Could differential under-reporting of loneliness between men and women bias the gender-specific association between loneliness duration and rate of memory decline? A probabilistic bias analysis of effect modification.

Gender is an observed effect modifier of the association between loneliness and memory aging. However, this effect modification may be a result of information bias due to differential loneliness under-reporting by gender. We applied probabilistic bias analyses to examine whether effect modification of the loneliness-memory decline relationship by gender is retained under three simulation scenarios with various magnitudes of differential loneliness under-reporting between men and women. Data were from biennial interviews with adults aged 50+ in the US Health and Retirement Study from 1996-2016 (5,646 women and 3,386 men). Loneliness status (yes vs. no) was measured from 1996-2004 using the CES-D loneliness item and memory was measured from 2004-2016. Simulated sensitivity and specificity of the loneliness measure were informed by a validation study using the UCLA Loneliness Scale as a gold standard. The likelihood of observing effect modification by gender was higher than 90% in all simulations, although the likelihood reduced with an increasing difference in magnitude of the loneliness under-reporting between men and women. The gender difference in loneliness under-reporting did not meaningfully affect the observed effect modification by gender in our simulations. Our simulation approach may be promising to quantify potential information bias in effect modification analyses.

Time-lagged associations between two adverse childhood experiences and later-life cognitive function through educational attainment and stroke.

OBJECTIVE: Adverse childhood experiences (ACEs) have been associated with worse cognitive health in older adulthood. This study aimed to extend findings on the specificity, persistence, and pathways of associations between two ACEs and cognition by using a comprehensive neuropsychological battery and a time-lagged mediation design. METHOD: Participants were 3304 older adults in the Health and Retirement Study Harmonized Cognitive Assessment Protocol. Participants retrospectively reported whether they were exposed to parental substance abuse or experienced parental physical abuse before age 18. Factor scores derived from a battery of 13 neuropsychological tests indexed cognitive domains of episodic memory, executive functioning, processing speed, language, and visuospatial function. Structural equation models examined self-reported years of education and stroke as mediators, controlling for sociodemographics and childhood socioeconomic status. RESULTS: Parental substance abuse in childhood was associated with worse later-life cognitive function across all domains, in part via pathways involving educational attainment and stroke. Parental physical abuse was associated with worse cognitive outcomes via stroke independent of education. CONCLUSIONS: This national longitudinal study in the United States provides evidence for broad and persistent indirect associations between two ACEs and cognitive aging via differential pathways involving educational attainment and stroke. Future research should examine additional ACEs and mechanisms as well as moderators of these associations to better understand points of intervention.

Mechanisms underlying the association between adverse childhood experiences and racial disparities in later-life cognition.

OBJECTIVE: Adverse childhood experiences (ACEs) may be a risk factor for later-life cognitive disorders such as dementia; however, few studies have investigated underlying mechanisms, such as cardiovascular health and depressive symptoms, in a health disparities framework. METHOD: 418 community-dwelling adults (50% nonHispanic Black, 50% nonHispanic White) aged 55+ from the Michigan Cognitive Aging Project retrospectively reported on nine ACEs. Baseline global cognition was a z-score composite of five factor scores from a comprehensive neuropsychological battery. Depressive symptoms were assessed using the Center for Epidemiologic Studies Depression Scale. Cardiovascular health was operationalized through systolic blood pressure. A mediation model controlling for sociodemographics, childhood health, and childhood socioeconomic status estimated indirect effects of ACEs on global cognition via depressive symptoms and blood pressure. Racial differences were probed via t-tests and stratified models. RESULTS: A negative indirect effect of ACEs on cognition was observed through depressive symptoms [ß = -.040, 95% CI (-.067, -.017)], but not blood pressure, for the whole sample. Black participants reported more ACEs (Cohen's d = .21), reported more depressive symptoms (Cohen's d = .35), higher blood pressure (Cohen's d = .41), and lower cognitive scores (Cohen's d = 1.35) compared to White participants. In stratified models, there was a negative indirect effect through depressive symptoms for Black participants [ß = -.074, 95% CI (-.128, -.029)] but not for White participants. CONCLUSIONS: These results highlight the need to consider racially patterned contextual factors across the life course. Such factors could exacerbate the negative impact of ACEs and related mental health consequences and contribute to racial disparities in cognitive aging.

Specific aspects of religious involvement protect against depressive symptoms among immigrant versus U.S.-born, Hispanic older adults.

