Associations of Lifetime Cognitive Enrichment With Incident Alzheimer Disease Dementia, Cognitive Aging, and Cognitive Resilience

Abstract

Background and Objectives

The effects of lifetime cognitive enrichment on later-life cognitive outcomes are not comprehensively investigated. The aim of this study was to test the association of lifetime cognitive enrichment with Alzheimer disease (AD) dementia and cognitive decline and in an autopsied deceased subset to explore the association between lifetime enrichment and AD and related dementia (ADRD) pathologic indices and cognitive resilience that is, decline after adjusting for common ADRD pathologies.

Methods

This was a longitudinal clinicopathologic study involving older individuals from Northeastern Illinois who participated in the Rush Memory and Aging Project, were free of dementia at baseline, completed surveys reflecting lifetime enrichment, and had annual clinical evaluations. We constructed a composite measure reflecting lifetime cognitive enrichment and tested its association with incident AD dementia in proportional hazards models, mean age of AD dementia onset in an accelerated failure time model, and cognitive decline using linear mixed-effects models. In a deceased subset, we tested the association of lifetime cognitive enrichment with 9 ADRD pathologies and cognitive resilience.

Results

Participants (n = 1,939, 75% female, mean baseline age = 79.6) completed an average of 7.6 years of follow-up, during which 551 participants developed AD dementia. One unit higher in lifetime enrichment was associated with 38% lower hazards of developing AD dementia (hazard ratio 0.62, 95% CI 0.52–0.73, p < 0.001). High lifetime enrichment (90th percentile) compared with low (10th percentile) was associated with a mean of 5 years delayed onset of AD dementia. Lifetime enrichment was positively associated with cognitive function at baseline (estimate = 0.31, SE = 0.02, p < 0.001) and a slower rate of cognitive decline (estimate = 0.02, SE = 0.01, p = 0.002). In the deceased subset (n = 948), lifetime cognitive enrichment did not show meaningful associations with neuropathologic indices, but remained associated with higher cognitive function proximate to death (estimate = 0.32, SE = 0.06, p < 0.001) and a slower rate of cognitive decline after adjusting for pathology (estimate = 0.014, SE = 0.01, p = 0.02).

Discussion

Lifetime exposure to cognitive enrichment was related to lower risk of AD dementia and a slower rate of cognitive decline, including after adjustment for common ADRD pathologies, indicating higher resilience provided by lifetime enrichment. Our results suggest that cognitive health in later life is in part the product of lifetime exposure to cognitive enrichment.

Introduction

Cognitive enrichment is a well-established marker of cognitive reserve and resilience.¹ It may reflect socioeconomic status (SES), access to diverse resources such as subscription to a daily newspaper, and active engagement in cognitive activities, such as reading books. Most of these measures have been previously studied in late-life and associated with lower dementia incidence, better cognitive function, and slower rate of cognitive decline.^2–7. However, the later-life cognitive benefits reaped by cognitive enrichment likely reflect lifetime effects.

In previous work, we showed that more frequent engagement in cognitive activities throughout the lifespan are associated with slower later-life cognitive decline⁸ including after adjusting for neuropathologic indices,⁹ and that higher engagement in cognitive activities in older age alone is associated with lower odds of Alzheimer disease (AD) dementia, mild cognitive impairment (MCI), and with slower rate of cognitive decline.¹⁰ However, in these studies, we did not fully account for measures that reflect access to resources, such as SES.^8–10 Yet, we separately showed that higher economic wealth during every life period from young adulthood into old age¹¹ and access to cognitive resources in midlife¹² are associated with late-life cognitive function. In addition, more frequent engagement in cognitive activities in mid-life is associated with slower cognitive decline in later-life, even after adjusting for effects of common AD and related dementias (ADRD) neuropathologic indices.⁹ Finally, we previously showed that a composite measure reflecting early life cognitive enrichment composed of measures for SES, access to cognitive resources, and frequency in cognitively stimulating activities in childhood, is related to later-life cognitive function.¹³ However, these measures were studied in isolation, and we did not comprehensively investigate how lifetime cognitive enrichment might relate to these outcomes.

