Being married may protect late-life cognition. Less is known about living arrangement among unmarried adults and mechanisms such as brain health (BH) and cognitive reserve (CR) across race and ethnicity or sex/gender. The current study examines (1) associations between marital status, BH, and CR among diverse older adults and (2) whether one’s living arrangement is linked to BH and CR among unmarried adults.

Cross-sectional data come from the Washington Heights-Inwood Columbia Aging Project (N = 778, 41% Hispanic, 33% non-Hispanic Black, 25% non-Hispanic White; 64% women). Magnetic resonance imaging (MRI) markers of BH included cortical thickness in Alzheimer’s disease signature regions and hippocampal, gray matter, and white matter hyperintensity volumes. CR was residual variance in an episodic memory composite after partialing out MRI markers. Exploratory analyses stratified by race and ethnicity and sex/gender and included potential mediators.

Marital status was associated with CR, but not BH. Compared to married individuals, those who were previously married (i.e., divorced, widowed, and separated) had lower CR than their married counterparts in the full sample, among White and Hispanic subgroups, and among women. Never married women also had lower CR than married women. These findings were independent of age, education, physical health, and household income. Among never married individuals, living with others was negatively linked to BH.

Marriage may protect late-life cognition via CR. Findings also highlight differential effects across race and ethnicity and sex/gender. Marital status could be considered when assessing the risk of cognitive impairment during routine screenings.

3q29 deletion syndrome (3q29del) is a rare (~1:30 000) genomic disorder associated with a wide array of neurodevelopmental and psychiatric phenotypes. Prior work by our team identified clinically significant executive function (EF) deficits in 47% of individuals with 3q29del; however, the nuances of EF in this population have not been described.

We used the Behavior Rating Inventory of Executive Function (BRIEF) to perform the first in-depth assessment of real-world EF in a cohort of 32 individuals with 3q29del (62.5% male, mean age = 14.5 ± 8.3 years). All participants were also evaluated with gold-standard neuropsychiatric and cognitive assessments. High-resolution structural magnetic resonance imaging was performed on a subset of participants (n = 24).

We found global deficits in EF; individuals with 3q29del scored higher than the population mean on the BRIEF global executive composite (GEC) and all subscales. In total, 81.3% of study subjects (n = 26) scored in the clinical range on at least one BRIEF subscale. BRIEF GEC T scores were higher among 3q29del participants with a diagnosis of attention deficit/hyperactivity disorder (ADHD), and BRIEF GEC T scores were associated with schizophrenia spectrum symptoms as measured by the Structured Interview for Psychosis-Risk Syndromes. BRIEF GEC T scores were not associated with cognitive ability. The BRIEF-2 ADHD form accurately (sensitivity = 86.7%) classified individuals with 3q29del based on ADHD diagnosis status. BRIEF GEC T scores were correlated with cerebellar white matter and subregional cerebellar cortex volumes.

Together, these data expand our understanding of the phenotypic spectrum of 3q29del and identify EF as a core feature linked to both psychiatric and neuroanatomical features of the syndrome.

To compare the agreement and cost of two recall methods for estimating children’s minimum dietary diversity (MDD).

We assessed child’s dietary intake on two consecutive days: an observation on day one, followed by two recall methods (list-based recall and multiple-pass recall) administered in random order by different enumerators at two different times on day two. We compared the estimated MDD prevalence using survey-weighted linear probability models following a two one-sided test equivalence testing approach. We also estimated the cost-effectiveness of the two methods.

Cambodia (Kampong Thom, Siem Reap, Battambang, and Pursat provinces) and Zambia (Chipata, Katete, Lundazi, Nyimba, and Petauke districts).

Children aged 6–23 months: 636 in Cambodia and 608 in Zambia.

MDD estimations from both recall methods were equivalent to the observation in Cambodia but not in Zambia. Both methods were equivalent to the observation in capturing most food groups. Both methods were highly sensitive although the multiple-pass method accurately classified a higher proportion of children meeting MDD than the list-based method in both countries. Both methods were highly specific in Cambodia but moderately so in Zambia. Cost-effectiveness was better for the list-based recall method in both countries.

The two recall methods estimated MDD and most other infant and young child feeding indicators equivalently in Cambodia but not in Zambia, compared to the observation. The list-based method produced slightly more accurate estimates of MDD at the population level, took less time to administer and was less costly to implement.