OBJECTIVES: This study investigates religious involvement and depressive symptoms in Hispanic older adults in the United States. We hypothesized that private prayer, religious attendance, and religious belief would have an inverse association with depressive symptoms, and that these associations would be stronger among immigrants, compared to U.S.-born participants. METHOD: This cross-sectional, within-group study included 1,566 participants from the Health and Retirement Study. Multivariate linear regression evaluated the association between religious involvement and depressive symptoms in the whole sample and in subgroups stratified by immigrant status. RESULTS: Overall, only more frequent religious attendance was associated with fewer depressive symptoms. Stratified models revealed an additional inverse association between private prayer and depressive symptoms only in the immigrant group. CONCLUSION: These findings may help incorporate religious preferences into mental health prevention and treatment to reduce depressive symptoms among older Hispanic adults.

Does education moderate gender disparities in later-life memory function? A cross-national comparison of harmonized cognitive assessment protocols in the United States and India.

INTRODUCTION: We compared gender disparities in later-life memory, overall and by education, in India and the United States (US). METHODS: Data (N = 7443) were from harmonized cognitive assessment protocols (HCAPs) in the Longitudinal Aging Study of India-Diagnostic Assessment of Dementia (LASI-DAD; N = 4096; 2017-19) and US Health and Retirement Study HCAP (HRS-HCAP; N = 3347; 2016-17). We derived harmonized memory factors from each study using confirmatory factor analysis. We used multivariable-adjusted linear regression to compare gender disparities in memory function between countries, overall and by education. RESULTS: In the United States, older women had better memory than older men (0.28 SD-unit difference; 95% CI: 0.22, 0.35). In India, older women had worse memory than older men (-0.15 SD-unit difference; 95% CI: -0.20, -0.10), which attenuated with increasing education and literacy. CONCLUSION: We observed gender disparities in memory in India that were not present in the United States, and which dissipated with education and literacy.

Histories of neighborhood socioeconomic status contribute to race differences in later-life cognition.

INTRODUCTION: Neighborhood characteristics are increasingly implicated in cognitive health disparities, but no research has investigated how the historical context of neighborhoods shapes these disparities. METHODS: Four hundred sixty-four Black (55%) and White older adults (Mage = 63.6) were drawn from the Michigan Cognitive Aging Project, a community-based, prospective study of older adults. Participants' addresses at baseline (2017-2020) were geocoded and linked to 2000-2017 measures of neighborhood socioeconomic status (NSES): disadvantage [NDis] and affluence [NAff]. Latent class growth analysis (LCGA) characterized 18 interpolated year trajectories of NSES across 1344 census tracts. Path analysis examined whether NSES trajectory classes mediated the association between race and a global cognition composite. RESULTS: LCGA identified three NDis and two NAff trajectory classes, which were associated with participant race. Only one NDis class was associated with cognition, and it mediated the association between the Black race and cognition. DISCUSSION: Disinvestment in neighborhoods may be particularly salient in race disparities in cognitive function. HIGHLIGHTS: Race is implicated in the likelihood of living in more disadvantaged neighborhoods. Historical trends in neighborhood disadvantage are associated with cognitive function in older adulthood. Identifying patterns of neighborhood change may inform neighborhood-level interventions.

Neighborhood racial income inequality and cognitive health.

INTRODUCTION: Neighborhood socioeconomic status (SES) has been linked to dementia, but the distribution of SES within a neighborhood may also matter. METHODS: Data from 460 (47% Black, 46% White) older adults from the Michigan Cognitive Aging Project were linked to census tract-level data from the National Neighborhood Data Archive (NaNDA). Neighborhood SES included two composites reflecting disadvantage and affluence. Neighborhood racial income inequality was the ratio of median incomes for White versus Black residents. Generalized estimating equations examined associations between neighborhood factors and cognitive domains. RESULTS: Neighborhood racial income inequality was uniquely associated with worse cognitive health, and these associations did not differ by participant race. Neighborhood disadvantage was only associated with worse cognitive health among Black participants. DISCUSSION: Both the level and racial distribution of SES within a neighborhood may be relevant for dementia risk. Racial differences in the level and impact of neighborhood SES contribute to dementia inequalities. HIGHLIGHTS: Black participants lived in neighborhoods with lower socioeconomic status (SES) than White participants, on average. Neighborhood SES and racial income inequality were associated with worse cognition. Effects of neighborhood racial income inequality did not differ across racial groups. Effects of neighborhood SES were only evident among Black participants.

Links between early-life contextual factors and later-life cognition and the role of educational attainment.

