults (Wen & Sin, 2022). Additionally, there is strong support for the negative association between chronic stress exposure and lower hippocampal volume (Kim et al., 2015). If Black and Hispanic older adults perceive constraints as less stressful than White adults (which has been demonstrated for chronic stressor exposure; Brown et al., 2020), the neurophysiological impact of constraints may be lessened among these groups and may be why White older adults selectively showed an association between constraints and hippocampal volume decline. Future work should incorporate indicators of perceived stress and consider additional sources of resilience when testing associations between constraints and brain health. Of note, the unexpected finding of greater mastery among Hispanic and Black older adults is consistent with the mental health paradox (Mezuk et al., 2013) and may reflect group-based differences in coping and/or other resources (e.g., social support, ethnic identity) (Glymour & Manly, 2008).

Perceived Control and Cognitive Reserve Across Race and Ethnicity

Using complementary approaches to demonstrating cognitive reserve (i.e., “residual” and “moderation”), we found support for our second and third hypotheses when considering constraints, but not mastery, in relation to cognitive reserve trajectories. Using the “residual” approach, greater constraints related to worse initial cognition and faster cognitive decline when controlling for brain health trajectories. Using the “moderation” approach, we found that greater constraints related to a stronger association between changes in brain health and changes in cognition. Guidelines on operationalizations of cognitive reserve suggest that the moderation approach is a stronger test of cognitive reserve compared to models that control for brain health only (as done in the residual approach) (Stern et al., 2020, 2023). Further, the residual approach is subject to alternative interpretations other than cognitive reserve (Stern et al., 2020), such as perceived control influencing cognition through mechanisms other than the brain health indicators included in the composite. Thus, the more limited findings that emerged among the moderation approach (i.e., for longitudinal associations and among Hispanic and Black older adults) may point to these associations as most strongly indicative of cognitive reserve.

There has been little research on perceived control and cognitive reserve. Previous work identified that other psychosocial factors (e.g., personality traits and eudemonic well-being) were positively related to cognitive reserve (i.e., better than expected cognition given pathology; Graham et al., 2021; Willroth et al., 2023). However, neither previous study on psychosocial factors and cognitive reserve assessed differences in how psychosocial factors relate to cognitive reserve by race or ethnicity. Our study extends a previous cross-sectional study using an alternative approach to investigating cognitive reserve that found greater overall perceived control more strongly related to intact memory performance despite hippocampal volume loss among minoritized older adults compared to White older adults (Zahodne et al., 2018). Our longitudinal results indicate that greater constraints more strongly related to longitudinal cognitive reserve among Hispanic and Black older adults compared to White older adults.

It could be that White older adults with greater constraints have access to other resources that buffer the effect of structural atrophy on cognitive decline, nullifying the moderating role of constraints on longitudinal cognitive reserve. Contrarily, minoritized older adults with greater constraints may lack access to these other buffering resources, due to the pervasive and multilevel nature of structural and social determinants of health (e.g., exposure to pollutants, fewer socioeconomic opportunities, and limited healthy food/space for physical activity/health care access impacting health and health behaviors) (Adkins-Jackson et al., 2023), contributing to the stronger relationship between brain atrophy and cognitive decline among Black and Hispanic older adults. Thus, reducing environmental constraints (e.g., structural/social determinants of health, including structural racism), has high potential for maintaining cognitive health, regardless of brain health, for minoritized racial and ethnic individuals who have higher dementia risk. Notably, the lack of findings between mastery and cognitive reserve points to the importance of considering these perceived control dimensions separately, as these constructs have differential associations with cognitive health. Additionally, the preponderance of findings for constraints compared with mastery highlights the potential importance of macroenvironmental (as opposed to microenvironmental) interventions for cognitive disparities.