Women have a greater lifetime risk of developing Alzheimer’s disease (AD) dementia than men, a sex/gender disparity that cannot be explained by female longevity alone. There is substantial evidence for sex differences in the effects of APOE £4 on risk for AD. While APOE e4 increases AD risk in both sexes, women who carry APOE e4 are disproportionately vulnerable to cognitive impairment and AD compared to their counterpart men. In contrast to APOE e4, APOE £2 is associated with slower cognitive decline and a lower risk of AD. Although a less robust literature, APOE e2 may also have sex-specific effects. Because APOE e2 is the rarest major APOE allele, well-powered studies are needed to examine sex-specific effects. The objective of the present study was to examine sex-specific associations of APOE e2 carriage with longitudinal cognitive decline in a large cohort of clinically unimpaired adults.

We used observational data from two sources: the National Alzheimer’s Coordinating Center (NACC) and the Rush Alzheimer’s Disease Center (ROS/MAP/MARS) studies. We included data from clinically unimpaired adults who were >50 years old at baseline who self-identified as non-Hispanic White (NHW) or non-Hispanic Black (NHB). Participants were categorized as APOE £2, £4, or £3/e3 carriers. APOE e2/e4 carriers were excluded. The same battery of neuropsychological tests was used to assess global cognition in participants from both data sources. Linear mixed models examined interactive associations of genotype (£2 or £4 vs. £3/£3), sex, and time on longitudinal cognition in NHW and NHB participants separately. Analyses were first performed in a pooled sample of NACC and ROS/MAP/MARS participants and if significant they were repeated separately in each data source.

Across both data sources, 9,766 NHW (mean (SD) age=73.0(9.00) years, mean (SD) education=16.3(2.83) years, n(%) women=6,344(65.0)) and 2,010 NHB participants (mean(SD) age=71.3(7.59) years, mean(SD) education=14.9(3.10) years, n(%) women=1,583(78.8)) met inclusion criteria. Sex modified the association between APOE £2 and cognitive decline in NHW (ß=0.097, 95% CI: 0.023-0.172, pint=.01) but not NHB participants (ß=-0.011, 95% CI: -0.153-0.131, pint=.9). In sex-stratified analyses of NHW participants, APOE £2 (vs. £3/£3) carriage was associated with attenuated cognitive decline in men (ß=0.096, 95% CI: 0.037-0.155, p=.001), but not women (ß=-0.001, 95% CI: -0.044-0.043, p=.97). In analyses comparing men and women APOE £2 carriers, men exhibited slower cognitive decline than women (ß=0.120, 95% CI: 0.051-0.190, p=.001). Analyses performed separately in NACC and ROS/MAP revealed the same pattern of male-specific APOE £2 protection in NHW participants in both data sources.

In light of the longstanding view that APOE £2 protects against AD and dementia, our results provide evidence that APOE £2 is associated with attenuated cognitive decline in men but not women among NHW adults. This male-specific protection may contribute to sex differences in AD-related cognitive decline. Our findings have important implications for understanding the biological drivers of sex differences in AD risk, which is crucial for developing sex-specific strategies to prevent and treat AD dementia.

People whose parents had dementia or memory impairment are at higher risk for later-life cognitive impairment themselves. One goal of our research is to identify factors that either increase the risk of or protect against family history of dementia over the life course. External locus of control has been associated with lower cognitive function in middle-aged and older adults. Previous findings have shown that adults racialized as Black have relatively high levels of external locus of control due to inequity and racism. We hypothesized that lower parental memory would be associated with lower offspring memory among Non-Latinx Black and Non-Latinx White (hereafter Black and White, respectively) adults; and associations would be stronger among participants with higher levels of external locus of control.

Participants comprised 594 adults racialized as Black or White (60.3% Black; 62% women; aged 56.1 ± 10.4; 15.3 ± 2.7 years of education) from the Offspring Study who are the adult children of participants in the Washington Heights Inwood Columbia Aging Project (WHICAP). Parental memory was residualized for age (74.3 ± 6.0) and education (13.7 ± 3.1). Self-reported external locus of control was assessed using 8 items from the the perceived control questionnaire. Memory was assessed with the Selective Reminding Test, and a composite of total and delayed recall scores were computed. Linear regression quantified the interaction between parental memory and external locus of control on memory in models stratified by race, and adjusted for age, sex/gender, and number of chronic health diseases.