OBJECTIVE: Educational attainment is a well-documented predictor of later-life cognition, but less is known about upstream contextual factors. This study aimed to identify which early-life contextual factors uniquely predict later-life global cognition and whether educational attainment mediates these relationships. METHOD: Participants were drawn from the Michigan Cognitive Aging Project (N = 485; Mage = 63.51; SDage = 3.13; 50% non-Hispanic Black). Early-life exposures included U.S. region of elementary school (Midwest, South, Northeast), average parental education, household composition (number of adults (1, 2, 3+), number of children), school racial demographics (predominantly White, predominantly Black, diverse), self-reported educational quality, and school type (public/private). Later-life global cognition was operationalized with a factor score derived from a comprehensive neuropsychological battery. Sequential mediation models controlling for sociodemographics estimated total, direct, and indirect effects of early-life contextual factors on cognition through educational attainment (years). RESULTS: Higher educational quality, higher parental education, and attending a private school were each associated with better cognition; attending a predominantly Black or diverse school and reporting three or more adults in the household were associated with lower cognition. After accounting for educational attainment, associations remained for educational quality, school type, and reporting three or more adults in the household. Indirect effects through educational attainment were observed for school region, educational quality, school racial demographics, and parental education. CONCLUSIONS: School factors appear to consistently predict later-life cognition more than household factors, highlighting the potential long-term benefits of school-level interventions for cognitive aging. Future research should consider additional mediators beyond educational attainment such as neighborhood resources and childhood adversity.

Social network characteristics moderate associations between cortical thickness and cognitive functioning in older adults.

INTRODUCTION: Prior research suggests that the strength of association between Alzheimer's disease (AD) pathology and lower cognitive performance is influenced by modifiable psychosocial factors, such as social network size. However, little is known about distinct social relationship types. METHODS: The current cross-sectional study used data from the Washington Heights-Inwood Columbia Aging Project to examine whether social network characteristics (i.e., total size, spouse/partner, number of children, other relatives, friends) moderate associations between cortical thickness in regions implicated in AD and cognitive performance. RESULTS: Lower cortical thickness was associated with worse global cognition among individuals with smaller friend networks, but not among individuals with larger friend networks. This pattern of results was most prominent for language and speed/executive functioning. DISCUSSION: Longitudinal and intervention studies are needed to determine whether these cross-sectional findings reflect a protective effect of later-life friendships for maintaining cognitive performance in the context of poorer brain health.

Positive Psychosocial Factors and Cognitive Decline in Ethnically Diverse Older Adults.

OBJECTIVES: Previous cross-sectional studies have documented associations between positive psychosocial factors, such as self-efficacy and emotional support, and late-life cognition. Further, the magnitudes of concurrent associations may differ across racial and ethnic groups that differ in Alzheimer's disease risk. The goals of this longitudinal study were to characterize prospective associations between positive psychosocial factors and cognitive decline and explicitly test for differential impact across race and ethnicity. METHODS: 578 older adults (42% non-Hispanic Black, 31% non-Hispanic White, and 28% Hispanic) in the Washington Heights-Inwood Columbia Aging Project completed cognitive and psychosocial measures from the NIH Toolbox and standard neuropsychological tests over 2.4 years. Latent difference scores were used to model associations between positive psychosocial factors and cognitive decline controlling for baseline cognition, sociodemographics, depressive symptoms, physical health, and other positive psychosocial factors. Multiple-group modeling was used to test interactions between the positive psychosocial factors and race/ethnicity. RESULTS: Higher NIH Toolbox Friendship scores predicted less episodic memory decline. One standard deviation increase in friendship corresponded to 6 fewer years of memory aging. This association did not significantly differ across racial/ethnic groups. CONCLUSIONS: This longitudinal study provides support for the potential importance of friendships for subsequent episodic memory trajectories among older adults from three ethnic groups. Further study into culturally informed interventions is needed to investigate whether and how friend networks may be targeted to promote cognitive health in late life.

Mediators and Moderators of the Association Between Perceived Stress and Episodic Memory in Diverse Older Adults.

