The relationship between marital status and cognitive function in Chinese older adults: The mediating role of depressive symptoms

Abstract

Background

Emerging research indicates a possible link between marital status and cognitive performance among elderly individuals, yet limited investigations have explored the potential mediating effects of depressive symptoms within this association.

Objective

The present study aimed to examine the intermediary function of depressive symptoms in the connection between marital status and cognitive abilities within China's aging population.

Methods

The data for this study were obtained from the 2015 CHARLS survey wave, which included 5671 elderly people aged ≥60 years. To assess the associations among marital status, cognitive performance, and depressive symptoms, we employed linear regression modeling. Subsequently, mediation analysis was conducted to investigate whether depressive symptoms might serve as an intermediary factor in the connection between marital status and cognitive functioning. Subgroup analyses were performed according to marital status, sex, residential area, and education in order to evaluate the robustness and heterogeneity of the mediation effect.

Results

The findings indicated that marital status was positively associated with cognitive function in the elderly (β = 0.49, 95% CI: 0.17, 0.81, p = 0.0023), with married individuals demonstrating superior cognitive function. Depressive symptoms accounted for approximately 27.39% of the total effect of the correlation between marital status and cognitive impairment (95% CI: 14.16%, 79.68%, p = 0.002).

Conclusions

Our research indicates that depressive symptoms serve as a mediating factor in the relationship between marital status and cognitive function in aging populations. Older adults with cognitive impairment should be assessed for their marital status, and early identification and intervention for depressive symptoms are recommended.

Keywords

Alzheimer's disease cognitive impairment depressive symptoms marital status mediation analysis

Introduction

Cognitive function encompasses multiple domains, including attention, learning and memory, executive function, visuospatial ability, social cognition, and language ability.¹ Mild cognitive impairment is characterized by performance below the expected level for an individual's age and education in at least one cognitive domain. Dementia is a more serious form of decline in cognitive function, often accompanied by a decline in activities of daily living.^1,2 With global population aging, the prevalence of both mild cognitive impairment and dementia has been steadily increasing, creating a substantial public health challenge affecting millions worldwide.^3,4 For families, caring for a patient with dementia often means a huge emotional and financial burden. Studies indicate that the burden on family caregivers increases with the degree of cognitive decline in patients, especially when patients exhibit obvious behavioral and psychological symptoms.^5,6 From a societal perspective, the social and economic burden of dementia continues to grow amid global population aging. The complex care needs of individuals living with dementia typically demand substantial medical resources and social support, presenting a major challenge to public health systems worldwide.⁷

Due to the shrinking social circle of older adults, they tend to be more dependent on close relationships. Compared to other family members, children, or friends, spouses in a marital relationship are more likely to provide assistance with daily living activities,⁸ emotional support,⁹ and health-related behaviors.^10–12 Evidence suggests that marriage may confer cognitive benefits through the provision of social support, mental engagement, and companionship, all of which are associated with reduced risk of cognitive impairment.¹³ Consistently, studies indicate that being in a positive marital relationship correlates with better cognitive health.¹⁴ Conversely, individuals who never married, or who are separated, divorced, or widowed, are at an increased risk of cognitive decline.^15–18

Individuals without a spouse may face significant social pressures alone, while those experiencing marital breakdown must also cope with the negative emotions stemming from marital issues. These groups are more likely to develop unhealthy psychological states and even depressive symptoms than those in stable marriages.¹⁹ Extensive evidence demonstrates that depressive symptoms negatively impact the overall cognitive function of older adults, impairing their attention, memory, and executive function.^20,21 Additionally, elderly people with depressive symptoms experience a faster rate of brain aging, which is closely related to their cognitive performance and functional impairment.²² Furthermore, depressive symptoms can impair the ability of the elderly to carry out daily activities and participate in social activities by affecting cognitive function.²³ However, the role of depressive symptoms as a mediator in the relationship between marriage and cognition remains unclear.