Among Black participants (n=358), there were no main effects of parental memory or locus of control on offspring memory. However, lower parental memory was associated with lower offspring memory among Black participants with high levels of external locus of control (standardized estimate=0.36, p=0.02, 95%CI [0.05, 0.67]). Associations were attenuated and non-significant at lower levels of control. Among White participants (n=236), there was a main effect of parental memory on offspring memory, and this association did not vary by levels of external locus of control.

Poor parental memory, which reflects risk for later-life cognitive impairment and dementia, was associated with lower memory performance among White middle-aged participants. Among Black participants, this association was observed among those with high levels of external locus of control only. Economic and social constraints shape levels of external locus of control and are disproportionately experienced by Black adults. In the face of greater external locus of control, a cascade of psychological and biological stress-related processes may be triggered and make Black adults’ memory function more vulnerable to the detrimental impact of parent-related dementia risk. Longitudinal analyses are needed to clarify temporal associations. Nonetheless, these findings suggest that reducing social and economic inequities disproportionately experienced by Black adults may dampen the effect of intergenerational transmission of dementia risk on cognition.

We assess the association between self-reported history of periodontal disease diagnosis with self-reported cognitive and functional decline at age ∼60. We also investigate (1) the roles of social background, demographic characteristics, education, and adolescent test scores in confounding that association and (2) the role of cardiovascular disease in mediating that association.

We use data from a nationally representative sample of 13,525 people who participated in the 2021 wave of the High School & Beyond (HSB) cohort study. HSB began in 1980 with a nationally-representative sample of American 10th and 12th grade students; these students have been followed up on six occasions since 1980, yielding extraordinary and prospectively-collected life course data on all key measures for a large, multicultural sample.

In 2021, HSB sample members were evaluated with neuropsychological tests that evaluated list learning and memory, semantic and letter fluency, and working memory. They were also asked to self-report memory and functional decline using the AD8, using a cutoff of 2 or more items for significant concerns. Mild Cognitive Impairment will be identified using an algorithm validated in a similar sample of middle aged participants.

Sample members were also asked in 2021 whether a medical professional had ever diagnosed them with periodontal disease; those responding affirmatively were then asked the years in which they started and stopped having periodontal disease.

Measures of social and economic background; demographic characteristics; and educational contexts, opportunities, and attainments were measured prospectively—and in great detail—in the surveys administered in the 1980s. Critically, almost all sample members completed a series of cognitive tasks during adolescence, allowing us to address a key set of confounders in the relationship between periodontal disease and MCI. Markers of cardiovascular disease were measured in both 2013 and 2021.

We estimate logistic regression models predicting significant cognitive and functional concerns as a function of periodontal disease history; we also estimate models that account for confounders, including social background, demographic characteristics, education, and cognitive skills during adolescence; finally, we estimate models that account for the mediating role of cardiovascular disease. All models account for the clustered sampling design of HSB and employ sampling weights to account for HSB’s complex sampling design and selective attrition from the panel.

About 15% of the cohort has been diagnosed with periodontal disease, and nearly one in five had significant cognitive and functional concerns.

People with a history of periodontal disease were more likely to report significant cognitive and functional concerns. This association remains substantive and statistically significant after adjusting for confounders. All else equal, the odds of people with a history of periodontal disease having an AD8 score of 2 or higher were about 60% greater than the odds of those not reporting periodontal disease. Very little of this association can be attributed to cardiovascular disease as a mediating pathway.

People with a history of periodontal disease are at greatly elevated risk of self-reported cognitive and functional concerns at age ∼60. This supports evidence— never before collected at this scale in a long-term, representative cohort study—that oral pathogens may contribute to cognitive well-being over the life course.

Socioeconomic disadvantage is a chronic stressor associated with several biological markers of health (e.g., inflammation) as well as early-onset cognitive aging. Studies examining socioeconomic status (SES) and its link with health outcomes exhibit no uniformity in the way in which SES is measured and defined. Also, studies have found that subjective socioeconomic status (SSES), defined by a subjective SES scale, was more consistently and strongly related to psychological functioning and health-related outcomes than objective socioeconomic status (OSES), defined by a composite score of education, household income, and occupation. The goal of the current study was to assess whether SSES mediates the relationship between OSES and neuropsychological test performance similarly across racial and ethnic groups.