Strengths, Limitations, and Future Directions

Our study has several strengths. First, we included a well-characterized sample of racially and ethnically diverse older adults, which is particularly relevant given well-established racial and ethnic disparities in cognitive aging. Second, we included at least three measures of both brain health and cognition and used multiple indicators for these factors, allowing us to robustly estimate longitudinal changes. Third, we operationalized cognitive reserve across two approaches that have been previously recommended (Stern et al., 2020, 2023) but rarely examined in tandem to better identify mechanisms linking perceived control to cognition. Finally, we considered mastery and constraints independently, allowing for the detection of their unique contributions, which can guide intervention efforts.

We also recognize some study limitations. First, WHICAP is a regional sample based in northern Manhattan, which may limit generalizability. Second, our sample may be selective because of their participation in the MRI sub-study and their advanced age. The WHICAP MRI sub-study was found to be younger, more likely to be men, and more likely to be Black compared to individuals who were not scanned (Brickman et al., 2008). Thus, those included in the current study may not be fully representative of the entire WHICAP cohort, as is typically the case in MRI studies. However, we sought to minimize additional selection by requiring participants only to have a baseline assessment to be included, with full information maximum likelihood estimation accounting for missing data in subsequent waves. Regarding selection due to age, participants’ average age was close to older adults’ life expectancies from these racial and ethnic groups in 2022 (Ndugga et al., 2024). The age of the sample is particularly important when considering racial and ethnic disparities, as selective survival can influence the magnitude of disparities and causal inferences (Zahodne et al., 2016). Thus, all participants may have been particularly resilient, especially Black participants whose average age was above the average life expectancy for this group. Sensitivity analyses supported the possibility that the unexpected association between more constraints and greater initial hippocampal volume was attributable to selective survival. Third, although we used longitudinal data, the data were observational, so we are unable to confirm causality.

Future work should attempt replication with other regional and nationally representative samples that include adults from earlier in the life course to increase generalizability and limit the effects of selective survival. Although some research suggests perceived control is relatively stable over time (Windsor & Anstey, 2008), others suggest that it changes (Anderson et al., 2018; Du et al., 2024) and that changes towards more internal control relate to better cognition. Thus, future research could assess changes in perceived control in relation to changes in brain health and cognitive reserve. Indicators of pathophysiology could be assessed. Additionally, we assessed objective, performance-based measures of cognition, but future work could include functional measures or dementia as outcomes to further understand the clinical implications of our findings. Finally, although we included a racially and ethnically diverse sample of older adults, associations among adults from other races and ethnicities (e.g., Asian, Middle Eastern/North African) should also be considered, as well as within-group heterogeneity (e.g., Hispanic adults from Mexico, Cuba, and Puerto Rico).

Conclusion

We assessed whether different aspects of perceived control were longitudinally related to brain health and/or cognitive reserve among community-dwelling Hispanic, Black, and White older adults. We found more evidence for cognitive reserve mechanisms than brain health mechanisms in explaining the link between perceived control and later-life cognitive health, that constraints, but not mastery, contributed to cognitive reserve, and links between constraints and cognitive reserve were strongest among older adults from minoritized racial and ethnic groups. Thus, targeting environmental constraints may be an effective way to maintain older adults’ cognitive health and reduce racial and ethnic disparities in dementia risk through cognitive resilience mechanisms.

Supplementary Material

Supplementary Material

Acknowledgements

Data collection and sharing for this project was supported by the Washington Heights-Inwood Columbia Aging Project (WHICAP, PO1AG07232, R01AG037212, RF1AG054023) funded by the National Institute on Aging. This manuscript has been reviewed by WHICAP investigators for scientific content and consistency of data interpretation with previous WHICAP Study publications. We acknowledge the WHICAP study participants and the WHICAP research and support staff for their contributions to this study.

Funding

This publication was supported by the National Center for Advancing Translational Sciences, National Institutes of Health (UL1TR001873). Additionally, this work was supported by National Institute on Aging (R01AG054520 to A. M. Brickman and L. B. Zahodne, K01AG073588 to K. Sol, and F31 AG082506 to J. D. Palms). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Footnotes

Conflicts of Interest

None.

Data Availability

Qualified researchers may request access to de-identified data through the WHICAP data request form. This study was not preregistered.