OBJECTIVE: Stress is a risk factor for numerous negative health outcomes, including cognitive impairment in late-life. The negative association between stress and cognition may be mediated by depressive symptoms, which separate studies have identified as both a consequence of perceived stress and a risk factor for cognitive decline. Pathways linking perceived stress, depressive symptoms, and cognition may be moderated by sociodemographics and psychosocial resources. The goal of this cross-sectional study was to identify modifying factors and enhance understanding of the mechanisms underlying the stress-cognition association in a racially and ethnically diverse sample of older adults. METHOD: A linear regression estimated the association between perceived stress and episodic memory in 578 older adults (Mage = 74.58) in the Washington Heights-Inwood Columbia Aging Project. Subsequent models tested whether depressive symptoms mediated the stress-memory relationship and whether sociodemographics (gender, race, and ethnicity) or perceived control moderated these pathways. RESULTS: Independent of sociodemographics and chronic diseases, greater perceived stress was associated with worse episodic memory. This relationship was mediated by more depressive symptoms. Higher perceived control buffered the association between stress and depressive symptoms. There was no significant moderation by gender, race, or ethnicity. CONCLUSION: Depressive symptoms may play a role in the negative association between perceived stress and cognition among older adults; however, longitudinal analyses and studies using experimental designs are needed. Perceived control is a modifiable psychological resource that may offset the negative impact of stress.

Comparison of Education and Episodic Memory as Modifiers of Brain Atrophy Effects on Cognitive Decline: Implications for Measuring Cognitive Reserve.

OBJECTIVE: This study compared the level of education and tests from multiple cognitive domains as proxies for cognitive reserve. METHOD: The participants were educationally, ethnically, and cognitively diverse older adults enrolled in a longitudinal aging study. We examined independent and interactive effects of education, baseline cognitive scores, and MRI measures of cortical gray matter change on longitudinal cognitive change. RESULTS: Baseline episodic memory was related to cognitive decline independent of brain and demographic variables and moderated (weakened) the impact of gray matter change. Education moderated (strengthened) the gray matter change effect. Non-memory cognitive measures did not incrementally explain cognitive decline or moderate gray matter change effects. CONCLUSIONS: Episodic memory showed strong construct validity as a measure of cognitive reserve. Education effects on cognitive decline were dependent upon the rate of atrophy, indicating education effectively measures cognitive reserve only when atrophy rate is low. Results indicate that episodic memory has clinical utility as a predictor of future cognitive decline and better represents the neural basis of cognitive reserve than other cognitive abilities or static proxies like education.

Longitudinal Associations Between Contact Frequency with Friends and with Family, Activity Engagement, and Cognitive Functioning.

OBJECTIVES: Social engagement may be an important protective resource for cognitive aging. Some evidence suggests that time spent with friends may be more beneficial for cognition than time spent with family. Because maintaining friendships has been demonstrated to require more active maintenance and engagement in shared activities, activity engagement may be one underlying pathway that explains the distinct associations between contact frequency with friends versus family and cognition. METHODS: Using two waves of data from the national survey of Midlife in the United States (n = 3707, Mage = 55.80, 51% female at baseline), we examined longitudinal associations between contact frequency with friends and family, activity engagement (cognitive and physical activities), and cognition (episodic memory and executive functioning) to determine whether activity engagement mediates the relationship between contact frequency and cognition. RESULTS: The longitudinal mediation model revealed that more frequent contact with friends, but not family, was associated with greater concurrent engagement in physical and cognitive activities, which were both associated with better episodic memory and executive functioning. CONCLUSION: These findings suggest that time spent with friends may promote both cognitively and physically stimulating activities that could help to preserve not only these social relationships but also cognitive functioning.

Psychological predictors of memory decline in a racially and ethnically diverse longitudinal sample of older adults in the United States.

OBJECTIVES: In the United States, racial and ethnic disparities in memory dysfunction and Alzheimer disease are evident even after accounting for many risk factors. Psychological factors, such as psychological well-being, perceived control, depressive symptoms, and negative affect, may influence memory dysfunction, and associations may differ by race and ethnicity. This study examined whether psychological factors are differentially associated with episodic memory trajectories across racial and ethnic groups in the United States. METHODS/DESIGN: The National Health and Aging Trends Study (NHATS), is a US-representative, longitudinal study of Medicare-eligible adults 65+ years old. Analyses of 5 years of data, included a total of 9411 participants without dementia at baseline. Adjusting for relevant covariates, a linear mixed model estimated the associations between psychological predictors and a composite of immediate and delayed trials from a word list memory test. RESULTS: More depressive symptoms (B = -0.02), lower psychological well-being (B = 0.03), and lower perceived control (B = 0.05) were independently associated with lower initial memory. Depressive symptoms were associated with faster rate of memory decline (B = -0.01). Black (B = -0.34) and Hispanic (B = -0.28) participants evidenced lower initial memory level than whites, but only Hispanic (B = -0.04) participants evidenced faster memory decline than whites. There were no significant interactions between the psychological variables and race and ethnicity. CONCLUSIONS: Results extend previous studies showing racial and ethnic disparities in episodic memory trajectories, and the longitudinal effects of depressive symptoms on episodic memory in US samples. Epidemiological studies of cognitive aging should incorporate more psychological factors clarify cognitive decline and disparities.