Here, we leveraged data from the Rush Memory and Aging Project (MAP) to extend previous work in 3 important ways. We constructed a composite that reflects lifetime cognitive enrichment and examined its associations with incident AD dementia and baseline level and rate of cognitive decline in late life. In a subset who died and underwent autopsy we then examined the associations of lifetime cognitive enrichment with ADRD neuropathologic indices. Finally, we investigated the association between lifetime cognitive enrichment and cognitive decline after accounting for ADRD neuropathologic indices, thus investigating if lifetime enrichment is associated with cognitive resilience.

Methods

Participants

Study participants are from the Rush MAP, an ongoing longitudinal study that began in 1997. Participants live in northeastern Illinois (the United States) and were recruited from personal accommodations, subsidized housing, and retirement facilities. The recruitment strategy is designed to capture the full spectrum of conditions in an aging population. The only requirement is the ability to sign the consent and an Anatomic Gift Act for brain donation. All data collection is in the participants’ place of residence to minimize healthy volunteer bias. There we no other inclusion or exclusion criteria.

In this study, we included participants without dementia at baseline who completed the enrichment surveys and had at least 1 follow-up neuropsychological assessment (eFigure 1a). Neuropathologic analyses were restricted to deceased participants who completed brain autopsy and neuropathologic assessment (eFigure 1b).

Assessment of Lifetime Cognitive Enrichment

At baseline, participants completed questions on cognitive enrichment during early, mid, and late life. For each life stage, composites were created as the mean of the z-scored indicators.

Early-life enrichment included 4 indicators¹³: (1) SES (parental education and number of siblings); (2) cognitive resources at age 12 (range 0–8) years, based on household items such as newspaper subscription, encyclopedia, globe, or atlas; (3) frequency of cognitively stimulating activities (range 0–5) at age 6 (e.g., being read to) and 12 years (e.g., reading books); (4) and early-life foreign language instruction, based on self-reported years of foreign language instruction before age 18 years.

Midlife cognitive enrichment included the following 3 indicators: (1) income level age 40 using a “show-card” method (range $0–$4,999–≥$75,000); (2) cognitive resources (range 0–8) based on household items for example, magazines, dictionary, or library card; and (3) frequency of cognitively stimulating activities at age 40 (range 0–5) years.

Late-life enrichment included (1) frequency of engagement in cognitive activities (e.g., reading, writing, playing games, range 0–5) and (2) all sources of income at study enrollment, including wages, salaries, social security, retirement benefits, help from relatives, rent from property, etc. (range $0–$4,999 to ≥$75,000).

Diagnoses of AD Dementia and MCI

Cognitive status was determined annually based on computer-scoring of cognitive tests, clinical judgment by a neuropsychologist, and diagnostic classification by an experienced clinician.^14–16 AD dementia was diagnosed using criteria of the joint working group of the National Institute of Neurologic and Communicative Disorders/Stroke/AD and Related Disorders Association^15,17,18 requiring a history of cognitive decline and impairment ≥2 domains of cognitive function, one of which must be memory.¹⁵ MCI was classified based on impairment in ≥1 cognitive domain without meeting criteria for dementia.^15,19

Assessment of Cognitive Function

A battery of 21 neuropsychological tests was administered. The Mini-Mental State Examination assessed global cognition, and Complex Ideas aided diagnostic classification. The remaining 19 tests included 7 measures of episodic memory (Word List Memory, Word List Recall, and Word List Recognition and immediate and delayed recall of the East Boston Story and Logical Memory), 3 measures of semantic memory (15-item form of the Boston Naming Test, Verbal Fluency, and a 15-item word recognition test), 3 measures of working memory (Digit Span Forward and Digit Span Backward and a modified version of Digit Ordering), 4 measures of perceptual speed (Number Comparison, the oral version of the Symbol Digit Modalities Test, and 2 measures from a modified version of the Stroop Neuropsychological Screening Test: the number of color names correctly read minus the number of errors and the number of colors correctly named minus the number of errors), and 2 measures of visuospatial ability (15-item form of the Judgment of Line Orientation Test and a 16-item form of Standard Progressive Matrices).^20,21 A global composite score was derived by standardizing the 19 tests using baseline means and standard deviations.