References

  1. Adkins-Jackson PB, George KM, Besser LM, Hyun J, Lamar M, Hill-Jarrett TG, Bubu OM, Flatt JD, Heyn PC, Cicero EC, Zarina Kraal A, Pushpalata Zanwar P, Peterson R, Kim B, Turner RW, Viswanathan J, Kulick ER, Zuelsdorff M, Stites SD, … Babulal G (2023). The structural and social determinants of Alzheimer’s disease related dementias. Alzheimer’s & Dementia, 19(7), 3171–3185. 10.1002/alz.13027 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Anderson E, Cochrane A, Golding J, & Nowicki S (2018). Locus of control as a modifiable risk factor for cognitive function in midlife. Aging, 10(7), 1542–1555. 10.18632/aging.101490 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Avila JF, Rentería MA, Jones RN, Vonk JMJ, Turney I, Sol K, Seblova D, Arias F, Hill-Jarrett T, Levy S, Meyer O, Racine AM, Tom SE, Melrose RJ, Deters K, Medina LD, Carrión CI, Díaz-Santos M, Byrd DR, … Manly JJ (2021). Education differentially contributes to cognitive reserve across racial/ethnic groups. Alzheimer’s & Dementia, 17(1), 70–80. 10.1002/alz.12176 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bailey ZD, Krieger N, Agénor M, Graves J, Linos N, & Bassett MT (2017). Structural racism and health inequities in the USA: Evidence and interventions. The Lancet, 389(10077), 1453–1463. 10.1016/S0140-6736(17)30569-X [DOI] [PubMed] [Google Scholar]
  5. Brickman AM, Muraskin J, & Zimmerman ME (2009). Structural neuroimaging in Alzheimer’s disease: Do white matter hyperintensities matter? Dialogues in Clinical Neuroscience, 11(2), 181–190. 10.31887/dcns.2009.11.2/ambrickman [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brickman AM, Provenzano FA, Muraskin J, Manly JJ, Blum S, Apa Z, Stern Y, Brown TR, Luchsinger JA, & Mayeux R (2012). Regional white matter hyperintensity volume, not hippocampal atrophy, predicts incident Alzheimer disease in the community. Archives of Neurology, 69(12), 1621–1627. 10.1001/archneurol.2012.1527 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Brickman AM, Schupf N, Manly JJ, Luchsinger JA, Andrews H, Tang MX, Reitz C, Small SA, Mayeux R, DeCarli C, & Brown TR (2008). Brain Morphology in Older African Americans, Caribbean Hispanics, and Whites From Northern Manhattan. Archives of Neurology, 65(8), 1053–1061. 10.1001/archneur.65.8.1053 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brickman AM, Sneed JR, Provenzano FA, Garcon E, Johnert L, Muraskin J, Yeung L-K, Zimmerman ME, & Roose SP (2011). Quantitative approaches for assessment of white matter hyperintensities in elderly populations. Psychiatry Research: Neuroimaging, 193(2), 101–106. 10.1016/j.pscychresns.2011.03.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Brown LL, Mitchell UA, & Ailshire JA (2020). Disentangling the Stress Process: Race/Ethnic Differences in the Exposure and Appraisal of Chronic Stressors Among Older Adults. The Journals of Gerontology: Series B, 75(3), 650–660. 10.1093/geronb/gby072 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dickerson BC, Bakkour A, Salat DH, Feczko E, Pacheco J, Greve DN, Grodstein F, Wright CI, Blacker D, Rosas HD, Sperling RA, Atri A, Growdon JH, Hyman BT, Morris JC, Fischl B, & Buckner RL (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. Cerebral Cortex, 19(3), 497–510. 10.1093/cercor/bhn113 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Du C, Wu B, Peng C, Dong X, Li M, Pernice FM, & Wang Y (2024). Disaggregating between- and within-person associations of mastery and cognitive function: Age as a moderator. BMC Geriatrics, 24(1), 722. 10.1186/s12877-024-05256-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Glymour MM, & Manly JJ (2008). Lifecourse social conditions and racial and ethnic patterns of cognitive aging. Neuropsychology Review, 18(3 SPEC. ISS.), 223–254. 10.1007/s11065-008-9064-z [DOI] [PubMed] [Google Scholar]
