Interactions among neighborhood conditions, sleep quality, and episodic memory across the adult lifespan

Abstract

Objectives

On average, adults racialized as non-Hispanic Black and Hispanic sleep more poorly than adults racialized as non-Hispanic White (hereafter, Black, Hispanic, White), but associations between factors that may moderate sleep-memory associations in these groups, such as neighborhood conditions, are unclear. Poorer neighborhood conditions (e.g., lower neighborhood cohesion) may be negatively associated with sleep quality and multiplicatively influence sleep-memory associations. We hypothesized lower ratings of neighborhood conditions would be associated with poorer sleep quality and moderate the association between sleep quality and episodic memory, especially in Black and Hispanic adults, who are disproportionately situated in poor neighborhood conditions.

Design

Seven-hundred-thirty-six adults across the adult lifespan (27–89 years) were recruited from the northern Manhattan community as a part of the Offspring Study of Racial and Ethnic Disparities in Alzheimer’s disease. Sleep quality was assessed using a modified version of the Pittsburgh Sleep Quality Index, and episodic memory was evaluated with the Buschke Selective Reminding Test. With multiple regression models, we measured associations between perceived neighborhood conditions and sleep quality and the interaction between sleep quality and neighborhood conditions on episodic memory stratified by racial/ethnic and gender identity groups.

Results

Overall, poorer neighborhood conditions were associated with poorer sleep quality. In Black and Hispanic women, the sleep-memory association was moderated by neighborhood conditions. With more favorable neighborhood conditions, Black women showed an association between higher sleep quality and higher memory performance, and Hispanic women showed a protective effect of neighborhood (higher memory even when sleep quality was poor).

Conclusion

Poorer neighborhood experiences may contribute to poorer sleep quality across groups. In Black and Hispanic women, the association between sleep quality and episodic memory performance was dependent upon neighborhood conditions. These findings may inform tailored, structural level sleep interventions, aimed to improve neighborhood experiences and thereby sleep quality and episodic memory.

Introduction

Sleep has been mechanistically linked with episodic memory performance, including encoding, consolidation, and retrieval processes.^1–3 Sleep quality is one modifiable factor that may help mitigate disparities in memory performance.⁴ Racialized differences in associations between sleep quality and memory performance may stem from structural inequities, as these inequities in minoritized racial and ethnic groups and women are the root of many health disparities.^5,6 Below, we will discuss racial and gender disparities in sleep quality, differences in sleep-memory associations by racial and ethnic groups, and the potential moderating role of structural factors, specifically the neighborhood environment, in sleep-memory associations.

Sleep disparities by gender and racial and ethnic group

As people age, both sleep quality and episodic memory decline.^7,8 As compared to older men, older women report poorer sleep quality (measured with questionnaires), however, they demonstrate higher objective sleep quality (e.g., sleep efficiency, measured with polysomnography) than men.⁹ Differing experiences by gender identity may affect sleep quality, including higher gender-based discrimination¹⁰ and caretaking expectations for women.¹¹

Regardless of age, both men and women who are racialized as non-Hispanic Black and Hispanic experience poorer sleep quality and lower sleep duration (i.e., poor sleep) than those racialized as non-Hispanic White (hereafter, these groups will be referred to as Black, Hispanic, and White).^12,13 These minoritized racial and ethnic groups may experience particularly poor self-reported sleep quality due to racism-related stress and related structural factors (e.g., financial stress, chronic stress, noise).^14,15

Associations among sleep quality, memory performance, and racial and ethnic differences

Poor sleep quality (e.g., low sleep efficiency and polysomnography-measured slow wave activity) has been associated with Alzheimer’s Disease and Related Dementias (ADRD) symptomology, including worse levels of amyloid and tau and poor memory performance.^1,16 Poorer sleep quality may be more strongly associated with lower memory-related neural activity and episodic memory performance in Black adults than White adults.^17–19 For example, findings demonstrated lower electroencephalography-measured neural oscillations, indicative of detailed memory representations (alpha desynchronization),²⁰ were associated with lower actigraphy-measured sleep consistency in Black adults, but not White adults. Emerging research suggests actigraphy-measured sleep quality (e.g., higher sleep continuity) is associated with higher memory performance in Black adults,^18,19,21 and one recent study demonstrated an association between self-reported daytime sleepiness and poorer memory in middle-aged to older Black adults.²² However, longitudinal associations showed greater daytime sleepiness was associated with lower decline in memory performance, and Black adults demonstrated no meaningful associations between self-reported sleep quality and amyloid burden.²² These few prior studies suggest a link among sleep quality, episodic memory performance, and brain health. Poorer sleep quality may underlie worse ADRD symptomology and lower episodic memory retrieval performance in minoritized racial and ethnic groups.