Assessment of Postmortem Brain Neuropathology

Neuropathologic assessments were performed blinded to all clinical data and followed a standard protocol for tissue preservation, tissue sectioning, and quantification of pathologic findings.^22–24 Pathologic indices included a global measure of AD (mean of amyloid-β and PHF-tau in 8 regions),²⁵ hippocampal sclerosis, TAR DNA-binding protein 43, neocortical Lewy bodies, macroinfarcts, microinfarcts, cerebral amyloid angiopathy, atherosclerosis, and arteriolosclerosis.²⁶

Statistical Procedure

We examined correlations using Pearson correlation coefficients.

Incident AD Dementia and MCI

Proportional hazards models examined the association between lifetime cognitive enrichment and odds of developing incident dementia and MCI. Accelerated failure time models calculated mean time to dementia and MCI. Covariates included age at baseline, sex, and education.

Level of Cognition and Rate of Decline

Linear mixed-effects (LME) models tested whether lifetime cognitive enrichment was related to baseline level of cognition or rate of cognitive change. Models included terms for age at baseline, sex, education, lifetime enrichment, time in years since baseline, and the interaction of each measure with time. The term for enrichment estimated baseline cognitive function, and its interaction with time estimated annual change per SD increase in enrichment. The primary outcome was global cognition, and analyses were repeated using the 5 cognitive domains as outcomes.

ADRD Neuropathologic Analyses

We used separate linear and logistic regression models to examine the association between lifetime cognitive enrichment and ADRD pathologic indices, adjusting for age at death, sex, and education.

Cognitive Resilience

We ran LME models further adjusting for 9 ADRD neuropathologies and their interaction with time before death. The interaction term with time tests the effect of enrichment on cognitive change independent of pathology.

Secondary Analyses

Because the enrichment composites reflect both socioeconomic opportunity and behavioral engagement, we ran additional LME models testing each indicator’s effect on cognition individually and jointly, to identify which indicators were consistently associated with cognition. Models were adjusted for age, sex, and education.

Models testing resilience were further adjusted for 9 neuropathologic indices.

Standard Protocol Approvals, Registrations, and Patient Consents

The Rush MAP study was approved by an Institutional Review Board of Rush University Medical Center.¹⁴ All study participants provided written informed consent and an Anatomical Gift Act for organ donation.

Data Availability

All data included in these analyses are available at the Rush Alzheimer’s Disease Center Resource Sharing Hub. Descriptions of the studies can be found in this hub. Qualified investigators may create an account and submit requests for deidentified data.

Results

Participants

A total of 1,939 participants were included (mean baseline age ~80 years, highly educated, two-thirds female; Table 1). Participants were followed up for an average of 7.6 years (SD = 5 years, range 4–15).

Table 1.

Characteristics of the Full Sample (N = 1,939)

	Means (SD)	Range
Age at baseline, y (SD)	79.6 (7.5)	53.3, 100.5
Education, y (SD)	15.0 (3.3)	0, 30
Female, n (%)	1,458 (75)	-
Global cognition at baseline (SD)	0.10 (0.5)	−1.9, 1.8
Retired at enrollment (%)	1,741 (87)	−
Early-life socioeconomic statusa (SD)	0.01 (0.8)	−3.1, 4.7
Availability of cognitive resources at age 12b (SD)	4.4 (2.1)	0.0, 8.0
Frequency of participation in cognitively stimulating activities at age 6 and 12 c	3.0 (0.7)	1.0, 4.7
Early life foreign language instructiond (%)	1,385 (73)	−
Income at age 40e (SD)	6.1 (2.5)	1.0, 10
Cognitive resources at age 40f (SD)	6.1 (1.7)	0, 8.0
Cognitive activity frequency at age 40g (SD)	3.3 (0.6)	1.0, 4.8
Income at study enrollmenth (SD)	7.1 (2.6)	1.0, 10
Cognitive activity frequency at enrollmenti (SD)	3.2 (0.6)	1.0, 4.8
Lifetime enrichment j	0.066 (0.6)	−2.2, 1.5
Early-life enrichment j	0.052 (0.7)	−3.9, 1.8
Mid-life enrichment j	0.01 (0.8)	−3.5, 1.6
Late-life enrichment j	0.091 (0.8)	−3.2, 1.9

^aA composite index based on paternal and maternal education and number of children in the family (number of children is multiplied by −1). The index is the average of the 3 z scores.