  13. Graham EK, James BD, Jackson KL, Willroth EC, Boyle P, Wilson R, Bennett DA, & Mroczek DK (2021). Associations Between Personality Traits and Cognitive Resilience in Older Adults. The Journals of Gerontology: Series B, 76(1), 6–19. 10.1093/geronb/gbaa135 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hashimoto T, Takeuchi H, Taki Y, Sekiguchi A, Nouchi R, Kotozaki Y, Nakagawa S, Miyauchi CM, Iizuka K, Yokoyama R, Shinada T, Yamamoto Y, Hanawa S, Araki T, Hashizume H, Kunitoki K, & Kawashima R (2015). Neuroanatomical correlates of the sense of control: Gray and white matter volumes associated with an internal locus of control. NeuroImage, 119, 146–151. 10.1016/j.neuroimage.2015.06.061 [DOI] [PubMed] [Google Scholar]
  15. Jack CR, Petersen RC, Xu Y, O’Brien PC, Smith GE, Ivnik RJ, Tangalos EG, & Kokmen E (1998). Rate of medial temporal lobe atrophy in typical aging and Alzheimer’s disease. Neurology, 51(4), 993–999. 10.1212/WNL.51.4.993 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kessler RC (1979). Stress, Social Status, and Psychological Distress. Journal of Health and Social Behavior, 20(3), 259. 10.2307/2136450 [DOI] [PubMed] [Google Scholar]
  17. Kim EJ, Pellman B, & Kim JJ (2015). Stress effects on the hippocampus: A critical review. Learning & Memory, 22(9), 411–416. 10.1101/lm.037291.114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Lachman ME, Neupert SD, & Agrigoroaei S (2011). The Relevance of Control Beliefs for Health and Aging. In Handbook of the Psychology of Aging (pp. 175–190). Elsevier. 10.1016/B978-0-12-380882-0.00011-5 [DOI] [Google Scholar]
  19. Lachman ME, & Weaver SL (1998). The sense of control as a moderator of social class differences in health and well-being. Journal of Personality and Social Psychology, 74(3), 763–773. 10.1037/0022-3514.74.3.763 [DOI] [PubMed] [Google Scholar]
  20. Manly JJ, Bell-McGinty S, Tang M-X, Schupf N, Stern Y, & Mayeux R (2005). Implementing diagnostic criteria and estimating frequency of mild cognitive impairment in an urban community. Archives of Neurology, 62(11), 1739–1746. 10.1001/archneur.62.11.1739 [DOI] [PubMed] [Google Scholar]
  21. Mezuk B, Abdou CM, Hudson D, Kershaw KN, Rafferty JA, Lee H, & Jackson JS (2013). “White Box” Epidemiology and the Social Neuroscience of Health Behaviors: The Environmental Affordances Model. Society and Mental Health, 3(2), 79–95. 10.1177/2156869313480892 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Muthén LK, & Muthén BO (2017). Mplus User’s Guide (Eighth). Muthén & Muthén. [Google Scholar]
  23. Ndugga N, Hill L, & Artiga S (2024). Key Data on Health and Health Care by Race and Ethnicity (Racial Equity and Health Policy). KFF. https://www.kff.org/key-data-onhealth-and-health-care-by-race-and-ethnicity/ [Google Scholar]
  24. Nkwata AK, Zhang M, Song X, Giordani B, & Ezeamama AE (2021). The Relationship of Race, Psychosocial Stress and Resiliency Indicators to Neurocognitive Impairment among Older Americans Enrolled in the Health and Retirement Survey: A Cross-Sectional Study. International Journal of Environmental Research and Public Health, 18(3), 1358. 10.3390/ijerph18031358 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Nkwata AK, Zhang M, Song X, Giordani B, & Ezeamama AE (2022). Toxic Psychosocial Stress, Resiliency Resources and Time to Dementia Diagnosis in a Nationally Representative Sample of Older Americans in the Health and Retirement Study from 2006–2016. International Journal of Environmental Research and Public Health, 19(4), 2419. 10.3390/ijerph19042419 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Nyberg L, Lövdén M, Riklund K, Lindenberger U, & Bäckman L (2012). Memory aging and brain maintenance. Trends in Cognitive Sciences, 16(5), 292–305. 10.1016/j.tics.2012.04.005 [DOI] [PubMed] [Google Scholar]