This study aimed to systematically explore the pathways and factors underlying the association between marital status and cognitive ability in older adults through mediation analysis, with a focus on the mediating role of depressive symptoms. We used data from the China Health and Retirement Longitudinal Study (CHARLS) and constructed a theoretical path model linking marital status, depressive symptoms, and cognitive function. The following hypotheses were proposed: (1) A relationship exists between marital status and cognitive function in elderly people; (2) Depressive symptoms are associated with both marital status and cognition; (3) Depressive symptoms act as a mediator of the association between marital status and cognition. By quantifying the mediating contribution of depressive symptoms, this study provides empirical evidence for understanding the psychological mechanisms through which social relationships influence cognitive aging and offers theoretical support for developing cognitive intervention strategies targeting specific marital status groups. The findings are expected to fill the current gap in research on the mechanisms underlying the relationship of marital status with cognition in the Chinese population, particularly highlighting the important role of sociopsychological factors in this context.

Methods

Participants

The data for this study were collected from the 2015 wave of CHARLS. CHARLS is a prospective national cohort study that provides dependable data on the health and associated factors of middle-aged and elderly individuals in China.

Initially, data from 21,097 participants were included. However, 37 samples with missing marital status data, 4275 samples with missing cognitive ability assessment data, 9220 samples from individuals aged under 60 or with missing age data, and 1892 samples with missing or erroneous covariate data were eliminated. This resulted in a final sample size of 5671 participants.

Cognitive function assessment

In the CHARLS, participants underwent in-person assessments across four cognitive function domains: memory, orientation, drawing, and calculation. The Telephone Interview for Cognitive Status (TICS) was employed to assess orientation and calculation. Orientation assessment included day, month, year, current season, and day of the week. The orientation dimension was given a total score of 5 points, with 1 point allocated for each item. Participants were asked to subtract 7 from 100 five times consecutively, with one point awarded for each correct answer. For memory assessment, the assessor read 10 words randomly to each participant, and immediate memory was assessed by the number of words recalled immediately. After participants completed other tasks, delayed word recall was assessed. The total memory score was calculated by summing the immediate and delayed recall scores (maximum 20 points). For drawing ability, participants were asked to copy two overlapping five-pointed star images, scoring 1 point for a correct copy. The total cognitive score was the sum of memory (20 points), orientation (5 points), drawing (1 point), and calculation (5 points), for a total of 31 points.

Marital status

In CHARLS, marital status was categorized into seven types. This study dichotomized marital status into two groups: Unmarried, including divorced, separated, never married, and widowed; and Married, including married and living with spouse, married but temporarily not living with spouse, and cohabited.

Depression assessment

The Center for Epidemiological Studies Depression Scale (CES-D-10) was used to evaluate the presence of depressive symptoms. The CES-D-10 consists of 10 questions with 4 response options, scored from 0 to 3. The total score ranges from 0 to 30, with lower scores indicating lower levels of depressive symptoms. A CES-D score of ≥10 was used to indicate the presence of depressive symptoms or an elevated risk of depression.²⁴

Covariates

We chose established risk factors for cognitive decline as possible confounders.²⁵ Specifically, this includes: continuous variables: age, body mass index (BMI) (kg/m²). Categorical variables: sex, residential area (urban and rural), education (illiterate, primary school, middle school, high school or above), ever smoked (yes/no), ever alcohol use (yes/no), hypertension (physician-diagnosed), hyperglycemia (physician-diagnosed diabetes, impaired glucose tolerance, or elevated fasting glucose), stroke (physician-diagnosed) and dyslipidemia (physician-informed diagnosis based on elevated low-density lipoprotein (LDL), triglycerides, or total cholesterol and/or low high-density lipoprotein (HDL)).

Statistical analysis

All data were analyzed using R (version 4.2) software. Demographic characteristics are presented according to marital status. Categorical variables are expressed as numbers and percentages, and continuous variables as means ± standard deviation (SD). Stepwise regression analysis was conducted. First, we analyzed the overall consequences of marital status for cognitive function. Second, we examined the effects of marital status on depressive symptoms. Finally, the effect of depressive symptoms on cognitive function was investigated after controlling for marital status. Three models were used: an unadjusted model, Model I (adjusted for age, sex, residential area), and Model II (age, sex, residential area, education, BMI, ever smoked, ever alcohol use, hypertension, hyperglycemia, stroke, and dyslipidemia). When analyzing the relationship between depressive symptoms and cognition, we also adjusted for marital status.