The neighborhood environment is associated with sleep quality and memory performance in minoritized racial and ethnic groups

Neighborhood and Sleep

The neighborhood environment may be one factor that modifies associations between sleep quality and memory performance. Due to historic neighborhood conditions that influenced residential segregation, people from minoritized racial and ethnic groups often reside in neighborhoods with crime-producing factors, including high levels of poverty.²³ Such neighborhoods are also disproportionately surveilled, increasing risk for adverse encounters with police that may increase hypervigilance, reduce poor mental health, and likely reduce sleep quality.¹⁵ People who live in higher quality neighborhood conditions, such as safer and more socially cohesive neighborhoods, also have self-reported and actigraphy-measured higher quality sleep than those who live in lower quality neighborhood conditions.²⁴ Compared to other racial and ethnic groups, Black adults have shown stronger associations between perceived neighborhood social cohesion and greater actigraphy-measured sleep duration.^25,26 The subjective experience of the neighborhood environment may influence sleep quality. When people perceive their neighborhoods as safer and less violent, they have higher self-reported²⁷ and actigraphy-measured sleep quality.²⁸ The neighborhood environment may be associated with sleep quality because of barriers (e.g., crime, violence) to and facilitators (social cohesion or support) of high sleep quality.^25,27 When people feel unsafe in an environment, they may be more highly vigilant and experience difficulty falling asleep and staying asleep throughout the night. By contrast, people who feel safer and more supported in their environments experience higher sleep quality (e.g., actigraphy-measured sleep efficiency),²⁸ as these low-stress conditions are conducive to sleep initiation and maintenance. Thus, one’s experience within the neighborhood environment is associated with sleep quality.

Neighborhood and Memory Performance

Racial and ethnic differences in neighborhood conditions may contribute to racial and ethnic differences in sleep, episodic memory, and moderate sleep-memory associations; the moderation effect of neighborhood on sleep-memory associations may show differential impact for minoritized racial and ethnic groups as compared to White adults. In the Midlife Development of the United States cohort, disparities in cognition between Black and White adults were partially explained by cumulative stressors across the lifecourse, including neighborhood stress.²⁹ Lower perceptions of neighborhood safety, trust, and support, along with other stressors (e.g., financial stress, work stress), were associated with lower episodic memory in Black as compared to White adults. Chronic stress, which is disproportionately high in minoritized racial and ethnic groups relative to White adults, contributes to lower hippocampal volume and may lead to poorer episodic memory.³⁰ Conversely, positive neighborhood perceptions may benefit episodic memory. In middle-aged to older adults in the China Health and Retirement Longitudinal Study, retrospective childhood perceptions of higher neighborhood social cohesion (e.g., trust, friendliness) were associated with higher episodic memory.³¹ Taken together, limited evidence suggests neighborhood perceptions influence memory and may contribute to associations between sleep quality and memory.

The neighborhood environment may moderate sleep-memory associations

The association between poorer neighborhood conditions and poorer sleep quality has been established.^25,27,28 Associations among poorer neighborhood conditions,²⁹ poorer sleep quality,³ and poorer episodic memory are understudied, particularly in Black and Hispanic adults who experience high levels of structural discrimination. The impact of poorer sleep quality on poorer episodic memory performance is disproportionately higher in minoritized racial and ethnic groups.^17–19 Additionally, structural discrimination often places Black and Hispanic adults in poorer neighborhood environments with potentially multiplicative effects with poorer sleep quality and poorer memory performance. Higher levels of discrimination and stress that are associated with sleep in these minoritized groups¹⁵ may moderate the association between sleep quality and episodic memory performance, such that more negative experiences strengthen the association between poorer sleep and poorer memory and positive ones buffer the sleep-memory association. For example, feeling unsafe within the neighborhood may be associated with intrusive thoughts during memory tasks, which may disrupt encoding and retrieval processes. By contrast, feeling supported by neighbors may be associated with higher self-efficacy and facilitate successful encoding and retrieval.³² Black and Hispanic adults with poorer experiences within their neighborhoods may show an exacerbated association between poorer sleep quality and poorer memory performance. Conversely, Black and Hispanic adults who report positive neighborhood experiences may demonstrate a protective effect of neighborhood, such that more favorable neighborhood conditions attenuate the association between poorer sleep quality and poorer memory performance. Disparities related to intersections between racial and ethnic identity and gender identity may be associated with relationships among, sleep, memory, and neighborhood conditions. However, little research has investigated intersectionality in relation to neighborhood conditions and cognition³³ in middle and older age, during which both sleep quality⁷ and episodic memory⁸ decline.

