Skip to main content
NCBI home page
JAMA Network logoLink to JAMA Network
. 2023 Aug 4;6(8):e2327421. doi: 10.1001/jamanetworkopen.2023.27421

Association of Childhood and Midlife Neighborhood Socioeconomic Position With Cognitive Decline

Anna M Kucharska-Newton 1,2,✉, James Russell Pike 3, Jinyu Chen 3, Josef Coresh 3, A Richey Sharret 3, Thomas Mosley 4,5, Priya Palta 6
PMCID: PMC10403777  PMID: 37540511

Key Points

Question

Is neighborhood socioeconomic position (nSEP) associated with cognitive decline?

Findings

In this longitudinal cohort study of 5711 men and women, the rate of cognitive decline from midlife to older adulthood decreased by 9.2% with each 1-SD increase in childhood nSEP. No associations of midlife nSEP with cognitive decline were observed.

Meaning

This study suggests that early childhood experiences are associated with cognitive trajectories in adulthood.

Abstract

Importance

Early-life socioeconomic adversity may be associated with poor cognitive health over the life course.

Objective

To examine the association of childhood and midlife neighborhood socioeconomic position (nSEP) with cognitive decline.

Design, Setting, and Participants

This cohort study included 5711 men and women enrolled in the community-based Atherosclerosis Risk in Communities (ARIC) Study with repeated cognitive data measured over a median 27.0 years (IQR, 26.0-27.9 years) (1990-2019). Statistical analysis was performed from December 2022 through March 2023.

Exposure

Residence addresses for ARIC Study cohort participants were obtained at midlife (1990-1993) and as recalled addresses at 10 years of age (childhood). A composite nSEP z score was created as a sum of z scores for US Census–based measures of median household income; median value of owner-occupied housing units; percentage of households receiving interest, dividend, or net rental income; percentage of adults with a high school degree; percentage of adults with a college degree; and percentage of adults in professional, managerial, or executive occupations. Childhood nSEP and midlife nSEP were modeled as continuous measures and discretized into tertiles.

Main Outcomes and Measures

A factor score for global cognition was derived from a battery of cognitive tests administered at 5 in-person visits from baseline to 2019. The rate of cognitive decline from 50 to 90 years of age was calculated by fitting mixed-effects linear regression models with age as the time scale and adjusted for race, sex, birth decade, educational level, and presence of the apolipoprotein E ε4 allele.

Results

Among 5711 ARIC Study participants (mean [SD] baseline age, 55.1 [4.7] years; 3372 women [59.0%]; and 1313 Black participants [23.0%]), the median rate of cognitive decline was −0.33 SDs (IQR, −0.49 to −0.20 SDs) per decade. In adjusted analyses, each 1-SD-higher childhood nSEP score was associated with a slower (β, −9.2%; 95% CI, −12.1% to −6.4%) rate of cognitive decline relative to the sample median. A comparable association was observed when comparing the highest tertile with the lowest tertile of childhood nSEP (β, −17.7%; 95% CI, −24.1% to −11.3%). Midlife nSEP was not associated with the rate of cognitive decline.

Conclusions and Relevance

In this cohort study of contextual factors associated with cognitive decline, childhood nSEP was inversely associated with trajectories of cognitive function throughout adulthood.

This cohort study uses data from the Atherosclerosis Risk in Communities Study to examine the association of childhood and midlife neighborhood socioeconomic position with cognitive decline.

Introduction

Early-life experiences are foundational to brain development and have strong implications for the trajectory of cognitive function throughout life.1 Such trajectories may be direct, through the effect of childhood life circumstances on brain development2 during sensitive cognitive developmental periods,3 or may be corollary to the accumulation of adverse vascular risk factors that begins in childhood.4 Early-life experiences are highly correlated with childhood socioeconomic position (SEP), which is often defined according to the educational level and/or occupation of the individual’s parents or as a contextual characteristic of the environment in which children are born and live.5 Beginning with early studies by Barker et al,6,7 which highlighted the importance of prenatal and early-childhood SEP to cardiovascular health in adulthood, there now exists extensive research documenting the risk of adverse cardiovascular outcomes in adulthood in association with childhood SEP that is not mitigated by adulthood SEP.8,9 Likewise, multiple studies provide evidence for a greater risk of low cognitive performance in adulthood and a subsequently greater likelihood of dementia at a younger age in association with low SEP in childhood.10,11 However, the few studies examining the association of childhood SEP with the rate of cognitive decline are conflicting.10,12,13,14 Moreover, most of those studies have operationalized childhood SEP exposure through the lens of parental educational level, occupation, and financial stability without consideration of the neighborhood context, which, when measured in adulthood, is known to be associated with cognitive function.15,16,17 In this investigation, we used life course SEP measures combined with repeated assessments of cognitive function available from the Atherosclerosis Risk in Communities (ARIC) Study cohort18 to examine the association of neighborhood SEP (nSEP), measured from childhood through midlife, with the rate of cognitive decline from midlife to older adulthood.