^bTotal number of availabilities of 7 items (e.g., dictionary, world atlas, etc) and number of books in the home (responses are adjusted to range from 0 to 1). Overall score ranges from 0 to 8 and represents the sum of all items.

^cAn 11-item composite measure of frequency of participation in cognitively stimulating activities at age 6 (3 items, e.g., how often did someone read to you when you were 6?) and at age 12 (8 items, e.g., how often did you visit a library?).

^dParticipants are asked if they received no, <5 years, or 5–18 years of foreign language instruction.

^eParticipants were asked to select 1 of 10 levels of total family income when they were 40 years old using the “show-card” method. Ranges per level started from $0 to $4,999 (score of 1) and went up to $75,000 (score of 10) and over. A mean score of 6.1 indicates that mean income ranged between $25,000 and $29,999.

^fTotal number of availabilities of 7 items (e.g., subscription to a daily newspaper, possession of library card, etc) that support cognitive ability and number of books in the home (responses are adjusted to range from 0 to 1). Overall score ranges from 0 to 8 and represents the sum of all items.

^gA 9-item 5-point scale of frequency of participation of cognitively stimulating activities (e.g., visiting a museum, visiting a library, playing games). The sum is averaged and the measure ranges from 1 to 5.

^hParticipants were asked to select one of the levels of total family income using the “show-card” method. Ranges per level started from $0 to $4,999 (score of 1) and went up to $75,000 (score of 10) and over. A mean score of 7.1 indicates that mean income ranged between $30,000 and $34,999.

ⁱA 9-item 5-point scale of frequency of participation of cognitively stimulating activities (e.g., visiting a museum, visiting a library, playing games). The sum is averaged and the measure ranges from 1 to 5.

^jLifetime enrichment is the z-scored mean of early-, mid-, and late-life enrichment. Early-life enrichment is the z-scored mean of early-life SES, availability of cognitive resources at age 12, frequency of participation in cognitively stimulating activities at age 6 and 12, and early-life foreign language instuction. Mid-life enrichment is the z-scored mean of income at age 40 and cognitive resources at age 40. Late-life enrichment is the z-scored mean of income at study enrollment and cognitive activity frequency at enrollment.

Indicators within each life stage were moderately and significantly correlated (eTables 1–3). Early-life enrichment was moderately correlated with mid- (r = 0.51, p < 0.001) and late-(r = 0.43, p < 0.001) life enrichment; mid- and late-life enrichment were also moderately correlated (r = 0.56, p < 0.001). A lifetime enrichment composite was created by averaging the mean of the 3 composites.

Lifetime enrichment was moderately correlated with education (r = 0.56, p < 0.001) and global cognition (r = 0.43, p < 0.001), but not age (r = −0.03, p = 0.31).

Cognitive Enrichment and Incident AD Dementia and MCI

During follow-up, 551 persons developed AD dementia. Every 1-point higher in lifetime enrichment was associated with 38% lower dementia risk (hazard ratio [HR] 0.62, 95% CI 0.52–0.73, p < 0.001). eFigure 2 shows cumulative hazards by enrichment percentile.

Early-, mid-, and late-life cognitive enrichment were each associated with 20% (HR 0.80, 95% CI 0.70–0.93, p = 0.002), 21% (HR 0.79, 95% CI 0.70–0.90, p < 0.001), and 29% (HR 0.71, 95% CI 0.63–0.79, p < 0.001) lower risk of AD dementia, respectively, supporting a life-course model of risk reduction.

Higher level of lifetime cognitive enrichment was associated with later AD dementia onset (estimate = 0.035, 95% CI 0.021–0.046, p < 0.001). Individuals in the 90th (vs 10th) enrichment percentile developed dementia at a mean age of 93.8 years (vs 88.4 years), with a mean difference of 5.4 years (Figure 1).