The present study

We use the National Institute on Aging (NIA) Health Disparities Research Framework to better understand racial and ethnic differences in sleep-memory associations, as it provides guidelines for assessing multiple factors that may impact racial and ethnic health disparities in aging at the environmental, sociocultural, behavioral, and biological levels.³⁴ Here, we focus on environmental factors (e.g., neighborhood safety and neighborhood cohesion) as related to biobehavioral factors (self-reported sleep quality) with episodic memory performance as the outcome of interest. This study is among the first to investigate neighborhood perceptions that may contribute to individual differences in sleep-memory associations in Black, Hispanic, and White adults. In alignment with the NIA Health Disparities Research Framework, the present study takes an intersectional approach (race and ethnicity by gender identity) to explore associations among neighborhood perceptions, self-reported sleep quality, and episodic memory in the Offspring Study of Racial and Ethnic Disparities in Alzheimer’s disease (Offspring), a community-based cohort of middle-aged to older adults.³⁵ This cross-sectional study includes participants who are in middle age, allowing us to better capture factors that may aid ADRD prevention. We hypothesize (1) more favorable neighborhood perceptions are associated with higher sleep quality and (2) neighborhood conditions moderate the association between sleep quality and episodic memory performance. We expect these associations to be present in Black and Hispanic adults, but not White adults. The findings may further vary by intersections within gender identity group (Black women, Hispanic women, Black men, and Hispanic men).

Method

Study Design and Sample

The current study used available cross-sectional data from the Offspring Study, an ongoing community-based cohort study designed to identify structural, social, and neural pathways of racial and ethnic cognitive disparities.³⁵ Those who identified as Hispanic are of predominantly Dominican descent, but the Hispanic group also includes adults of Puerto Rican and Cuban descent. At the time of data analysis, 736 adults (19% Black, 59% Hispanic; 22% White; <1% Other [multiple racial and ethnic identities/refused]) were included in the analytic sample, which included those with complete data on our primary exposure and outcome variables at baseline (see Measures). Upon participation, participants completed informed consent forms approved by the Institutional Review Board at Columbia University (protocol number: AAAR1120).

Measures

Exposures

Race and Ethnicity.

Race and ethnicity in this study are used to understand ongoing effects of racialization and racially patterned systems of differential resource allocation (e.g., high quality neighborhood environments) on sleep quality and episodic memory. In the Offspring study, the measures of race and ethnicity are collected as follows. Participants self-select their racial and ethnic categories from several options (i.e., American Indian or Alaskan Native, Asian, Black or African American or African, Hispanic/Latino or Spanish, Middle Eastern or North African, Native Hawaiian or other Pacific Islander, and White) and additionally select from other ethnic identifications (e.g., South African, Costa Rican, Jamaican, Mexican, Senegalese). Participants in the present analysis were categorized as Black (non-Hispanic Black and any other racial category), Hispanic (Hispanic), and White (non-Hispanic White). A few participants who did not identify as any of these three categories (n=3) were included in overall analyses (across groups) but were not included in analyses stratified by racial and ethnic group.

Neighborhood Conditions.

Participants self-reported positive (cohesion, safety) and negative (problems, disorder) aspects of their neighborhood experiences on 29 widely-used questionnaire items,^36,37 which were summed on subscales for each aspect of the neighborhood. Neighborhood cohesion provides an index of trust and belonging among neighbors and was assessed with four items on a Likert-type scale form 1 (higher cohesion) to 7 (lower cohesion). Neighborhood safety was assessed with two items regarding feeling safe (1) or unsafe (4) while walking in the neighborhood during the daytime and nighttime. Neighborhood problems were measured with 19 items from 1 (big problem) to 3 (not a problem), including issues with tension, unemployment, and policing. These three measures were reversed scored such that higher scores reflected higher cohesion, safety, and problems. Neighborhood disorder was estimated using four items, such as the presence of abandoned houses, vandalism, and cleanliness. Responses ranged from 1 (lower disorder) to 4 (higher disorder).

Sleep.