Methods

Study Population

Details describing the ARIC Study cohort have been published elsewhere.18 In brief, 15 792 men and women 45 to 64 years of age were recruited through population-based sampling conducted in 4 geographically distinct communities of Washington County, Maryland; Jackson, Mississippi; Forsyth County, North Carolina; and selected suburbs of Minneapolis, Minnesota. ARIC Study participants completed a clinic-based examination at visit 1 (1987-1989) and were reexamined at visit 2 (1990-1992), visit 3 (1993-1995), and visit 4 (1996-1998). The ARIC Neurocognitive Study (ARIC-NCS) was initiated during visit 5 (2011-2013) and continued during visit 6 (2016-2017) and visit 7 (2018-2019). The protocol for each visit was approved by the institutional review board of each participating field center (University of North Carolina at Chapel Hill, University of Mississippi Medical Center, Johns Hopkins University, University of Minnesota, and Wake Forest University), and oral informed consent was obtained at each visit. This cohort study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.19

Exposure

Residence addresses for ARIC Study participants were obtained at visit 2 (1990-1993; midlife) and as recalled address in childhood (at 10 years of age) as part of the ARIC Lifecourse Socioeconomic Status Study conducted from 2002 to 2004.20 Geocoded addresses were used to place participants within geographically defined neighborhoods. A composite nSEP z score was calculated as a sum of z scores for US Census–based measures of (1) median household income; (2) median value of owner-occupied housing units; (3) percentage of households receiving interest, dividend, or net rental income; (4) percentage of adults with a high school degree; (5) percentage of adults with a college degree; and (6) percentage of adults in professional, managerial, or executive occupations.21 The childhood nSEP z score was derived from the 1930-1950 US Census data, which is available at only the county level. Midlife nSEP was based on US Census data available at the Census-tract level. Childhood nSEP and midlife nSEP were examined as continuous variables winsorized at the 1st and 99th percentiles and as tertiles of the overall distribution.

Outcome

An in-person battery of neuropsychological tests was conducted in a quiet room by trained examiners using standardized protocols at visits 2, 4, 5, 6, and 7. During visits 2 and 4, the battery included the Word Fluency Test,22 Digit Symbol Substitution Test,23 and Delayed Word Recall Test.24 Beginning at visit 5, the battery additionally incorporated the Digit Span Backward Test,23 Boston Naming Test,25 Animal Naming Test,22 Trail Making Tests A and B,26 Incidental Learning Test,27 and Logical Memory Test.23 The tests available at each visit were used to compute a global cognition factor score at each visit.28

Covariates

Sex, race, date of birth, and educational attainment were self-reported at visit 1. Date of birth was used to calculate the age in years at each visit. A categorical classification of race and center (hereafter referred to as race-center) was specified to account for the design of ARIC as well as possible systematic differences in health care resources and access associated with race and geography. Blood samples obtained during clinic visits were analyzed for apolipoprotein E (APOE) ε4 alleles using the TaqMan assay (Applied Biosystems).29 Additional information regarding study participants’ characteristics was ascertained at visit 2 and included high-density and low-density lipoprotein cholesterol level, body mass index, systolic blood pressure, hypertension, diabetes, multimorbidity, and physical activity (eMethods in Supplement 1).

Statistical Analysis

Statistical analysis was performed from December 2022 through March 2023. The analytic sample (eFigure in Supplement 1) was restricted to participants enrolled in the ARIC-NCS (2011-2013; n = 6538), excluding those missing childhood nSEP (n = 795); those participants of race other than Black or White and, due to small sample sizes, Black participants in Maryland and Minnesota (n = 22); and those participants with an unknown level of educational attainment (n = 10). The baseline was defined as the first neuropsychological evaluation conducted at visit 2 (1990-1992). The median follow-up time was 27.0 years (IQR, 26.0-27.9 years) across a mean (SD) of 4.6 (0.7) neuropsychological evaluations.

To mitigate bias caused by informative attrition, multivariate imputation by chained equations with auxiliary variables was used to impute predeath cognitive factor scores among participants who did not complete a neuropsychological evaluation at a given visit (eMethods in Supplement 1).30 Ten imputed data sets were generated, exceeding the 6 imputations suggested by a quadratic formula.31 Imputed data sets were analyzed in a 2-stage process.

During the first stage, a linear mixed-effects model (LMM) estimated the rate of cognitive decline for each participant. The LMM specified time from 50 years of age as the time scale, used the maximum likelihood estimation, used an unstructured variance-covariance matrix, and incorporated a random intercept and slope. Sex and race-center were included as fixed effects, and an interaction with time was specified. The median rate of decline in the analytic sample was −0.33 SDs (IQR, −0.49 to −0.20 SDs) per decade.

In the second stage, a percentage difference was calculated as the rate of cognitive decline per participant divided by the median rate of decline in the analytic sample. Negative percentages represent slower cognitive decline, while positive percentages denote more rapid decline. Linear mixed-effects models with a random intercept at the neighborhood level were used to estimate the association between nSEP (continuous measures and discretized as tertiles) and the percentage difference from the median rate. Participant-specific rates of cognitive decline were also discretized into quintiles and analyzed in multinomial logistic mixed-effects models (MLMMs) that estimated the odds of belonging to a specific quintile relative to the third quintile. Specifically, we estimated the odds of having the least cognitive decline (fifth quintile) relative to the median (third quintile) or the most cognitive decline (first quintile) relative to the median (third quintile).

The childhood nSEP and midlife nSEP were examined in 2 models. Model 1 adjusted for sex and race-center. Model 2 additionally adjusted for birth decade, educational level, and presence of APOE ε4 alleles. All person-level covariates were group mean centered at the neighborhood level.32 The neighborhood-level mean values were reintroduced to accurately adjust for within- and between-neighborhood effects.33 An interaction between time and each covariate was specified. Parameter estimates from the analysis of imputed data were combined according to Rubin’s rules.34

Exploratory analyses examined moderation and effect modification by race, birth decade, sex, and multimorbidity in fully adjusted models. Two-sided P values for additive interactions in LMMs and multiplicative interactions in MLMMs were calculated by specifying a cross-level interaction between nSEP and a person-level covariate. P values for additive interactions in MLMMs were computed by calculating the relative excess risk due to interaction.35 Statistical significance was defined as P < .05. Point estimates and 95% CIs from stratified models were used to evaluate effect modification. Race-stratified restricted cubic spline models with knots placed at the 5th, 27.5th, 50th, 72.5th, and 95th percentiles assessed how the nonnormal distribution of nSEP altered the association with cognitive decline. All analyses were executed in SAS, version 9.4 (SAS Institute Inc), with the exception of the multiple imputation, which was performed in Stata, version 14.0 (StataCorp LLC).