Figure 1.

AD-Free Survival Probabilities Associated With Different Levels of Lifetime Cognitive Enrichment

Relation of low, medium, and high lifetime cognitive enrichment to probability of remaining dementia-free from AD, from an accelerated failure time model. AD = Alzheimer disease.

Mean onset was delayed by 2.9, 3.5, and 5.5 years for individuals in the 90th vs 10th percentile of early-, mid-, and late-life enrichment.

After excluding 495 participants because of prevalent MCI at baseline, 1,444 participants with similar demographic characteristics remained for incident MCI analyses (eTable 4). Of those, 719 developed MCI over follow-up.

The hazard of developing MCI was reduced by 36% for every 1-point increased in lifetime enrichment (HR 0.67, 95% CI 0.58–0.78, p < 0.001). Early-, mid-, and late-life enrichment were associated with 17% (HR 0.83, 95% CI 0.74–0.94, p = 0.003), 20% (HR 0.80, 95% CI 0.72–0.89, p < 0.001), and 24% (HR 0.76, 95% CI 0.69–0.84, p < 0.001) lower hazards of incident MCI, respectively.

More lifetime enrichment was associated with later MCI onset (estimate = 0.063, 95% CI 0.024–0.103, p = 0.002). Individuals in the 90th (vs 10th) enrichment percentile developed MCI at a mean age of 84.5 years (vs 77.5 years), with a mean difference of 7 years. Mean onset was delayed by 4.8, 4.1, and 7.8 years for individuals in the 90th enrichment percentiles during early-, mid-, and late-life, respectively (Table 2).

Table 2.

Odds of Incident AD Dementia and Mild Cognitive Impairment Associated With Lifetime Cognitive Enrichment, and With Cognitive Enrichment During Each Life Period

	Model term	Model A	Model B	Model C	Model D
		Estimate (SE), p value	Estimate (SE), p value	Estimate (SE), p value	Estimate (SE), p value
Incident AD dementia (N = 551)	Lifetime enrichment	0.62 (0.52–0.73), <0.001	—	—	—
	Early-life enrichment	—	0.80 (0.70–0.93), 0.002	—	—
	Mid-life enrichment	—	—	0.79 (0.70–0.90), 0.002	—
	Late-life enrichment	—	—	—	0.71 (0.63–0.79), <0.001
Incident MCI (N = 495)	Lifetime enrichment	0.67 (0.58–0.78), <0.001	—	—	—
	Early-life enrichment	—	0.83 (0.74–0.94), 0.003	—	—
	Mid-life enrichment	—	—	0.80 (0.72–0.89) <0.001	—
	Late-life enrichment	—	—	—	0.76 (0.69–0.84), <0.001

Abbreviations: AD = Alzheimer disease; MCI = mild cognitive impairment.

All models are adjusted for age at baseline, sex, and education.

Cognitive Enrichment, Level of Cognition, and Rate of Cognitive Decline

Higher lifetime cognitive enrichment was associated with better cognition at baseline (estimate = 0.31, SE = 0.02, p < 0.001) and a slower rate of decline (estimate = 0.02, SE = 0.01, p = 0.002) (Table 3). Relative to a person in the 50th percentile of enrichment, decline was 14% faster in the 10th percentile and 10% slower in the 90th percentile (Figure 2).

Table 3.

Association of Lifetime Cognitive Enrichment and Enrichment During Each Life Period with Global Cognitive Level and Slope (N = 1,939)

Model term	Model A	Model B	Model C	Model D
	Estimate (SE), p value	Estimate (SE), p value	Estimate (SE), p value	Estimate (SE), p value
Time (from baseline)	−0.09 (0.003), <0.001	−0.094 (0.003), <0.001	−0.093 (0.003), <0.001	−0.092 (0.003), <0.001
Lifetime enrichment	0.313 (0.02), <0.001	—	—	—
Time × lifetime enrichment	0.015 (0.005), 0.002	—	—	—
Early-life enrichment	—	0.178 (0.018), <0.001	—	—
Time × early-life enrichment	—	0.001 (0.004), 0.153	—	—
Mid-life enrichment	—	—	0.140 (0.016), <0.001	—
Time × mid-life enrichment	—	—	0.009 (0.004), 0.016	—
Late-life enrichment	—	—	—	0.229 (0.015), <0.001
Time × late-life enrichment	—	—	—	0.011 (0.004), 0.002

Each column is a linear-mixed effects model that controlled for age at baseline, sex, and education.