Sleep quality was measured with a modified version of the Pittsburgh Sleep Quality Index (PSQI).³⁸ For the analyses presented here, our primary sleep measure is a raw score of self-reported sleep quality (range: 0 to 21). We estimated sleep quality with sleep component 5, which is a measure of sleep disruption (e.g., trouble staying asleep and factors interfering with sleep, including temperature, pain, coughing, bathroom visits). Lower values of sleep disruption indicate higher sleep quality. Sleep quality was assessed as a continuous variable and was used as both an outcome (for neighborhood conditions) and exposure (for episodic memory) measure. Prior research has assessed individual components of the PSQI,^39,40 and has shown associations between the sleep disruption component and actigraphy-measured sleep disruption (i.e., wake after sleep onset).³⁹ This sleep disruption measure may also be sensitive to racial and ethnic differences in associations between sleep quality and episodic memory performance.⁴⁰ See Supplementary Materials for all sleep items measured for the present analyses.

Primary Outcome

Episodic Memory.

The Selective Reminding Test was used to assess delayed memory recall.⁴¹ During the test, six learning trials are administered, and for each trial, participants are asked to learn a list of 12 words. After each trial, participants are given reminders only for the words that they did not recall. The total number of words recalled over all six learning trials is recorded. Lastly, after a 15-minute delay, free recall of the words is assessed (delayed recall; range: 0 to 12 words). Higher scores indicate higher episodic memory performance. Our outcome measure for episodic memory is the raw score for delayed call.

Covariates and Moderators

We adjusted for age, years of education, household income level (e.g., participant and others contributing to household income), and immediate retrieval performance. Age and years of education were treated as continuous variables. We stratified our results by racial and ethnic and gender identity group (e.g., Black women, White women). Participants self-reported their gender identities as woman (n=474), man (n=259), transgender woman (n=2), and transgender man (n=1). We binarized gender into women (women/transgender women) and men (men/transgender men) to stratify our analyses with sufficient participants in each group. See Supplementary Table 1 for all reported gender identities. Income was trichotomized into annual household income lower than $30,000, between $30,000 and $89,999, and higher than $90,000. For the income categories, we considered estimates of the poverty line in the city of New York and the United States, which averaged $36,262 and $26,926, respectively, in 2019. We also adjusted for immediate recall performance, which was defined as the number of words recalled during the sixth recall trial (see Episodic Memory section above).

Analytic Approach

For each hypothesis, we performed regression analyses (1) across racial and ethnic groups and gender identity groups and (2) stratified by race and ethnicity and gender identity group (Black women, Black men, Hispanic women, Hispanic men, White women, White men). We examined (1) if more favorable neighborhood conditions were associated with higher sleep quality and (2) if the association between sleep quality and memory performance was moderated by neighborhood conditions. In each moderation analysis, we include the main effect of sleep quality, the main effect of neighborhood condition, and the interaction effect between sleep quality and neighborhood conditions. Regression models were separately assessed for each neighborhood condition. We tested these associations using regression models adjusted for age, years of education, and income level. We present each of these models in the Results section.

Missing Data

Missing data were imputed using the Multiple Imputation by Chained Equations (MICE) package in R. We imputed 5 datasets for missing data for covariates, including years of education and income level. We repeated regression analyses in the complete dataset of 615 participants (see Supplementary Tables 3–4). The results for the imputed and complete dataset were similar. Thus, we report the results for the imputed dataset only.

Results

Of the 736 participants, 615 (84%) had complete data, and most of the missing data was because they did not report income (n=81, 11%), which is one of the covariates. Participants who did not wish to report income or any other covariate also reported lower neighborhood cohesion (−1.18, 95% CI [−2.26, −0.11], p = .031) than those who had complete data. Given the difference between participants with complete data and those with missing data, we conducted an imputation for missing covariates to reduce bias from having missing data not completely at random.

More Favorable Neighborhood Conditions Were Associated with Higher Sleep Quality Across the Full Sample

In the full sample, more positive perceptions of neighborhood environments, including neighborhood safety, social cohesion, neighborhood disorder, and neighborhood problems, were associated with higher sleep quality (see Table 2 for the details of each regression analysis).

Table 2.