Participants were 1,912 middle-aged older adults (13% non-Hispanic white, 17% non-Hispanic Black, 69% Hispanic/Latinx) from the Offspring study. Participants are the adult children of participants in the Washington Heights Inwood Columbia Aging Project, a community-based cohort study of aging and dementia representing the ethnic/racial diversity of upper Manhattan. Participants on average were 56.5 years of age and 67% were women. Measures of verbal learning and memory (SRT immediate and delayed recall), verbal fluency (animal and letter fluency), and attention/working memory (digit span forward and backward) were administered. OSES was characterized by years of formal education completed. SSES was measured by the MacArthur Scale of Subjective Social Status. The scale measures perceptions of one’s social standing relative to others. We conducted separate stratified mediation analyses for each neuropsychological outcome across each racial and ethnic group. All models were adjusted for age.

Participants with higher OSES demonstrated higher neuropsychological test scores (effect size associations ranged from .29 to .45) and reported higher SSES (b=.109 95% CI: .08, .14). Lower SSES was associated with lower neuropsychological test scores (effect-size range .06 to .13). In stratified analyses, the relationship between OSES and SSES was strongest for White participants (b=.13 [.01, .24]) compared with Latinx (b=.06 [.02, .11]) and Black (b=.06 [-.03, .16]) participants. Associations between SSES and neuropsychological outcomes were only reliable for White participants on SRT Immediate and Delayed Recall and Animal Fluency and for Black participants on Digit Span Forward. In mediation analyses, SSES mediated the relationship between OSES and Immediate Recall (indirect effect b=.18 [.001, .45]; 39% mediated), Delayed Recall (indirect effect b=.05 [.004, .09]; 44% mediated), and Animal Fluency (indirect effect b=.09 [.01, .20]; 22% mediated) for White participants. There was no evidence of mediation in Black or Latinx participants.

The relationship between OSES and SSES was strongest for White participants compared to Black and Latinx participants. Even though perception of social status predicted lower cognitive test scores among Black and Latinx adults, it is only a part of the indirect pathway linking OSES to cognitive function among White adults. It is likely that mechanisms related to tangible resources that benefit health (as opposed to perceived inequity) are in the pathway linking education to cognition among Black and Latinx, and thus intervening on systems of inequality throughout the life course has the most promise for improving brain health in those communities.

The AD8 is a validated screening instrument for functional changes that may be caused by cognitive decline and dementia. It is frequently used in clinics and research studies because it is short and easy to administer, with a cut off score of 2 out of 8 items recommended to maximize sensitivity and specificity. This cutoff assumes that all 8 items provide equivalent “information” about everyday functioning. In this study, we used item response theory (IRT) to test this assumption. To determine the relevance of this measure of everyday functioning in men and women, and across race, ethnicity, and education, we conducted differential item functioning (DIF) analysis to test for item bias.

Data came from the 2021 follow up of the High School & Beyond cohort (N=8,690; mean age 57.5 ± 1.2; 55% women), a nationally representative, longitudinal study of Americans who were first surveyed in 1980 when they were in the 10th or 12th grade. Participants were asked AD8 questions about their own functioning via phone or internet survey. First, we estimated a one-parameter (i.e., differing difficulty, equal discrimination across items) and two-parameter IRT model (i.e., differing difficulty and differing discrimination across items). We compared model fit using a likelihood-ratio test. Second, we tested for uniform and non-uniform DIF on AD8 items by sex, race and ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic), education level (high school or less, some college, BA degree or more), and survey mode (phone or internet). We examined DIF salience by comparing the difference between original and DIF-adjusted AD8 scores to the standard error of measurement of the original score.

The two-parameter IRT model fit the data significantly better than the one-parameter model, indicating that some items were more strongly related to underlying everyday functional ability than others. For example, the “problems with judgment” item had higher discrimination (more information) than the “less interest in hobbies/activities” item. There were significant differences in item endorsement by race/ethnicity, education, and survey mode. We found significant uniform and non-uniform DIF on several items across each of these groups. For example, for a given level of functional decline (theta) White participants were more likely to endorse “Daily problems with thinking/memory” than Black and Hispanic participants. The DIF was salient (i.e., caused AD8 scores to change by greater than the standard error of measurement for a large portion of respondents) for those with a college degree and phone respondents.

In a population representative sample of Americans ∼age 57, the items on the AD8 contributed differing levels of discrimination along the range of everyday functioning that is impacted by later life cognitive impairment. This suggests that a simple cut-off or summed score may not be appropriate since some items yield more information about the underlying construct than others. Furthermore, we observed significant and salient DIF on several items by education and survey mode, AD8 scores should not be compared across education groups and assessment modes without adjustment for this measurement bias.