Figure 2.

Association of Lifetime Cognitive Enrichment to Later-Life Cognitive Decline From Baseline

Predicted paths of decline in global cognition from baseline up to 14 years of follow-up for individuals with high (90th percentile, green line), middle (50th percentile, blue line), and low (10th percentile, red line) lifetime cognitive enrichment, adjusted for age at death, sex, and education.

Higher scores on early, mid-, and late-life enrichment were associated with higher level of cognitive function at baseline; however, longitudinally, only mid- and late-life cognitive enrichment were associated with a slower rate of cognitive decline.

Lifetime enrichment was associated with a slower rate of cognitive decline across all 5 cognitive domains (eTable 5). Although enrichment during each period of life was associated with higher cognitive level across domains, mid-life enrichment was further associated with a slower rate of decline in working memory and visuospatial orientation, and late-life enrichment was associated with a slower rate of decline across all domains (eTable 6).

Cognitive Enrichment and ADRD Pathologies

The postmortem subset (N = 948) had similar demographics to the full sample (eTable 7).

Higher lifetime enrichment (odds ratio [OR] 0.71, 95% CI 0.55–0.93, p = 0.01) and separately late-life enrichment (OR 0.78, 95% CI 0.65–0.94, p = 0.01) were associated with fewer gross chronic infarcts.

Higher early-life enrichment was negatively associated with amyloid-β (estimate = −0.146, SE = 0.07, p = 0.044), PHF-tau (estimate = −0.174, SE = 0.08, p = 0.023), and overall AD pathology (estimate = −0.083, SE = 0.03, p = 0.013) reflecting our previous work.¹³

We did not find any associations for mid-life enrichment.

These associations are of marginal significance given the multiple comparisons with multiple pathologic indices across different life periods, although we note that many of the pathologies are highly correlated as are the 3 enrichment epochs.

Cognitive Enrichment and Cognitive Resilience

Every 1-point higher on the lifetime enrichment was associated higher global cognition (estimate = 0.32, SE = 0.06, p < 0.001) and slower rate of decline (estimate = 0.014, SE = 0.01, p = 0.02) proximate of death, independent of neuropathology (Table 4). Relative to lifetime enrichment at the 50th percentile, cognition declined 20% faster at the 10th percentile and 16% slower at the 90th percentile (Figure 3).

Table 4.

Association of Lifetime Cognitive Enrichment and Cognitive Enrichment During Each Life Period With Global Cognitive Level and Slope After Adjustment for 9 ADRD Neuropathologic Indices (N = 948)

Model term	Model A	Model B	Model C	Model D
	Estimate (SE), p value	Estimate (SE), p value	Estimate (SE), p value	Estimate (SE), p value
Time (before death)	−0.016 (0.008), 0.05	−0.016 (0.008), 0.04	−0.015 (0.008), 0.06	−0.016 (0.008), 0.05
Lifetime enrichment	0.316 (0.06), <0.001	—	—	—
Time before death × lifetime enrichment	0.014 (0.006), 0.02	—	—	—
Early-life enrichment	—	0.14 (0.05), 0.002	—	—
Time × early-life enrichment	—	0.005 (0.005), 0.300	—	—
Mid-life enrichment	—	—	0.184 (0.043), <0.001	—
Time × mid-life enrichment	—	—	0.012 (0.004), 0.006	—
Late-life enrichment	—	—	—	0.201 (0.04), <0.001
Time × late-life enrichment	—	—	—	0.006 (0.004), 0.183

Abbreviations: AD = Alzheimer disease; ADRD = AD and related dementia.

Each column is a separate mixed-effects models that controlled for age at death, sex, education, and postmortem measures of a summary measure of AD pathology, hippocampal sclerosis, TAR DNA-binding protein 43, Lewy bodies, macroinfarcts, microinfarcts, cerebral angiopathy, atherosclerosis, and arteriolosclerosis.