  27. Pruessner JC, Baldwin MW, Dedovic K, Renwick R, Mahani NK, Lord C, Meaney M, & Lupien S (2005). Self-esteem, locus of control, hippocampal volume, and cortisol regulation in young and old adulthood. NeuroImage, 28(4), 815–826. 10.1016/j.neuroimage.2005.06.014 [DOI] [PubMed] [Google Scholar]
  28. Ross CE, & Mirowsky J (2013). The Sense of Personal Control: Social Structural Causes and Emotional Consequences. In Aneshensel CS, Phelan JC, & Bierman A (Eds.), Handbook of the Sociology of Mental Health (pp. 379–402). Springer; Netherlands. 10.1007/978-94-007-4276-5_19 [DOI] [Google Scholar]
  29. Siedlecki KL, Manly JJ, Brickman AM, Schupf N, Tang M-X, & Stern Y (2010). Do neuropsychological tests have the same meaning in Spanish speakers as they do in English speakers? Neuropsychology, 24(3), 402–411. 10.1037/a0017515 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Stern Y (2002). What is cognitive reserve? Theory and research application of the reserve concept. Journal of the International Neuropsychological Society: JINS, 8(3), 448–460. [PubMed] [Google Scholar]
  31. Stern Y, Albert M, Barnes CA, Cabeza R, Pascual-Leone A, & Rapp PR (2023). A framework for concepts of reserve and resilience in aging. Neurobiology of Aging, 124, 100–103. 10.1016/j.neurobiolaging.2022.10.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Stern Y, Arenaza-Urquijo EM, Bartrés-Faz D, Belleville S, Cantilon M, Chetelat G, Ewers M, Franzmeier N, Kempermann G, Kremen WS, Okonkwo O, Scarmeas N, Soldan A, Udeh-Momoh C, Valenzuela M, Vemuri P, Vuoksimaa E, & and the Reserve, Resilience and Protective Factors PIA Empirical Definitions and Conceptual Frameworks Workgroup. (2020). Whitepaper: Defining and investigating cognitive reserve, brain reserve, and brain maintenance. Alzheimer’s & Dementia, 16(9), 1305–1311. 10.1016/j.jalz.2018.07.219 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Sutin AR, Stephan Y, & Terracciano A (2018). Psychological Distress, Self-Beliefs, and Risk of Cognitive Impairment and Dementia. Journal of Alzheimer’s Disease, 65(3), 1041–1050. 10.3233/JAD-180119 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Tang MX, Cross P, Andrews H, Jacobs DM, Small S, Bell K, Merchant C, Lantigua R, Costa R, Stern Y, & Mayeux R (2001). Incidence of AD in African-Americans, Caribbean Hispanics, and Caucasians in northern Manhattan. Neurology, 56(1), 49–56. 10.1212/wnl.56.1.49 [DOI] [PubMed] [Google Scholar]
  35. Turney IC, Lao PJ, Rentería MA, Igwe KC, Berroa J, Rivera A, Benavides A, Morales CD, Rizvi B, Schupf N, Mayeux R, Manly JJ, & Brickman AM (2023). Brain Aging Among Racially and Ethnically Diverse Middle-Aged and Older Adults. JAMA Neurology, 80(1), 73. 10.1001/jamaneurol.2022.3919 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Voevodskaya O, Simmons A, Nordenskjöld R, Kullberg J, Ahlström H, Lind L, Wahlund L-O, Larsson E-M, Westman E, & Alzheimer’s Disease Neuroimaging Initiative. (2014). The effects of intracranial volume adjustment approaches on multiple regional MRI volumes in healthy aging and Alzheimer’s disease. Frontiers in Aging Neuroscience, 6. 10.3389/fnagi.2014.00264 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Ward JB, Gartner DR, Keyes KM, Fliss MD, McClure ES, & Robinson WR (2019). How do we assess a racial disparity in health? Distribution, interaction, and interpretation in epidemiological studies. Annals of Epidemiology, 29, 1–7. 10.1016/j.annepidem.2018.09.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Wen JH, & Sin NL (2022). Perceived control and reactivity to acute stressors: Variations by age, race and facets of control. Stress and Health, 38(3), 419–434. 10.1002/smi.3103 [DOI] [PubMed] [Google Scholar]