The longitudinal association between social network composition and episodic memory in older adulthood: the importance of contact frequency with friends.

Objectives: The composition of one's social network has been associated with cognition such that a greater proportion of family is associated with worse cognition compared to a greater proportion of friends. It is not clear whether this association between network composition and cognitive aging is driven by potential negative effects of family interactions or positive effects of friend interactions.Methods: Using the Health and Retirement Study (T1: 2006/2008, T2: 2010/2012, T3: 2012/2014), a longitudinal mediation model was conducted to test the effects of composition on episodic memory and latent change in memory through contact frequency with friends and family.Results: Analyses revealed indirect effects of composition on both T2 memory and latent change in memory through contact frequency with friends. A greater proportion of family in one's network was associated with lower contact frequency with friends and in turn lower memory. Composition was also associated with higher contact frequency with family; however, contact frequency with family was not associated with memory.Conclusions: These findings suggest that spending time with family may not affect episodic memory in older adulthood, but spending time with friends may be beneficial. Potential mechanisms and implications regarding the importance of friendships in later life are discussed.

Inflammatory mechanisms underlying the effects of everyday discrimination on age-related memory decline.

BACKGROUND/OBJECTIVES: Previous research suggests that everyday discrimination is associated with worse episodic memory and partially mediates Black-White disparities in memory aging. The biological mechanisms underlying the link between everyday discrimination and memory are unclear but may involve inflammatory processes. This study aimed to determine whether systemic inflammation, indexed by blood levels of C-reactive protein (CRP), mediates associations between everyday discrimination and episodic memory over 6 years. DESIGN: A longitudinal mediation model quantified associations between baseline everyday discrimination, 4-year change in CRP, and 6-year change in episodic memory. SETTING: The Health and Retirement Study (HRS). PARTICIPANTS: 12,624 HRS participants aged 51 and older. MEASUREMENTS: Everyday Discrimination Scale, high-sensitivity CRP assays of dried blood spots, composite scores of immediate and delayed recall of a word list. RESULTS: Black participants reported greater everyday discrimination. Greater discrimination was associated with lower baseline memory and faster memory decline. Higher CRP at baseline partially mediated the negative association between discrimination and baseline memory, but CRP change did not mediate the association between discrimination and memory decline. CONCLUSION: This U.S.-representative longitudinal study provides evidence for deleterious effects of discrimination on subsequent episodic memory. The fact that elevated CRP only partially explained the concurrent association between discrimination and memory highlights the need for more comprehensive investigations of biological mechanisms underlying the link between social stress and age-related memory decline in order to better characterize potential intervention targets to reduce racial inequalities in memory aging.

Predictors of Episodic Memory Performance Across Educational Strata: Multiple-Group Comparisons.

OBJECTIVES: Low educational attainment is a risk factor for more rapid cognitive aging, but there is substantial variability in cognitive trajectories within educational groups. The aim of this study was to determine the factors that confer resilience to memory decline within educational strata. METHODS: We selected 2573 initially nondemented White, African American, and Hispanic participants from the longitudinal community-based Washington Heights/Inwood Columbia Aging Project who had at least two visits. We estimated initial memory (intercept) and the rate of memory decline (slope) using up to five occasions of measurement. We classified groups according to the educational attainment groups as low (≤5 years), medium (6-11 years), and high (≥12 years). We used a multiple-group latent growth model to identify the baseline predictors of initial memory performance and rate of memory decline across groups. The model specification considered the influence of demographic, socioeconomic, biomedical, and cognitive variables on the intercept and the slope of memory trajectory. RESULTS: Our results indicated that the three educational groups do not benefit from the same factors. When allowed to differ across groups, the predictors were related to cognitive outcomes in the highly educated group, but we found no unique predictor of cognition for the low educated older adults. CONCLUSIONS: These findings highlight that memory-protective factors may differ across older adults with distinct educational backgrounds, and the need to evaluate a broader range of potential resilience factors for older adults with few years of school.

Positive Psychosocial Factors and Cognition in Ethnically Diverse Older Adults.