Results

At baseline, the mean (SD) age of the 5711 participants was 55.1 (4.7) years, 3372 participants (59.0%) were women, 2339 participants (41.0%) were men, and 1313 participants (23.0%) were Black (Table). The mean (SD) of the nSEP z score was 0.5 (4.0) at childhood and 1.1 (4.8) at midlife. Compared with participants in the lowest tertile of childhood nSEP, participants in the intermediate or highest tertiles were more likely to be younger and White, had more years of formal education and higher cognitive scores, and exhibited lower cardiometabolic risk. A similar, albeit smaller, gradient of participant characteristics was observed across the tertiles of midlife nSEP (eTable 1 in Supplement 1).

Table. Baseline Characteristics of the Study Population by Tertiles of nSEP During Childhood: The Atherosclerosis Risk in Communities Study, 1990-2019a.

Characteristic Participants, No. (%) P value for trendb
All (N = 5711) Tertile of childhood nSEP, z score range
Low, −9.39 to −1.95 (n = 1906) Intermediate, −1.93 to 2.16 (n = 1901) High, 2.23 to 20.25 (n = 1904)
Age, mean (SD), y 55.1 (4.7) 58.1 (4.2) 54.3 (4.4) 52.8 (3.8) <.001
Decade of birth
1920-1929 687 (12.0) 424 (22.2) 182 (9.6) 81 (4.3) <.001
1930-1939 3355 (58.7) 1293 (67.8) 1087 (57.2) 975 (51.2)
1940-1949 1669 (29.2) 189 (9.9) 632 (33.2) 848 (44.5)
Sex
Female 3372 (59.0) 1118 (58.7) 1139 (59.9) 1115 (58.6) .95
Male 2339 (41.0) 788 (41.3) 762 (40.1) 789 (414)
Black race 1313 (23.0) 581 (30.5) 325 (17.1) 407 (21.4) <.001
Race and center
White, Forsyth County, NC 1169 (20.5) 488 (25.6) 279 (14.7) 402 (21.1) <.001
Black, Forsyth County, NC 84 (1.5) 33 (1.7) 14 (0.7) 37 (1.9)
White, Minneapolis, MN 1699 (29.7) 205 (10.8) 557 (29.3) 937 (49.2)
White, Washington County, MD 1530 (26.8) 632 (33.2) 740 (38.9) 158 (8.3)
Black, Jackson, MS 1229 (21.5) 548 (28.8) 311 (16.4) 370 (19.4)
Educational level
Less than high school 827 (14.5) 436 (22.9) 236 (12.4) 155 (8.1) <.001
High school, GED certification, or vocational school 2421 (42.4) 828 (43.4) 837 (44.0) 756 (39.7)
College, graduate, or professional school 2463 (43.1) 642 (33.7) 828 (43.6) 993 (52.2)
≥1 APOE ε4 allele, No./total No. (%) 1594/5527 (28.8) 512/1855 (27.6) 558/1826 (30.6) 524/1846 (28.4) .60
Systolic blood pressure
No. 5679 1893 1887 1899 NA
Mean (SD), mm Hg 117.5 (16.6) 119.7 (17.1) 116.3 (16.0) 116.3 (16.4) <.001
Hypertension 1783 (31.2) 715 (37.5) 533 (28.0) 535 (28.1) <.001
BMI
No. 5677 1893 1886 1898 NA
Mean (SD) 27.7 (5.1) 27.8 (5.2) 27.7 (5.0) 27.6 (5.2) .34
Diabetes, No/total No. (%) 497/5660 (8.8) 185/1885 (9.8) 160/1882 (8.5) 152/1893 (8.0) .05
Sports participation
No. 5708 1904 1901 1903 NA
Mean (SD), min/wk 668.1 (807.4) 598.3 (766.2) 675.0 (814.3) 730.9 (835.1) <.001
HDL cholesterol
No. 5644 1877 1879 1888 NA
Mean (SD), mg/dL 50.9 (16.7) 50.2 (15.9) 50.8 (17.3) 51.7 (17.0) .02
LDL cholesterol
No. 5576 1859 1855 1862 NA
Mean (SD), mg/dL 130.8 (35.0) 132.9 (33.8) 131.5 (35.8) 127.9 (35.1) <.001
Multimorbidity 473 (8.3) 205 (10.8) 129 (6.8) 139 (7.3) <.001
Childhood nSEP
No. 5711 1906 1901 1904 NA
Mean (SD) 0.5 (4.0) −3.7 (1.6) 0.4 (1.2) 4.9 (2.8) <.001
Midlife nSEP
No. 5538 1802 1861 1875 NA
Mean (SD) 1.1 (4.8) −0.2 (4.7) 1.1 (4.6) 2.2 (4.8) <.001
Global cognition, mean (SD)
No. 5688 1897 1893 1898 NA
Mean (SD) 0.0 (1.0) −0.3 (1.0) 0.1 (0.9) 0.2 (1.0) <.001
Death by 2019 1326 (23.2) 574 (30.1) 414 (21.8) 338 (17.8) <.001
Open in a new tab

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); GED, General Educational Development; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NA, not applicable; nSEP, neighborhood socioeconomic position.