A mediation analysis was conducted to determine whether the influence of marital status on cognitive ability was affected by depression score. We performed subgroup analyses to assess the generalizability and variation of the mediation effects. First, the analysis was stratified by marital status to examine pathway heterogeneity. Second, it was stratified by key demographic factors (sex, residence area, and education) to test the robustness of the effects across different subpopulations. The threshold for statistical significance was set at p < 0.05.

Results

Baseline characteristics of participants

A total of 5671 participants were included in this study, with a mean age of 67.54 ± 6.06 years; 45.25% were female. The mean score for cognitive function was 14.03 ± 5.16 points, and the mean CES-D 10 score was 8.03 ± 6.32 points. Based on marital status, participants were categorized into married (n = 4697) and unmarried (n = 974) groups. Compared to the unmarried group, the married cohort had a higher proportion of women, a higher educational attainment level, higher cognitive function scores, higher smoking and drinking rates, a lower average age, fewer depressive symptoms, and lower hypertension rates (p < 0.05). No statistically significant differences emerged in the prevalence of hyperglycemia, stroke, or dyslipidemia in the two groups (p > 0.05) (Table 1).

Table 1.

The baseline characteristics of participants.

Variable	Unmarried	married	Standardize diff.	p
Participants	974	4697
Age	70.83 ± 7.18	66.86 ± 5.56	0.62 (0.55, 0.69)	<0.001
Cognitive function score	12.48 ± 5.39	14.35 ± 5.05	0.36 (0.29, 0.43)	<0.001
CES-D 10	9.36 ± 6.90	7.75 ± 6.16	0.25 (0.18, 0.32)	<0.001
BMI	23.32 ± 4.07	23.60 ± 4.08	0.07 (−0.00, 0.14)	0.053
Sex			0.37 (0.30, 0.44)	<0.001
Male	589 (60.47%)	1977 (42.09%)
Female	385 (39.53%)	2720 (57.91%)
Residential area			0.03 (−0.04, 0.10)	0.418
Rural	392 (40.25%)	1825 (38.85%)
Urban	582 (59.75%)	2872 (61.15%)
Education			0.28 (0.21, 0.35)	<0.001
Illiterate	565 (58.01%)	2132 (45.39%)
Primary school	236 (24.23%)	1322 (28.15%)
Middle school	128 (13.14%)	823 (17.52%)
High school or above	45 (4.62%)	420 (8.94%)
Ever alcohol use			0.15 (0.08, 0.22)	<0.001
no	551 (56.57%)	2306 (49.10%)
yes	423 (43.43%)	2391 (50.90%)
Ever smoked			0.16 (0.09, 0.23)	<0.001
no	542 (55.65%)	2232 (47.52%)
yes	432 (44.35%)	2465 (52.48%)
Hypertension			0.08 (0.01, 0.15)	0.027
no	564 (57.91%)	2898 (61.70%)
yes	410 (42.09%)	1799 (38.30%)
Hyperglycemia			0.02 (−0.05, 0.09)	0.580
no	857 (87.99%)	4162 (88.61%)
yes	117 (12.01%)	535 (11.39%)
Stroke			0.02 (−0.05, 0.09)	0.627
no	930 (95.48%)	4501 (95.83%)
yes	44 (4.52%)	196 (4.17%)
Dyslipidemia			0.03 (−0.04, 0.10)	0.462
no	781 (80.18%)	3717 (79.14%)
yes	193 (19.82%)	980 (20.86%)

Association between marital status, depressive symptoms, and cognitive function

As shown in Table 2, marital status demonstrated a positive association with cognitive function scores, indicating that married individuals had higher cognitive function scores than unmarried individuals. This association remained statistically significant after adjusting for covariates. In the fully adjusted model, married individuals had an average cognitive function score that was 0.49 points higher than that of unmarried individuals (β = 0.49, 95% CI 0.17, 0.81, p = 0.0023).

Table 2.

The relationship between marital status and cognitive function score.

Variable	Non-adjusted	Adjust I	Adjust II
Unmarried	0	0	0
married	1.88 (1.52, 2.23) < 0.0001	0.79 (0.44, 1.13) < 0.0001	0.49 (0.17, 0.81) 0.0023

Adjust I was adjusted for age, sex, and residential area. Adjust II was adjusted for age, sex, residential area, education, BMI, ever smoked, ever alcohol use, hypertension, hyperglycemia, stroke, and dyslipidemia.