The current research framework recommends using biomarkers to further understand Alzheimer’s disease (AD) pathogenesis, including other contributing factors like cerebrovascular disease. In longitudinal studies of people with neuropathological examination after death, baseline loneliness was associated with lower cognition, faster cognitive decline, and future AD risk, independent of AD pathology. Examination of memory impairment along with AD and cerebrovascular biomarkers, could aid risk reduction efforts earlier in the lifecourse and among populations with more exposure to loneliness. We hypothesized that loneliness is associated with amyloid, vascular, and neurodegeneration biomarkers; with worse memory; and that loneliness increases the susceptibility to biomarker-related memory impairment.

A subset of cognitively unimpaired older adults with available amyloid PET, vascular MRI (white matter hyperintensity volume, WMH), structural MRI (cortical thickness in AD signature regions), neuropsychological testing (memory factor score), dichotomized loneliness data (one item from CES-D), and relevant medical data were drawn from the community-based Washington Heights-Inwood Columbia Aging Project (WHICAP; n=169; covariates included age=81±6 years; 63% women; 49/31/20% Non-Hispanic Black/Non-Hispanic White/Hispanic; education=13±4 years; 32% APOE-e4 carriers). General linear models in the overall sample and stratified by race and ethnicity tested the association between loneliness and AD and cerebrovascular biomarkers, loneliness and memory, and the interaction of loneliness and biomarkers on memory, adjusting for covariates.

Loneliness was endorsed in 18% of participants, marginally associated with older age (2.1 [-0.2, 4.4], p=0.08), was more likely in those with untreated diabetes (13/0.1% lonely/not lonely, p=0.001), associated with lower cortical thickness (-0.05 [-0.09, -0.02], p=0.01), and associated with lower memory (0.3 [-0.6, -0.001], p=0.05). In Non-Hispanic White participants, loneliness was associated with greater WMH volume (0.5 [0.07, 0.82], p=0.03), while in Hispanic participants, loneliness was associated with lower cortical thickness (-0.16 [-0.24, -0.08], p=0.0006). In Non-Hispanic Black participants, loneliness was associated with lower memory (-13 [-26, -0.5], p=0.05), and the association between lower cortical thickness and lower memory was stronger in those that endorse loneliness (5 [0.2, 10], p=0.05). In Hispanic participants, loneliness was associated with higher memory (13 [4, 22], p=0.009), but the association between higher amyloid burden and lower memory was stronger in those that endorse loneliness (-12 [-20, -4], p=0.006); further, loneliness was marginally associated with lower memory (-0.7 [-1.4, 0.1], p=0.09), independently of WMH.

Associations between loneliness and biomarkers may relate to health seeking behavior, reported as treatment status for diabetes, for cerebrovascular burden and general neurodegeneration, but might be more complex for amyloid. The degree to which loneliness increased the susceptibility to amyloid and neurodegeneration-related, but not cerebrovascular-related, memory impairment, specifically, may suggest that domains beyond memory should be considered. Future work should be longitudinal to disentangle the effects of loneliness from related constructs like depression and anxiety, incorporate other AD biomarkers such as hyperphosphorylated tau, and incorporate biological mechanisms (e.g., stress, inflammation) into models of loneliness and AD pathogenesis. Older adults from all backgrounds may be more susceptible to loneliness, which was associated with lower memory; culturally-humble, social support-based interventions may reduce the risk of cognitive impairment.

Non-Hispanic Black older adults experience a disproportionate burden of Alzheimer’s Disease and related dementias (ADRD) risk compared to non-Hispanic White older adults. It is necessary to identify mechanisms that may be contributing to inequities in cognitive aging. Psychosocial stressors that disproportionately affect Black adults (e.g., discrimination) have the potential to impact brain health through stress pathways. The brain’s white matter, which appears to be particularly important for ADRD risk among Black older adults, may be uniquely vulnerable to stress-related physiological dysfunction. To further understand whether and how discrimination can affect ADRD risk, this study aimed to examine associations between multiple forms of racial discrimination and white matter integrity, operationalized through diffusion tensor imaging.