Figure 3.

Association of Lifetime Cognitive Enrichment to Later-Life Cognitive Decline Before Death

Predicted paths of decline in global cognition during the last 14 years of life for individuals with high (90th percentile, green line), middle (50th percentile, blue line), and low (10th percentile, red line) lifetime cognitive enrichment, adjusted for age at death, sex, education, and 9 common ADRD pathologic indices. AD = Alzheimer disease; ADRD = AD and related dementia.

Early- (estimate = 0.14, SE = 0.05, p = 0.002), mid- (estimate = 0.18, SE = 0.04, p < 0.001), and late- (estimate = 0.20 SE = 0.04, p < 0.001) life enrichment were associated with higher cognitive function proximate to death independent of neuropathology. Mid-life enrichment was further associated with a slower rate of decline (estimate = 0.012, SE = 0.004, p = 0.006) (Table 4, eFigure 3).

Lifetime enrichment was positively associated with visuospatial orientation and perceptual speed (eTable 8). Enrichment in each life-epoch was positively associated with cognitive function proximate to death. Midlife enrichment was further associated with a slower rate of decline in visuospatial orientation and marginally with working memory (eTable 9).

Secondary Analyses

Each indicator was positively associated with late-life cognition (eFigure 4, panels A, C, E), and except for availability of childhood resources, these associations remained significant in fully adjusted models. Foreign language instruction (<5 years vs no instruction: estimate = 0.23, SE = 0.03, p < 0.001; ≥5 years estimate = 0.28, SE = 0.04, p < 0.001 vs no instruction) and cognitive activity participation (midlife: estimate = 0.15, SE = 0.02, p < 0.001; late-life estimate = 0.27, SE = 0.02, p < 0.001) had the largest effects, which persisted after adjustment for SES. Importantly, SES also demonstrated modest but consistent associations with late-life cognition over and above behavioral indicators (childhood: estimate = 0.08, SE = 0.02, p < 0.001; midlife: estimate = 0.01, SE = 0.01, p = 0.02, late-life: estimate = 0.04, SE = 0.01, p < 0.001), suggesting that socioeconomic opportunity contributed independently to preserved cognitive function in late-life. Midlife activity participation (estimate = 0.05, SE = 0.02, p = 0.02) was further associated with slower decline (eFigure 4, panel C). After adjusting for neuropathology, SES was no longer significant, while mid- (estimate = 0.01, SE = 0.01, p = 0.03) and late-life cognitive activities (estimate = 0.21, SE = 0.05, p < 0.001) continued to be associated with slower decline (eFigure 4, panels B, D, F).

Discussion

In this community-based prospective cohort study of almost 2,000 older adults without dementia at baseline, higher lifetime cognitive enrichment was associated with 38% lower risk of AD dementia, 36% lower risk of MCI, delayed onset of AD dementia and MCI by up to 5–7 years, higher later-life cognitive function, and slower cognitive decline. In postmortem analyses, individuals with higher lifetime enrichment retained higher cognitive function and slower decline proximate to death, independent of neuropathology. Individuals in the topmost decile of lifetime enrichment experienced over 43% less decline than individuals in the lowest decile, equivalent to being about 6 years younger. These results suggest that lifelong cognitive enrichment may delay the onset of AD dementia and protect cognition in the presence of neuropathology. Our results suggests that cognitive health in late life is in part the product of lifetime exposure to cognitive enrichment.

Our and others’ previous studies^{6,9,10,12,18,27–31} examined single dimensions of enrichment, such as frequency of cognitive activities rather than the broader lived-in environment.³² Some previous work has also controlled for early life factors, such as childhood SES, which fail to account for the full extent of life enrichment across the life course.³² Previous work has also tried to tease out specific activities that might be beneficial, such as doing crossword puzzles^33,34 or playing board games³⁵; however, most often, enriching experiences tend to cluster together within a combination of factors and a broader lifestyle that cannot be easily isolated. Previous work has often,³⁶ but not always,³⁴ reported associations of cognitive engagement with level, but not with slope of cognition, in part due to smaller sample sizes, high attrition,³ limited follow-ups (≤3),³⁷ and brief measures that do not capture the lifetime broader home environment.