  39. Wight RG, Aneshensel CS, Seeman M, & Seeman TE (2003). Late life cognition among men: A life course perspective on psychosocial experience. Archives of Gerontology and Geriatrics, 37(2), 173–193. 10.1016/S0167-4943(03)00046-3 [DOI] [PubMed] [Google Scholar]
  40. Willroth EC, James BD, Graham EK, Kapasi A, Bennett DA, & Mroczek DK (2023). Well-Being and Cognitive Resilience to Dementia-Related Neuropathology. Psychological Science, 34(3), 283–297. 10.1177/09567976221119828 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Windsor TD, & Anstey KJ (2008). A Longitudinal Investigation of Perceived Control and Cognitive Performance in Young, Midlife and Older Adults. Aging, Neuropsychology, and Cognition, 15(6), 744–763. 10.1080/13825580802348570 [DOI] [PubMed] [Google Scholar]
  42. Zahodne LB (2021). Biopsychosocial Pathways in Dementia Inequalities: Introduction to the Michigan Cognitive Aging Project. American Psychologist, 76(9), 1470–1481. 10.1037/amp0000936 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Zahodne LB, Manly JJ, Azar M, Brickman AM, & Glymour MM (2016). Racial Disparities in Cognitive Performance in Mid- and Late Adulthood: Analyses of Two Cohort Studies. Journal of the American Geriatrics Society, 64(5), 959–964. 10.1111/jgs.14113 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zahodne LB, Manly JJ, Brickman AM, Narkhede A, Griffith EY, Guzman VA, Schupf N, & Stern Y (2015). Is residual memory variance a valid method for quantifying cognitive reserve? A longitudinal application. Neuropsychologia, 77, 260–266. 10.1016/j.neuropsychologia.2015.09.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Zahodne LB, Manly JJ, Smith J, Seeman T, & Lachman ME (2017). Socioeconomic, health, and psychosocial mediators of racial disparities in cognition in early, middle, and late adulthood. Psychology and Aging, 32(2), 118–130. 10.1037/pag0000154 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Zahodne LB, Meyer OL, Choi E, Thomas ML, Willis SL, Marsiske M, Gross AL, Rebok GW, & Parisi JM (2015). External locus of control contributes to racial disparities in memory and reasoning training gains in ACTIVE. Psychology and Aging, 30(3), 561–572. 10.1037/pag0000042 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Zahodne LB, Schupf N, & Brickman AM (2018). Control beliefs are associated with preserved memory function in the face of low hippocampal volume among diverse older adults. Brain Imaging and Behavior, 12(4), 1112–1120. 10.1007/s11682-017-9776-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Zahodne LB, Sharifian N, Kraal AZ, Zaheed AB, Sol K, Morris EP, Schupf N, Manly JJ, & Brickman AM (2021). Socioeconomic and psychosocial mechanisms underlying racial/ethnic disparities in cognition among older adults. Neuropsychology, 35(3), 265–275. 10.1037/neu0000720 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Zahodne LB, Sol K, & Kraal Z (2019). Psychosocial Pathways to Racial/Ethnic Inequalities in Late-Life Memory Trajectories. The Journals of Gerontology: Series B, 74(3), 409–418. 10.1093/geronb/gbx113 [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 Material
NIHMS2146645-supplement-Supplementary_Material.docx (27.1KB, docx)

Data Availability Statement

Qualified researchers may request access to de-identified data through the WHICAP data request form. This study was not preregistered.

RESOURCES