Adjusted associations of perceived neighborhood conditions and sleep quality in all participants and stratified by each racial and ethnic and gender identity group (full sample n=736)

	Overall			Black Men			Black Women			Hispanic Men			Hispanic Women			White Men			White Women
	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value
Neighborhood Safety	−0.75	−1.0, −0.46	<0.001	−2.3	−3.6, −0.90	0.002	−0.27	−1.2, 0.61	0.5	−0.94	−1.6, −0.31	0.004	−0.64	−1.1, −0.21	0.004	−0.16	−1.5, 1.2	0.8	−0.3	−1.3, 0.67	0.5
Neighborhood Cohesion	−0.12	−0.18, −0.05	<0.001	−0.11	−0.35, 0.12	0.3	−0.1	−0.33, 0.12	0.4	−0.14	−0.27, −0.01	0.038	−0.1	−0.20, 0.00	0.048	−0.03	−0.25, 0.19	0.8	−0.15	−0.33, 0.03	0.11
Neighborhood Disorder	0.09	0.03, 0.16	0.004	0.12	−0.13, 0.36	0.3	0.06	−0.14, 0.27	0.6	0.12	−0.02, 0.26	0.089	0.06	−0.05, 0.16	0.3	0.03	−0.18, 0.25	0.8	0.18	−0.01, 0.37	0.057
Neighborhood Problems	0.08	0.03, 0.13	0.001	0.08	−0.11, 0.28	0.4	0.03	−0.12, 0.17	0.7	0.17	0.06, 0.29	0.004	0.02	−0.06, 0.10	0.6	0.25	0.10, 0.41	0.002	0.12	−0.01, 0.24	0.062

Note:

¹CI = Confidence Interval

Results by racial and ethnic and gender identity subgroups were varied. Black men, Hispanic men, and Hispanic women reported higher quality sleep when they perceived that their neighborhoods were safer. Both Hispanic men and Hispanic women reported higher quality sleep with higher neighborhood cohesion. Hispanic men and White men reported higher quality sleep when they perceived fewer neighborhood problems within their neighborhoods. No other statistically significant associations between neighborhood perceptions and sleep quality were found in these groups (see Table 2 and Figure 1 for the details of each regression analysis).

Figure 1.

Forest plots for adjusted associations of perceived neighborhood conditions and sleep quality in all participants and stratified by each racial and ethnic and gender identity group (full sample n=736).

Associations Between Sleep Quality and Episodic Memory Performance Are Dependent Upon Neighborhood Conditions in Black and Hispanic Women

When Black women perceived their neighborhood conditions more favorably, particularly when they experienced lower neighborhood disorder and problems, they demonstrated associations between higher sleep quality and higher memory performance (See Table 3 and Figure 2 for detailed interaction effects). When neighborhood disorder and problems were higher, Black women performed worse on delayed memory recall tests, even when their sleep quality was high. Similar evidence for this pattern emerged when neighborhood cohesion and neighborhood safety were higher, but these moderation effects did not reach statistical significance (See Figure 3 for simple effects for each moderation analysis).

Table 3.

Adjusted associations of the interaction effect of sleep quality and perceived neighborhood conditions on episodic memory in all participants and stratified by each racial and ethnic and gender identity group (full sample n=736)

Characteristic	Overall			Black Men			Black Women			Hispanic Men			Hispanic Women			White Men			White Women
Beta	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value	Beta	95% CI1	p-value
Sleep Quality × Safety	−0.02	−0.04, 0.00	0.11	−0.07	−0.26, 0.11	0.4	0.06	−0.02, 0.14	0.11	0.01	−0.05, 0.07	0.8	−0.03	−0.06, 0.01	0.12	0.14	−0.09, 0.38	0.2	−0.02	−0.14, 0.09	0.7
Sleep Quality × Cohesion	0.00	−0.01, 0.00	0.3	−0.02	−0.05, 0.01	0.15	0.02	0.00, 0.04	0.063	0.00	−0.01, 0.01	>0.9	−0.01	−0.02, 0.00	0.046	0.02	−0.02, 0.06	0.3	−0.01	−0.02, 0.00	0.14
Sleep Quality × Disorder	0.00	−0.01, 0.00	0.6	−0.02	−0.05, 0.01	0.2	0.02	0.01, 0.04	0.009	0.00	−0.02, 0.02	>0.9	−0.01	−0.02, 0.00	0.043	0.01	−0.02, 0.03	0.7	−0.02	−0.04, 0.00	0.1
Sleep Quality × Problems	0.00	−0.01, 0.00	0.3	0.00	−0.02, 0.02	0.9	−0.01	−0.03, 0.00	0.013	−0.01	−0.02, 0.01	0.5	0.00	0.00, 0.01	0.4	0.00	−0.03, 0.02	0.8	0.00	−0.01, 0.02	0.6

Note:

¹CI = Confidence Interval

Figure 2.

Forest plots of adjusted associations of the interaction effect of sleep quality and perceived neighborhood conditions on episodic memory in all participants and stratified by each racial and ethnic and gender identity group (full sample n=736).

Figure 3.