SI conversion factors: To convert HDL cholesterol and LDL cholesterol to millimoles per liter, multiply by 0.0259.

a

Study baseline (1990-1992) defined as first year in which participant in Atherosclerosis Risk in Communities Study completed neuropsychological assessment.

b

Univariate baseline differences in study variables were assessed using linear regression, Cochran-Armitage trend tests, and Cochran-Mantel-Haenszel trend tests as appropriate.

Rate of Cognitive Decline

The median rate of cognitive decline among the cohort was −0.33 SDs (IQR, −0.49 to −0.20 SDs) per decade. In models adjusted for sex and race-center, a 1-SD-higher nSEP during childhood was associated with a slower (β, −22.6%; 95% CI, −25.4% to −19.9%) rate of cognitive decline relative to the sample median. After additionally adjusting for birth decade, educational level, and the presence of APOE ε4 alleles (Figure 1), the association was attenuated but remained statistically significant (β, −9.2%; 95% CI, −12.1% to −6.4%). A comparable association was observed when comparing the highest tertile with the lowest tertile of childhood nSEP (β, −17.7%; 95% CI, −24.1% to −11.3%). In contrast, the association of midlife nSEP was considerably smaller and not statistically significant in fully adjusted models. This pattern persisted when both exposures were included in the same model (eTable 2 in Supplement 1) and was further supported by restricted cubic spline models that depict a largely linear association (Figure 2).

Figure 1. Estimates of Percentage Difference From Median Rate of Cognitive Decline Between 50 and 90 Years of Age by Neighborhood Socioeconomic Position (nSEP) During Childhood or Midlife: The Atherosclerosis Risk in Communities Study, 1990-2019 (N = 5711).

Figure 1.

Open in a new tab

A 2-stage process was used to estimate participant-specific cognitive decline between 50 and 90 years of age. Linear mixed-effects models with random intercepts at the neighborhood level estimated the association between nSEP and the percentage difference from the median rate of cognitive decline in the analytic sample. Separate models were fit for each measure of childhood and midlife nSEP. Each model was adjusted for sex, race and center, birth decade, education, and APOE ε4.

Figure 2. Histograms and Restricted Cubic Spline Model Estimates of the Percentage Difference From Median Rate of Cognitive Decline Between 50 and 90 Years of Age by Continuous Standardized Measures of Neighborhood Socioeconomic Position (nSEP) During Childhood or Midlife Stratified by Race: The Atherosclerosis Risk in Communities Study, 1990-2019 (N = 5711).

Figure 2.

Open in a new tab

A 2-stage process was used to estimate participant-specific cognitive decline between 50 and 90 years of age. Linear mixed-effects models with random intercepts at the neighborhood level and restricted cubic splines estimated the association between nSEP and the percentage difference from the median rate of cognitive decline in the analytic sample. Knots for the splines were placed at the 5th, 27.5th, 50th, 72.5th, and 95th percentiles. Separate models were fit for each measure of childhood and midlife nSEP. Effect modification was evaluated by stratifying the data set by race. The primary models were adjusted for sex, race and center, birth decade, education, and APOE ε4. Models fit to race-stratified data were adjusted for center instead of race and center. Histograms depict standardized nSEP during childhood or midlife. Solid lines indicate the point estimates, and dashed lines indicate the 95% CIs.

Quintiles of Cognitive Decline

When cognitive decline was discretized into quintiles (Figure 3; eTables 3-5 in Supplement 1), individuals in the top quintile experienced a mean (SD) cognitive decline of −0.10 (0.09) SDs per decade. A 1-SD-higher childhood nSEP was associated with higher odds (odds ratio, 1.21; 95% CI, 1.06-1.38) of belonging to the top (fifth) quintile (ie, the least amount of cognitive decline) and lower odds (odds ratio, 0.80; 95% CI, 0.69-0.92) of belonging to the bottom (first) quintile (ie, the greatest amount of cognitive decline) relative to the median (third) quintile. This finding was replicated when examining tertiles of nSEP and when examining childhood and midlife nSEP in the same model.

Figure 3. Estimates of the Odds of Belonging to a Specific Quintile of Cognitive Decline Between 50 and 90 Years of Age by Continuous Standardized Measures of Neighborhood Socioeconomic Position (nSEP) During Childhood or Midlife: The Atherosclerosis Risk in Communities Study, 1990-2019 (N = 5711).

Figure 3.

Open in a new tab

A 2-stage process was used to estimate participant-specific cognitive decline between 50 and 90 years of age. Multinomial logistic mixed-effects models with random intercepts at the neighborhood level estimated the association between nSEP and the odds of belonging to a specific quintile of cognitive decline. Separate models were fit for each measure of childhood and midlife nSEP. Each model was adjusted for sex, race and center, birth decade, education, and APOE ε4. aOR indicates covariate-adjusted odds ratio.

Exploratory Analyses

Exploratory analyses did not reveal consistent evidence of interactions (eTables 6 and 7 in Supplement 1) but did suggest that the protective association of higher childhood nSEP is greater among White participants who may also live in neighborhoods with a higher nSEP (childhood range, −9.3 to 20.2 SDs; midlife range, −8.6 to 14.8 SDs) than the maximum observed for Black participants (childhood range, −9.4 to 7.4 SDs; midlife range, −12.8 to 9.5 SDs).

Discussion

In this large, well-characterized cohort of community-dwelling adults, we found that higher nSEP in childhood at 10 years of age was associated with protection against cognitive decline between 50 and 90 years of age. Although membership in the high tertile compared with low tertile of the childhood nSEP was associated with a 17.7% slower rate of cognitive decline, the protective association of childhood nSEP was not limited to those within the narrow highest childhood nSEP category. The quintile analysis permitted us to assess whether there is an outlying group with exceptional characteristics within the more favorable childhood nSEP categories, such as individuals in the top quintile, for whom the mean (SD) cognitive decline was −0.10 (0.09) SDs per decade. On the contrary, the monotonic continuum of the odds ratios examining the association of childhood nSEP with membership in quintiles of the rate of cognitive decline suggests only a stepwise incremental benefit of improvements in childhood nSEP with cognitive health in adulthood.