OBJECTIVES: Cross-sectional and longitudinal evidence from largely non-Hispanic White cohorts suggests that positive psychosocial factors, particularly self-efficacy and social support, may protect against late-life cognitive decline. Identifying potentially protective factors in racial/ethnic minority elders is of high importance due to their increased risk of Alzheimer's disease. The overall goal of this study was to characterize cross-sectional associations between positive psychosocial factors and cognitive domains among Black, Hispanic, and White older adults. METHODS: A total of 548 older adults (41% Black, 28% Hispanic, 31% White) in the Washington Heights-Inwood Columbia Aging Project completed cognitive and psychosocial measures from the NIH Toolbox and standard neuropsychological tests. Multiple-group regressions were used to compare cross-sectional associations between positive psychosocial factors and cognition across racial/ethnic groups, independent of demographics, depressive symptoms, and physical health. RESULTS: Positive associations between self-efficacy and language did not significantly differ across race/ethnicity, although the bivariate association between self-efficacy and language was not significant among Hispanics. Additional positive associations were observed for Whites and Blacks, but not Hispanics. Negative associations between emotional support and purpose in life and working memory were seen only in Hispanics. CONCLUSIONS: Results confirm and extend the link between self-efficacy and cognition in late life, particularly for White and Black older adults. Previous studies on positive psychosocial factors in cognitive aging may not be generalizable to Hispanics. Longitudinal follow-up is needed to determine whether negative relationships between certain psychosocial factors and cognition in Hispanics reflect reverse causation, threshold effects, and/or negative aspects of having a strong social network. (JINS, 2018, 24, 294-304).

Subjective Social Status Predicts Late-Life Memory Trajectories through Both Mental and Physical Health Pathways.

BACKGROUND: Subjective social status (SSS) is associated with mental and physical health, independent of objective socioeconomic status (SES), but its association with late-life cognitive decline is unknown. OBJECTIVE: This study characterizes the association between SSS and late-life memory trajectories in a large, nationally representative sample of older adults in the United States. METHODS: Using data from 8,530 participants aged 65 years and older in the Health and Retirement Study, structural equation models tested associations between SSS, objective SES (i.e., educational attainment, occupation, income, and wealth), physical and mental health, and 6-year memory trajectories, controlling for sociodemographic characteristics. RESULTS: Independent of objective SES, lower SSS was associated with worse initial memory but not subsequent memory decline. The association between SSS and initial memory was separately mediated by chronic diseases, stroke, and depressive symptoms. CONCLUSION: Results provide preliminary behavioral evidence for the deleterious effects of social stress on cognitive aging. These results may help inform the development of policies and interventions to reduce cognitive morbidity among older adults who perceive a low position on the social hierarchy.

Comparing Variability, Severity, and Persistence of Depressive Symptoms as Predictors of Future Stroke Risk.

OBJECTIVE: Numerous studies show that depressive symptoms measured at a single assessment predict greater future stroke risk. Longer-term symptom patterns, such as variability across repeated measures or worst symptom level, might better reflect adverse aspects of depression than a single measurement. This prospective study compared five approaches to operationalizing depressive symptoms at annual assessments as predictors of stroke incidence. DESIGN: Cohort followed for incident stroke over an average of 6.4 years. SETTING: The Adult Changes in Thought cohort follows initially cognitively intact, community- dwelling older adults from a population base defined by membership in Group Health, a Seattle-based nonprofit healthcare organization. PARTICIPANTS: 3,524 individuals aged 65 years and older. MEASUREMENTS: We identified 665 incident strokes using ICD codes. We considered both baseline Center for Epidemiologic Studies-Depression scale (CES-D) score and, using a moving window of three most recent annual CES-D measurements, we compared most recent, maximum, average, and intra-individual variability of CES-D scores as predictors of subsequent stroke using Cox proportional hazards models. RESULTS: Greater maximum (hazard ratio [HR]: 1.18; 95% CI: 1.07-1.30), average (HR: 1.20; 95% CI: 1.05-1.36) and intra-individual variability (HR: 1.15; 95% CI: 1.06-1.24) in CES-D were each associated with elevated stroke risk, independent of sociodemographics, cardiovascular risks, cognition, and daily functioning. Neither baseline nor most recent CES-D was associated with stroke. In a combined model, intra-individual variability in CES-D predicted stroke, but average CES-D did not. CONCLUSIONS: Capturing the dynamic nature of depression is relevant in assessing stroke risk. Fluctuating depressive symptoms may reflect a prodrome of reduced cerebrovascular integrity.