Interaction effects neighborhood conditions and sleep quality on episodic memory recall in Black women. Note: These simple slope moderation effects were adjusted for covariates and generated with available data that was not imputed.

When Hispanic women perceived their neighborhood conditions more favorably, they showed a protective effect against poorer sleep quality. Specifically, when neighborhood cohesion was high, Hispanic women demonstrated higher memory performance, even when sleep quality was poorer (See Table 3 and Figure 2). Similarly, when neighborhood disorder was lower, Hispanic women showed higher memory performance at poorer levels of sleep quality. When they perceived their neighborhood conditions less favorably, poorer sleep quality was associated with poorer memory performance (See Figure 4 for simple effects for each moderation analysis).

Figure 4.

Interaction effects of neighborhood conditions and sleep quality on episodic memory recall in Hispanic women. Note: These simple slope moderation effects were adjusted for covariates and generated with available data that was not imputed.

Taken together, when Black women perceived their neighborhood conditions as poorer, they did not demonstrate associations between higher sleep quality and higher memory performance (See Figures 2 and 3). They experienced no benefit of high sleep quality when they perceived their neighborhood conditions more poorly (higher disorder and higher problems). However, Hispanic women showed a protective effect against poorer sleep quality when they perceived their neighborhoods more favorably (higher cohesion or lower disorder). We did not find moderation effects in the full sample across groups, Black men, or Hispanic men, or White adults for either gender identity group.

Discussion

We demonstrated that neighborhood perceptions are associated with sleep quality and memory performance in a cohort of middle-aged to older adults. Overall, more favorable neighborhood perceptions were associated with higher sleep quality and memory performance. Unexpectedly, in Black and Hispanic women only, neighborhood conditions moderated the association between higher sleep quality and higher memory performance.

Research in various populations, including Black and Hispanic adults, have shown associations between higher quality neighborhood conditions and higher quality sleep.^27,28,42 Consistent with this prior research, we found higher perceptions of neighborhood safety and cohesion and lower problems and disorder were associated with higher sleep quality in the full sample. Within the smaller groups, statistical significance varied by racial and ethnic and gender identity, however all associations were in the same direction, such that higher perceptions of neighborhood conditions were associated with higher quality sleep. Collectively, these findings suggest feeling more positively about one’s neighborhood environment is related to higher sleep quality, regardless of racial and ethnic or gender identity. The association between neighborhood conditions and sleep quality may be explained by stress and pre-sleep arousal (e.g., persistent worry).⁴³ For example, feelings of safety and security may reduce arousal (i.e., hypervigilance) that interferes with falling asleep and staying asleep throughout the night.⁴⁴ Equal access to well-resourced neighborhoods may reduce stress and improve sleep across various groups, reducing racial and ethnic sleep disparities. Living closer to neighborhood investments, including food markets and sidewalks, resulted in lower sleep quality decline in older Black adults as compared to those who lived further from investments.⁴⁵ Thus, poorer sleep in minoritized populations may be partially attributed to living in more poorly-resourced areas.

Sleep quality is associated with episodic memory performance,^1,9 but little research exists on sleep-memory associations in minoritized racial and ethnic groups.³ Although numerous studies have directly investigated associations between neighborhood conditions and sleep quality, few have investigated the consequences of poor sleep on episodic memory in Black and Hispanic adults.^18,19 Emerging research in this area suggests that associations between sleep quality and memory are stronger in Black adults as compared to White adults.^18,19 Our recent findings (under review) in the Offspring Study suggest Hispanic and Black adults may show associations between self-reported sleep quality and episodic memory recall, but White adults do not show this association. The present study extends this research by investigating the underlying causes of racial and ethnic differences in sleep-memory associations, one of which may be the neighborhood environment.

Neighborhood conditions that disrupt sleep may interfere with biological mechanisms that facilitate sleep-based memory consolidation^1,46 and thus lower episodic memory retrieval. The Active System Consolidation Hypothesis posits that sleep actively strengthens memory representations over time through reactivation of memories in the hippocampus that are gradually transferred to the neocortex for long-term storage.¹ The Synaptic Homeostasis Hypothesis accounts for the brain reorganization during sleep, which restores encoding capacity.² When these processes are disrupted by sleep deprivation or sleep disturbance, the encoding capacity of the hippocampus is reduced, resulting in lower learning and episodic memory retrieval.^46–48 Both total sleep deprivation⁴⁷ and partial sleep deprivation⁴⁸ have been associated with poor episodic memory encoding and retrieval performance. For example, adolescents who were restricted to five hours of sleep for five nights performed worse on a recognition memory test than those who slept normally, even after three nights of recovery sleep.⁴⁸ These findings evidence sleep loss reduces encoding ability. Chronically poor sleep quality from living in a poor neighborhood environment that disturbs sleep-based memory consolidation may reduce encoding ability and thereby reduce delayed memory retrieval.