Despite the importance of childhood as a foundational period for determining health outcomes over the life course and the recognized association of SEP with cognitive development, investigations of the association of childhood SEP at the neighborhood level with person-level cognitive decline are sparse. Data from several cohorts are inconclusive10,36,37,38,39; however, a cross-cohort examination of the Whitehall II study,40 the Health and Retirement Study,41 and the Kame Project42 provides evidence for the association of poor individual-level SEP during childhood, defined as low parental educational attainment and financial hardships, with a faster rate of cognitive decline.13 Our focus on neighborhood-level SEP adds an important contextual understanding of how the early-life socioeconomic environment, beyond individual-level socioeconomic factors, contributes to cognitive health over the life course. This potentially modifiable social factor can serve as a target for interventions aimed at delaying the onset of cognitive impairment and decreasing the population burden of dementia.

Our findings fit well within the life course model,43 which postulates the existence of pathways linking early life experiences44 with adult health. Our observations also conform to the conceptual framework of formative periods of intense brain development during childhood and through adolescence and young adulthood that are highly sensitive to the environment.45,46 Adverse childhood socioeconomic conditions can serve as a marker of deprivation in cognitive exposures,47 which may be associated with adverse changes in synaptic development and associated cognitive functions.48 Low childhood SEP has been found to be associated with low hippocampal49,50 and amygdala50 volumes and with white matter hyperintensities in adulthood,51 suggesting that early socioeconomic adversity may be associated with distinct biological brain development pathways52 that are associated with accelerated cognitive decline.53

An important factor that remains unclear is whether the association of childhood socioeconomic adversity with brain function in adulthood is direct54 or mediated by midlife SEP.38,39 Results of our analyses did not provide evidence for an association of midlife nSEP with cognitive decline, suggesting that childhood nSEP may have a direct association with cognitive decline. Furthermore, the finding that, in our study, the association of childhood nSEP with cognitive decline was robust to adjustment for individual educational attainment and midlife nSEP also bolsters this conclusion. However, we cannot rule out a more nuanced multilevel process in which individual-level SEP, but not neighborhood-level SEP, in midlife, beyond educational attainment, mediates the association between childhood nSEP and cognitive decline.55

Limitations

This study has some limitations. A key limitation was the lack of US Census information at the census-tract level during childhood, which may have obscured the observation of potential microgeographic variation in nSEP that may have attenuated the associations of childhood nSEP with cognitive decline relative to those observed for midlife nSEP. Informative attrition may or may not have introduced meaningful bias despite efforts to minimize bias using multiple imputation methods.56 Specifically, the observed median rate of cognitive decline was both slower57,58 and faster59 than the mean rate of decline reported in other studies, which may result in a misleading representation of the percentage difference because the magnitude is relative to the selected reference in our study population.

Conclusions

In this cohort study, the observed gradient in the association of childhood nSEP with the rate of cognitive decline has implications for health and social policy interventions that can benefit children during early developmental stages and promote the closing of the persistent socioeconomic gap in cognitive outcomes at critical developmental periods. Place-based interventions aimed at reducing neighborhood poverty and improving social cohesion provide promising examples.60

Supplement 1.

eMethods.

eTable 1. Baseline Characteristics of the Study Population by Tertiles of Neighborhood Socioeconomic Position During Midlife: The ARIC Study 1990-2019 (N=5538)

eTable 2. Combined Estimates of Percentage Difference From Median Rate of Cognitive Decline Between the Ages of 50 and 90 Years by Neighborhood Socioeconomic Position During Childhood and Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 3. Estimates of the Odds of Belonging to a Specific Quintile of Cognitive Decline Between the Ages of 50 and 90 Years by Tertiles of Neighborhood Socioeconomic Position During Childhood or Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 4. Combined Estimates of the Odds of Belonging to a Specific Quintile of Cognitive Decline Between the Ages of 50 and 90 Years by Continuous, Standardized Measures of Neighborhood Socioeconomic Position During Childhood and Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 5. Combined Estimates of the Odds of Belonging to a Specific Quintile of Cognitive Decline Between the Ages of 50 and 90 Years by Tertiles of Neighborhood Socioeconomic Position During Childhood and Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 6. Stratified Estimates of Percentage Difference From Median Rate of Cognitive Decline Between the Ages of 50 and 90 Years by Continuous, Standardized Measures of Neighborhood Socioeconomic Position During Childhood or Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 7. Stratified Estimates of the Odds of Belonging to a Specific Quintile of Cognitive Decline Between the Ages of 50 and 90 Years by Continuous, Standardized Measures of Neighborhood Socioeconomic Position During Midlife: The ARIC Study 1990-2019 (N=5711)

eFigure. Flowchart of Participants Selected for Analysis

eReferences.

Click here for additional data file. (242.5KB, pdf)
Supplement 2.