Cross-sectional data were obtained from 190 non-Hispanic Black residents aged 65+ without dementia in northern Manhattan. Racial discrimination was self-reported using the Everyday Discrimination and Major Experiences of Lifetime Discrimination scales. Example items from the Everyday Discrimination Scale include: “You are treated with less respect than other people”; “You are called names or insulted.” Example items from the Major Experiences of Lifetime Discrimination Scale include: “At any time in your life, have you ever been unfairly fired from a job?”; “Have you ever been unfairly denied a bank loan?” Racial discrimination was operationalized as experiences attributed to “race” or “skin color.” White matter integrity was assessed using fractional anisotropy (FA) via diffusion tensor imaging. Multivariable regression models evaluated the unique effects of everyday and major experiences of lifetime racial discrimination on mean FA in the whole brain and specific regions. Initial models controlled for age, sex/gender, intracranial volume, and white matter hyperintensities. Subsequent models additionally controlled for socioeconomic and health factors to consider potential confounders or mediators of the relationship between discrimination and white matter integrity.

Major experiences of lifetime discrimination were negatively associated with mean FA within the left cingulum cingulate gyrus and the right inferior fronto-occipital fasciculus. These associations persisted when controlling for additional covariates (i.e., education, depression, and cardiovascular diseases). In contrast, major experiences of lifetime discrimination were positively associated with mean FA within the right superior longitudinal fasciculus (temporal part). This association was attenuated when controlling for additional covariates. Everyday racial discrimination was not associated with mean FA in any regions.

These results extend prior work linking racial discrimination to brain health and provide evidence for both risk and resilience among Black older adults. Major experiences of lifetime racial discrimination, a proxy for institutional racism, may have a stronger effect on white matter integrity than everyday racial discrimination, a proxy for interpersonal racism. Educational opportunities and cardiovascular risk factors may represent mediators between racial discrimination and white matter integrity. White matter integrity within specific brain regions may be a mechanism through which racially patterned social stressors contribute to racial disparities in ADRD. Future research should characterize within-group heterogeneity in order to identify factors that promote resilience among Black older adults.

We ask about the degree to which the association between (1) the quantity and quality of people’s education and (2) midlife self-reported concerns about cognition and daily function is mediated by occupational complexity. The overarching hypothesis is that amount and quality of education provides people with access to better jobs, including jobs that are more cognitively complex. Occupational complexity, in turn, may be protective against cognitive impairment. If true, this means that part of the poorly-understood connection between education and cognitive impairment can be attributed to occupational complexity.

We use data from a nationally representative sample of 13,525 people who participated in the 2021 wave of the High School & Beyond (HSB) cohort study. HSB began in 1980 with a nationally-representative sample of American 10th and 12th grade students; these students have been followed up on six occasions since 1980, yielding extraordinary and prospectively-collected life course data on all key measures for a large, multicultural sample.

In 2021, HSB sample members were evaluated with neuropsychological tests that evaluated list learning and memory, semantic and letter fluency, and working memory. They were also asked to self-report memory and functional decline using the AD8, using a cutoff of 2 or more items for significant concerns. Mild Cognitive Impairment will be identified using an algorithm validated in a similar sample of middle aged participants.

HSB surveys gathered information about sample members’ labor force activities in every survey between 1980 and 2021, including information sufficient to code verbatim reports of occupations to the standards of the 2010 Standard Occupational Classification. We have linked these codes for sample members’ 2013 and 2021 occupations to the O*Net database, which includes extensive information about the cognitive complexity (and other attributes) of every occupation.

Measures of key confounders—including social and economic background; demographic characteristics; educational contexts, opportunities, and attainments that are associated with labor force outcomes; adolescent achievement test scores; and aspects of midlife occupations besides complexity (e.g., how well they pay)—were measured prospectively and in great detail in the surveys administered between the 1980s and 2021.

We estimate logistic regression models predicting significant cognitive and functional concerns as a function of educational contexts, opportunities, and outcomes; we also estimate models that account for the confounders listed above. Our main focus is on coefficients for education in models that do and do not include occupational complexity as a mediator. All models account for the clustered sampling design of HSB and use sampling weights to account for HSB’s complex sampling design and selective attrition from the panel.

Nearly one in five cohort members had significant cognitive and functional concerns; rates are lower for non-Latinx Whites and for better educated people. Associations between educational contexts, opportunities, and outcomes (including attainment) are robust, even after adjusting for confounders.

Between one quarter and one third of the conditional association between education and self-reported cognitive and functional concerns can be attributed to occupational complexity.

Occupational complexity is an important pathway through which more and better education protects people from concerns about cognitive and functional decline at about age 60.