Neighborhood-level factors may also be directly associated with cognition. Poor experiences from living in a neighborhood environment that feels unsafe, unfriendly, and disorderly may lead to chronic stress, which negatively impacts brain health and cognitive performance.³⁰ Racial and ethnic disparities in cognition have been partially explained by cumulative neighborhood stressors across the lifecourse.²⁹ Lower perceived neighborhood safety, trust, and support were associated with lower episodic memory in Black adults as compared to White adults. Similarly, adults aged 45 and older in the China Health and Retirement Longitudinal Study who retrospectively reported higher perceptions of childhood neighborhood social cohesion, such as trust and friendliness, demonstrated higher episodic memory.³¹

In the present study, Black women showed associations between higher sleep quality and higher episodic memory only when neighborhood conditions were also higher, suggesting the benefits of sleep quality are dependent upon more favorable perceptions of one’s neighborhood. Interestingly, Hispanic women showed a protective effect of neighborhood, such that favorable neighborhood conditions attenuated the negative association between poorer sleep quality poorer and poorer memory performance.

Black and Hispanic women reported experiencing at least two types of sleep disturbances (e.g., pain and uncomfortable temperature) more than three times per week. Some research shows women report poorer sleep, but they often show better objective sleep quality (e.g., actigraphy-measured sleep efficiency).⁹ The reasoning for this discrepancy is unclear. Many women have an increased burden of responsibilities (e.g., being a primary source of income, as well as the primary planner and caregiver for children and aging parents)¹¹ and experience gender-based discrimination that can interfere with sleep and may show greater cognitive consequences (lower memory) even when they sleep well. After the Civil War, racially restrictive housing policies, such as redlining, isolated racially minoritized individuals into communities and diverted resources away from these neighborhoods. Even after ending redlining (Fair Housing Act) and decades of growing diversity, most Americans still live in racially segregated neighborhoods. Residential segregation has resulted in persistent inequalities, including dilapidated housing in neighborhoods, in which minoritized people reside, and poor quality of social and built environments.²³ Sexism and racism may also interact, such that the neighborhood has a more salient influence on the relationship between sleep and memory among Black and Hispanic women than other groups. Perhaps the neighborhood environment has a stronger effect on memory performance in women from minoritized groups because of greater environmental concerns (e.g., safety concerns that lead to intrusive thoughts or rumination about the neighborhood community or dependents), along with higher caregiving and planning responsibilities,¹¹ that affect their ability to effectively encode and retrieve episodic memories.

Given differences in exposures of structural racism and thereby resource allocation, we expected significant moderation effects in each minoritized racial and ethnic group, including Black and Hispanic men. Other groups may be better able to compensate for poorer sleep quality because of lower experiences with task-irrelevant thoughts (e.g., intrusions regarding neighborhood safety). However, if other responsibilities, such as caretaking burdens, are indeed lower in our sample of men, these groups may not be as strongly impacted by the neighborhood environment as the women are. Thus, women from minoritized groups may be multiplicatively impacted by several forms of discrimination (structural, interpersonal) and higher responsibility burdens. Another explanation for the null findings in the other groups is the small sample size of these race/ethnicity and gender subgroups. Because the small sample size limits generalizability of the Offspring sample, our findings should be replicated in larger studies.

Strengths, Limitations, and Future Directions

The present study has several strengths. The analytic sample includes 736 adults (19% Black, 59% Hispanic; 22% White) across the adult lifespan, including middle age. We also use a detailed measure of sleep quality, the sleep disturbance component of the Pittsburgh Sleep Quality Index,³⁸ that encompasses several experiences that may disturb sleep. Moreover, episodic memory is assessed using the Selective Reminding Task.⁴¹ We used delayed recall as our memory metric of interest, which is sensitive to sleep disturbances, but does not capture sleep-dependent memory consolidation following sleep.⁴⁹ Our study also has limitations. The study was conducted with people from the Northern Manhattan community and may not generalize to other areas. While racially and ethnically diverse, our group comparisons are limited to Black, Hispanic, and White adults who identify as men and women. Future research should evaluate associations among neighborhood conditions, sleep, and memory in other racial and ethnic groups (e.g., Asian, Dominican, Mexican, African, African American) and include minoritized gender identity and sexuality groups. Even when groups are small, studies should aim to assess within-group variation within smaller groups.