Data Sharing Statement

Click here for additional data file. (15.9KB, pdf)

References

  • 1.McCormick BJJ, Caulfield LE, Richard SA, et al. ; MAL-ED NETWORK INVESTIGATORS . Early life experiences and trajectories of cognitive development. Pediatrics. 2020;146(3):e20193660. doi: 10.1542/peds.2019-3660 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Bick J, Nelson CA. Early adverse experiences and the developing brain. Neuropsychopharmacology. 2016;41(1):177-196. doi: 10.1038/npp.2015.252 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Fox SE, Levitt P, Nelson CA III. How the timing and quality of early experiences influence the development of brain architecture. Child Dev. 2010;81(1):28-40. doi: 10.1111/j.1467-8624.2009.01380.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Hakala JO, Pahkala K, Juonala M, et al. Cardiovascular risk factor trajectories since childhood and cognitive performance in midlife: the Cardiovascular Risk in Young Finns Study. Circulation. 2021;143(20):1949-1961. doi: 10.1161/CIRCULATIONAHA.120.052358 [DOI] [PubMed] [Google Scholar]
  • 5.Walsh D, McCartney G, Smith M, Armour G. Relationship between childhood socioeconomic position and adverse childhood experiences (ACEs): a systematic review. J Epidemiol Community Health. 2019;73(12):1087-1093. doi: 10.1136/jech-2019-212738 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Barker DJ, Winter PD, Osmond C, Margetts B, Simmonds SJ. Weight in infancy and death from ischaemic heart disease. Lancet. 1989;2(8663):577-580. doi: 10.1016/S0140-6736(89)90710-1 [DOI] [PubMed] [Google Scholar]
  • 7.Barker DJ. Maternal nutrition, fetal nutrition, and disease in later life. Nutrition. 1997;13(9):807-813. doi: 10.1016/S0899-9007(97)00193-7 [DOI] [PubMed] [Google Scholar]
  • 8.Galobardes B, Smith GD, Lynch JW. Systematic review of the influence of childhood socioeconomic circumstances on risk for cardiovascular disease in adulthood. Ann Epidemiol. 2006;16(2):91-104. doi: 10.1016/j.annepidem.2005.06.053 [DOI] [PubMed] [Google Scholar]
  • 9.Zanetti D, Tikkanen E, Gustafsson S, Priest JR, Burgess S, Ingelsson E. Birthweight, type 2 diabetes mellitus, and cardiovascular disease: addressing the Barker hypothesis with mendelian randomization. Circ Genom Precis Med. 2018;11(6):e002054. doi: 10.1161/CIRCGEN.117.002054 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Faul JD, Ware EB, Kabeto MU, Fisher J, Langa KM. The effect of childhood socioeconomic position and social mobility on cognitive function and change among older adults: a comparison between the United States and England. J Gerontol B Psychol Sci Soc Sci. 2021;76(suppl 1):S51-S63. doi: 10.1093/geronb/gbaa138 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.George KM, Lutsey PL, Kucharska-Newton A, et al. Life-course individual and neighborhood socioeconomic status and risk of dementia in the Atherosclerosis Risk in Communities Neurocognitive Study. Am J Epidemiol. 2020;189(10):1134-1142. doi: 10.1093/aje/kwaa072 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Staff RT, Hogan MJ, Whalley LJ. The influence of childhood intelligence, social class, education and social mobility on memory and memory decline in late life. Age Ageing. 2018;47(6):847-852. doi: 10.1093/ageing/afy111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Tsang RSM, Gallacher JE, Bauermeister S. The long arm of childhood socioeconomic deprivation on mid- to later-life cognitive trajectories: a cross-cohort analysis. Alzheimers Dement (Amst). 2022;14(1):e12322. doi: 10.1002/dad2.12322 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Racine Maurice S, Hébert A, Turcotte V, Potvin O, Hudon C, Duchesne S. Childhood socioeconomic status does not predict late-life cognitive decline in the 1936 Lothian Birth Cohort. Front Psychol. 2021;12:679044. doi: 10.3389/fpsyg.2021.679044 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Pase MP, Rowsthorn E, Cavuoto MG, et al. Association of neighborhood-level socioeconomic measures with cognition and dementia risk in Australian adults. JAMA Netw Open. 2022;5(3):e224071. doi: 10.1001/jamanetworkopen.2022.4071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Shih RA, Ghosh-Dastidar B, Margolis KL, et al. Neighborhood socioeconomic status and cognitive function in women. Am J Public Health. 2011;101(9):1721-1728. doi: 10.2105/AJPH.2011.300169 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Mehdipanah R, Briceño EM, Heeringa SG, et al. Neighborhood SES and cognitive function among Hispanic/Latinx residents: why where you live matters. Am J Prev Med. 2022;63(4):574-581. doi: 10.1016/j.amepre.2022.04.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Wright JD, Folsom AR, Coresh J, et al. The ARIC (Atherosclerosis Risk In Communities) Study: JACC Focus Seminar 3/8. J Am Coll Cardiol. 2021;77(23):2939-2959. doi: 10.1016/j.jacc.2021.04.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies. Int J Surg. 2014;12(12):1495-1499. doi: 10.1016/j.ijsu.2014.07.013 [DOI] [PubMed] [Google Scholar]
  • 20.Carson AP, Rose KM, Catellier DJ, et al. Cumulative socioeconomic status across the life course and subclinical atherosclerosis. Ann Epidemiol. 2007;17(4):296-303. doi: 10.1016/j.annepidem.2006.07.009 [DOI] [PubMed] [Google Scholar]
  • 21.Borrell LN, Diez Roux AV, Rose K, Catellier D, Clark BL; Atherosclerosis Risk in Communities Study. Neighbourhood characteristics and mortality in the Atherosclerosis Risk in Communities Study. Int J Epidemiol. 2004;33(2):398-407. doi: 10.1093/ije/dyh063 [DOI] [PubMed] [Google Scholar]
  • 22.Benton AL, Hamsher KD, Sivan AB. Multilingual Aphasia Examination. AJA Associates; 1989. [Google Scholar]
  • 23.Wechsler D. Wechsler Memory Scale–Revised. Psychological Corporation; 1987. [Google Scholar]
  • 24.Knopman DS, Ryberg S. A verbal memory test with high predictive accuracy for dementia of the Alzheimer type. Arch Neurol. 1989;46(2):141-145. doi: 10.1001/archneur.1989.00520380041011 [DOI] [PubMed] [Google Scholar]
  • 25.Williams BW, Mack W, Henderson VW. Boston Naming Test in Alzheimer’s disease. Neuropsychologia. 1989;27(8):1073-1079. doi: 10.1016/0028-3932(89)90186-3 [DOI] [PubMed] [Google Scholar]
  • 26.Reitan RM. Validity of the Trail Making Test as an indicator of organic brain damage. Percept Mot Skills. 1958;8:271-276. doi: 10.2466/pms.1958.8.3.271 [DOI] [Google Scholar]
  • 27.Ryan J, Lopez S. Wechsler Adult Intelligence Scale–III. In: Dorfman W, Hersen M, eds. Perspectives on Individual Differences. Kluwer Academic/Plenum Publishers; 2001. [Google Scholar]
  • 28.Gross AL, Power MC, Albert MS, et al. Application of latent variable methods to the study of cognitive decline when tests change over time. Epidemiology. 2015;26(6):878-887. doi: 10.1097/EDE.0000000000000379 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Hsu CC, Kao WH, Coresh J, et al. Apolipoprotein E and progression of chronic kidney disease. JAMA. 2005;293(23):2892-2899. doi: 10.1001/jama.293.23.2892 [DOI] [PubMed] [Google Scholar]
  • 30.van Buuren S. Multiple imputation of discrete and continuous data by fully conditional specification. Stat Methods Med Res. 2007;16(3):219-242. doi: 10.1177/0962280206074463 [DOI] [PubMed] [Google Scholar]