By creating a lifelong enrichment composite that broadly captures one’s lived-in environment, we extend the engagement hypothesis which only focuses on activity engagement only as a means to maintain cognition rather than protecting against its decline.³⁸ Consistent with previous work,^3,30,37 we observed that cognitive enrichment at each life period was more strongly associated with cognitive level than with rate of decline. This pattern suggests that the protective effects of enrichment may be driven by maintaining higher cognitive starting points and more modest contributions to the rate of cognitive decline. Although the effects were small, our results also showed that higher lifetime enrichment has significant protective effects on the rate of cognitive decline.

Our findings indicate that cognitive enrichment is not simply a proxy for socioeconomic advantage. Although SES and access to resources showed modest independent associations with late-life cognition, the enrichment composites also captured sustained behavioral engagement in intellectual activities across the life-course beyond SES effects. Importantly, indicators of cognitive activity engagement remained associated with slower cognitive decline even after adjusting for 9 ADRD neuropathologic indices, reinforcing that enrichment effects persist independent of socioeconomic context and neuropathologic burden supporting the cognitive reserve hypothesis.³⁹ Although we previously linked activity participation to cognitive resilience,⁹ the association with broader environmental enrichment is novel. Early-life enrichment was directly associated with AD pathology, supporting our,¹³ and others⁴⁰ previous work that brain integrity in old age partly reflects early environmental exposures. The persistence of mid-life effects even after adjustment for neuropathologic indices highlights midlife as a potentially sensitive period for intervention.

Enrichment paradigms have shown significant improvements in spatial learning capacity and in morphological and molecular brain alterations, including increased nerve growth factor, brain-derived neurotrophic factor, and neurotrophin-3, demonstrating significant alteration in protein levels across several brain regions in rodents housed in enriched environments relative to rodents housed in standard cages.^41,42 Exposure to enriched environments have also been shown to mitigate cognitive deficits in mice with amyloid-beta over-production and amyloid deposits.⁴³ Clinicopathologic research in the human brain is needed to fully characterize the molecular mechanisms related to these exposures.

The impact of life-long cognitive enrichment cannot be overestimated. We show here that life-long immersion in intellectual stimulation could reduce AD dementia incidence by almost 40%. Although our effect sizes were stronger for enrichment in mid- and late-life, substantial benefits may be reaped by optimizing enrichment in early life which may put an individual on a life-course trajectory that would be more beneficial in terms of later-life ADRD outcomes than any short-term gain produced by a later-life intervention. Reduction in the epidemic of dementia will come from public investments that expand access to cognitively enriching environments such as libraries, books, and extracurricular activities. Intervention programs, such as HeadStart,⁴⁴ and Experience Corps⁴⁵ that target socially disadvantaged youth and older adults, might not only enrich environments potentially during sensitive periods but also foster life-long learning, and mitigate ADRD risk.

Out study has strengths and limitations. Retrospective reports of enrichment introduce potential recall bias and reliability; however, our scales demonstrate sound psychometric properties.⁸ Cognitive activity measures exhibit high internal consistency (Cronbach α = 0.88), high test-retest correlation (r = 0.79),⁸ and construct validity as indicated by moderate positive correlations with participants’ educational attainment, parental education, and cognitive function.⁸ Construct validity for the cognitive resources¹² and for the early-life SES indicators⁴⁶ has also been documented, with each showing expected positive associations with educational attainment and cognitive performance. Future studies will need to investigate data with the inclusion of other indicators for example, occupational complexity. Participants are mostly nonHispanic White of European descent and highly educated; thus, results may carry potential ascertainment bias; however, our research supports previous work linking higher cognitive activity engagement to better ADRD outcomes in later life.^3,28,37 and underscore the need to be replicated in more diverse cohorts.

Glossary

AD

Alzheimer disease

ADRD

AD and related dementia

LME

linear mixed-effect

MAP

Memory and Aging Project

MCI

mild cognitive impairment

OR

odds ratio

SES

socioeconomic status