Our study is also limited by self-report measurements of sleep quality, which are subject to inaccuracies. We used a modified version of the PSQI and were not able to assess global sleep quality. Future studies should aim to collect both objective (e.g., actigraphy) and subjective (questionnaires) sleep measures for a comprehensive assessment of sleep health. The present study is cross-sectional and cannot determine causality. In addition to current levels of sleep quality, future research should assess change over time in sleep health and neighborhood perceptions in relation to memory outcomes. Sleep quality should be assessed as mediator of the association between neighborhood conditions and memory performance with longitudinal data. The present cross-sectional findings suggest higher sleep quality alone may not be sufficient to support higher memory performance, as poorer neighborhood perceptions may attenuate the benefits of higher quality sleep on memory, particularly in Black women. Current recommendations emphasize environmental-level interventions may be more effective than those focusing on the individual-level alone.⁵⁰ Longitudinal study designs would further explicate which factors are most important to target for cognitive interventions. Moreover, future research should collect residential data, which would allow for parallel objective assessments of neighborhood conditions (e.g., crime rate, walkability).

Conclusion

The present study is of the first to investigate factors that may influence sleep-memory associations in different racial and ethnic and gender identity groups in middle-aged to older adults. Our findings suggests that the association between sleep quality and episodic memory may be dependent on perceptions of one’s neighborhood environment, particularly in Black and Hispanic women. Improving neighborhood conditions in areas with higher populations of women from these minoritized groups may result in improved sleep quality and facilitate associations between better sleep and better episodic memory, thus narrowing racial and ethnic cognitive disparities. Future research should build on the present findings to determine if neighborhood-level interventions administered during middle and older age impact sleep and memory outcomes in racially and ethnically diverse populations. Such interventions, particularly during middle-age, may prevent cognitive decline and reduce disparities in sleep quality and memory performance.

Table 1.

Participant demographic and clinical characteristics for included participants in the Offspring Study of Racial and Ethnic Disparities in Alzheimer’s Disease.

	Black		Hispanic		White		Other
	Men, n=531	Women, n=861	Men, n=1401	Women, n=2941	Men, n=661	Women, n=941	Men, n=11	Women, n=21
Age	54 (11)	58 (9)	54 (11)	55 (10)	60 (13)	56 (12)	63	49 (15)
Years of Education	13.69 (2.37)	14.68 (2.48)	13.2 (3.1)	13.2 (3.5)	17.03 (1.82)	17.04 (1.74)	14	17.00 (1.41)
Missing	1	1	3	7
Income Level
<30k	19 / 50 (38%)	26 / 77 (34%)	43 / 124 (35%)	117 / 259 (45%)	4 / 60 (6.7%)	11 / 83 (13%)	0 / 1 (0%)	0 / 1 (0%)
30–89k	22 / 50 (44%)	38 / 77 (49%)	48 / 124 (39%)	94 / 259 (36%)	20 / 60 (33%)	24 / 83 (29%)	0 / 1 (0%)	0 / 1 (0%)
>90k	9 / 50 (18%)	13 / 77 (17%)	33 / 124 (27%)	48 / 259 (19%)	36 / 60 (60%)	48 / 83 (58%)	1 / 1 (100%)	1 / 1 (100%)
Missing	3	9	16	35	6	11	0	1
Neighborhood Safety	7.66 (0.88)	7.28 (1.27)	7.45 (1.15)	7.13 (1.36)	7.71 (0.63)	7.41 (0.77)	8	7.00 (1.41)
Neighborhood Cohesion	21.1 (5.6)	22.1 (5.1)	22.3 (5.5)	21.4 (6.1)	23.6 (4.0)	23.7 (4.3)	26	23.50 (6.36)
Neighborhood Disorder	10.0 (5.6)	9.8 (5.8)	8.6 (5.5)	9.6 (5.8)	8.1 (4.1)	8.1 (4.0)	5	11.0 (9.9)
Neighborhood Problems	26 (7)	26 (8)	25 (6)	25 (8)	23.2 (5.1)	23.2 (6.1)	19	29.0 (11.3)
Sleep Quality	5.1 (4.6)	7.0 (5.0)	4.5 (4.3)	6.5 (5.1)	4.7 (3.4)	5.1 (3.6)	3	10.5 (10.6)

Note:

¹Mean (SD); n / N (%)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author, JM, upon reasonable request.