  • 31.von Hippel PT. How many imputations do you need? a two-stage calculation using a quadratic rule. Sociol Methods Res. 2020;49(3):699-718. doi: 10.1177/0049124117747303 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Enders CK, Tofighi D. Centering predictor variables in cross-sectional multilevel models: a new look at an old issue. Psychol Methods. 2007;12(2):121-138. doi: 10.1037/1082-989X.12.2.121 [DOI] [PubMed] [Google Scholar]
  • 33.Yaremych HE, Preacher KJ, Hedeker D. Centering categorical predictors in multilevel models: best practices and interpretation. Psychol Methods. 2023;28(3):613-630. doi: 10.1037/met0000434 [DOI] [PubMed] [Google Scholar]
  • 34.Rubin DB. Multiple Imputation for Nonresponse in Surveys. John Wiley & Sons; 1987. doi: 10.1002/9780470316696 [DOI] [Google Scholar]
  • 35.VanderWeele TJ, Knol MJ. A tutorial on interaction. Epidemiol Method. 2014;3(1):33-72. doi: 10.1515/em-2013-0005 [DOI] [Google Scholar]
  • 36.Sha T, Yan Y, Cheng W. Associations of childhood socioeconomic status with mid-life and late-life cognition in Chinese middle-aged and older population based on a 5-year period cohort study. Int J Geriatr Psychiatry. 2018;33(10):1335-1345. doi: 10.1002/gps.4930 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Marden JR, Tchetgen Tchetgen EJ, Kawachi I, Glymour MM. Contribution of socioeconomic status at 3 life-course periods to late-life memory function and decline: early and late predictors of dementia risk. Am J Epidemiol. 2017;186(7):805-814. doi: 10.1093/aje/kwx155 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Zaninotto P, Batty GD, Allerhand M, Deary IJ. Cognitive function trajectories and their determinants in older people: 8 years of follow-up in the English Longitudinal Study of Ageing. J Epidemiol Community Health. 2018;72(8):685-694. doi: 10.1136/jech-2017-210116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Oi K, Haas S. Cardiometabolic risk and cognitive decline: the role of socioeconomic status in childhood and adulthood. J Health Soc Behav. 2019;60(3):326-343. doi: 10.1177/0022146519867924 [DOI] [PubMed] [Google Scholar]
  • 40.Marmot M, Brunner E. Cohort Profile: the Whitehall II study. Int J Epidemiol. 2005;34(2):251-256. doi: 10.1093/ije/dyh372 [DOI] [PubMed] [Google Scholar]
  • 41.Sonnega A, Faul JD, Ofstedal MB, Langa KM, Phillips JWR, Weir DR. Cohort Profile: the Health and Retirement Study (HRS). Int J Epidemiol. 2014;43(2):576-585. doi: 10.1093/ije/dyu067 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Borenstein AR, Wu Y, Bowen JD, et al. Incidence rates of dementia, Alzheimer disease, and vascular dementia in the Japanese American population in Seattle, WA: the Kame Project. Alzheimer Dis Assoc Disord. 2014;28(1):23-29. doi: 10.1097/WAD.0b013e3182a2e32f [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Ben-Shlomo Y, Kuh D. A life course approach to chronic disease epidemiology: conceptual models, empirical challenges and interdisciplinary perspectives. Int J Epidemiol. 2002;31(2):285-293. doi: 10.1093/ije/31.2.285 [DOI] [PubMed] [Google Scholar]
  • 44.Suglia SF, Appleton AA, Bleil ME, et al. Timing, duration, and differential susceptibility to early life adversities and cardiovascular disease risk across the lifespan: implications for future research. Prev Med. 2021;153:106736. doi: 10.1016/j.ypmed.2021.106736 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Thompson A, Steinbeis N. Sensitive periods in executive function development. Curr Opin Behav Sci. 2020;36:98-105. doi: 10.1016/j.cobeha.2020.08.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Nelson CA III, Gabard-Durnam LJ. Early adversity and critical periods: neurodevelopmental consequences of violating the expectable environment. Trends Neurosci. 2020;43(3):133-143. doi: 10.1016/j.tins.2020.01.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Bradley RH, Convyn RF, Burchinal M, McAdoo HP, Coll CG. The home environments of children in the United States, part II: relations with behavioral development through age thirteen. Child Dev. 2001;72(6):1868-1886. doi: 10.1111/1467-8624.t01-1-00383 [DOI] [PubMed] [Google Scholar]
  • 48.Sheridan MA, McLaughlin KA. Dimensions of early experience and neural development: deprivation and threat. Trends Cogn Sci. 2014;18(11):580-585. doi: 10.1016/j.tics.2014.09.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Decker AL, Duncan K, Finn AS, Mabbott DJ. Children’s family income is associated with cognitive function and volume of anterior not posterior hippocampus. Nat Commun. 2020;11(1):4040. doi: 10.1038/s41467-020-17854-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.McLachlan K, Zhou D, Little G, et al. Current socioeconomic status correlates with brain volumes in healthy children and adolescents but not in children with prenatal alcohol exposure. Front Hum Neurosci. 2020;14:223. doi: 10.3389/fnhum.2020.00223 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Murray AD, McNeil CJ, Salarirad S, Whalley LJ, Staff RT. Early life socioeconomic circumstance and late life brain hyperintensities—a population based cohort study. PLoS One. 2014;9(2):e88969. doi: 10.1371/journal.pone.0088969 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Biazoli CE Jr, Salum GA, Gadelha A, et al. Socioeconomic status in children is associated with spontaneous activity in right superior temporal gyrus. Brain Imaging Behav. 2020;14(4):961-970. doi: 10.1007/s11682-019-00073-z [DOI] [PubMed] [Google Scholar]
  • 53.Barnes LL, Wilson RS, Everson-Rose SA, Hayward MD, Evans DA, Mendes de Leon CF. Effects of early-life adversity on cognitive decline in older African Americans and Whites. Neurology. 2012;79(24):2321-2327. doi: 10.1212/WNL.0b013e318278b607 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Houweling TAJ, Oude Groeniger J, Jansen PW, et al. Trajectories of socioeconomic inequality in early child development: a cohort analysis. Int J Equity Health. 2022;21(1):79. doi: 10.1186/s12939-022-01675-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Preacher KJ, Zyphur MJ, Zhang Z. A general multilevel SEM framework for assessing multilevel mediation. Psychol Methods. 2010;15(3):209-233. doi: 10.1037/a0020141 [DOI] [PubMed] [Google Scholar]
  • 56.Euser SM, Schram MT, Hofman A, Westendorp RG, Breteler MM. Measuring cognitive function with age: the influence of selection by health and survival. Epidemiology. 2008;19(3):440-447. doi: 10.1097/EDE.0b013e31816a1d31 [DOI] [PubMed] [Google Scholar]
  • 57.Hayden KM, Reed BR, Manly JJ, et al. Cognitive decline in the elderly: an analysis of population heterogeneity. Age Ageing. 2011;40(6):684-689. doi: 10.1093/ageing/afr101 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Royall DR, Palmer R, Chiodo LK, Polk MJ. Normal rates of cognitive change in successful aging: the Freedom House Study. J Int Neuropsychol Soc. 2005;11(7):899-909. doi: 10.1017/S135561770505109X [DOI] [PubMed] [Google Scholar]
  • 59.Levine DA, Gross AL, Briceño EM, et al. Sex differences in cognitive decline among US adults. JAMA Netw Open. 2021;4(2):e210169. doi: 10.1001/jamanetworkopen.2021.0169 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Sandel M, Faugno E, Mingo A, et al. Neighborhood-level interventions to improve childhood opportunity and lift children out of poverty. Acad Pediatr. 2016;16(3)(suppl):S128-S135. doi: 10.1016/j.acap.2016.01.013 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1.

eMethods.

eTable 1. Baseline Characteristics of the Study Population by Tertiles of Neighborhood Socioeconomic Position During Midlife: The ARIC Study 1990-2019 (N=5538)

eTable 2. Combined Estimates of Percentage Difference From Median Rate of Cognitive Decline Between the Ages of 50 and 90 Years by Neighborhood Socioeconomic Position During Childhood and Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 3. Estimates of the Odds of Belonging to a Specific Quintile of Cognitive Decline Between the Ages of 50 and 90 Years by Tertiles of Neighborhood Socioeconomic Position During Childhood or Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 4. Combined Estimates of the Odds of Belonging to a Specific Quintile of Cognitive Decline Between the Ages of 50 and 90 Years by Continuous, Standardized Measures of Neighborhood Socioeconomic Position During Childhood and Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 5. Combined Estimates of the Odds of Belonging to a Specific Quintile of Cognitive Decline Between the Ages of 50 and 90 Years by Tertiles of Neighborhood Socioeconomic Position During Childhood and Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 6. Stratified Estimates of Percentage Difference From Median Rate of Cognitive Decline Between the Ages of 50 and 90 Years by Continuous, Standardized Measures of Neighborhood Socioeconomic Position During Childhood or Midlife: The ARIC Study 1990-2019 (N=5711)

eTable 7. Stratified Estimates of the Odds of Belonging to a Specific Quintile of Cognitive Decline Between the Ages of 50 and 90 Years by Continuous, Standardized Measures of Neighborhood Socioeconomic Position During Midlife: The ARIC Study 1990-2019 (N=5711)

eFigure. Flowchart of Participants Selected for Analysis

eReferences.

Click here for additional data file. (242.5KB, pdf)
Supplement 2.

Data Sharing Statement

Click here for additional data file. (15.9KB, pdf)

Articles from JAMA Network Open are provided here courtesy of American Medical Association

RESOURCES