Table 3 presents the correlation between marital status and the CES-D 10 scores. The results indicated that married individuals had lower CES-D 10 scores and less severe depressive symptoms compared to unmarried individuals. In the fully adjusted model, married individuals scored 1.17 points lower on the CES-D 10 scale than unmarried individuals (β = −1.17, 95% CI −1.61, −0.73, p < 0.0001), indicating a negative correlation of depressive symptoms with marital status.

Table 3.

The relationship between marital status and CES-D 10 score.

Variable	Non-adjusted	Adjust I	Adjust II
Unmarried	0	0	0
married	−1.61 (−2.05, −1.18) < 0.0001	−1.33 (−1.78, −0.89) < 0.0001	−1.17 (−1.61, −0.73) < 0.0001

As presented in Table 4, CES-D 10 scores demonstrated a significant negative correlation with cognitive function scores. Higher depressive symptom levels were consistently associated with poorer cognitive performance. This correlation remained statistically significant after adjustment for covariates including marital status. In the fully adjusted model, each 1-point increase in the CES-D 10 score was correlated with a 0.11-point decrease in cognitive function scores (β = −0.11, 95% CI −0.13, −0.09, p < 0.0001).

Table 4.

The relationship between CES-D 10 and cognitive function score.

Exposure	Non-adjusted	Adjust I	Adjust II
CESD10	−0.19 (−0.21, −0.17) < 0.0001	−0.15 (−0.17, −0.13) < 0.0001	−0.11 (−0.13, −0.09) < 0.0001

Mediating role of depressive symptoms

As shown in Table 5 and Figure 1, depressive symptoms (CES-D-10 scores) demonstrated a significant mediating role in the relationship between marital status and cognitive function. The total effect was 0.49 (95% CI: 0.18, 0.83, p = 0.002), indicating a significant positive correlation between marital status and cognitive function scores. Specifically, marital status had a direct effect on cognitive function of 0.36 (β = 0.36, 95% CI: 0.04, 0.69, p = 0.030), and the indirect effect mediated by depressive symptoms was 0.13 (95% CI: 0.08, 0.20, p < 0.0001), explaining 27.39% (95% CI: 14.16%, 79.68%, p = 0.002) of the total effect.

Figure 1.

Mediating model of depressive symptoms.

Table 5.

Mediation effects of the CES-D 10 score between marital status and cognitive function score.

Variable	Estimate	95% CI lower	95% CI upper	p
Total effect	0.491217	0.184450	0.830984	0.0020
Mediation effect	0.134525	0.076667	0.196099	<0.0001
Direct effect	0.356692	0.038498	0.689618	0.0300
Proportion mediated	0.273862	0.141567	0.796791	0.0020

The above results were obtained after adjusting for factors, including age, sex, residential area, education, BMI, ever smoked, ever alcohol use, hypertension, hyperglycemia, stroke, and dyslipidemia.

Mediation analysis stratified by marital status

Mediation analysis revealed substantial heterogeneity in the mediating effects of depressive symptoms (CES-D 10) on cognitive function across marital status groups. Widowed individuals demonstrated a significant indirect effect through depressive symptoms (β = −0.129, 95% CI: −0.195, −0.068), with a mediation proportion of 27.4% (95% CI: 12.6 to 88.0). Cohabiting participants exhibited the largest point estimate for the indirect effect (β = −0.578, 95% CI: −1.212, −0.026), accounting for 29.6% (95% CI: 12.7, 114.7) of the total effect. Never-married individuals showed the most substantial total effect (β = −1.519, 95% CI: −2.813, −0.229), with a mediation proportion of 17.1% (95% CI: 2.2, 100.1). Notably, married individuals not cohabiting with spouses demonstrated the highest mediation proportion (65.0%, 95% CI: −353.4, 733.1), underscoring the predominant role of depressive symptoms in cognitive outcomes for this group. In contrast, separated and divorced groups showed non-significant mediation effects, with confidence intervals spanning the null value. These findings highlight the differential pathways through which marital status influences cognitive health and emphasize the varying importance of depressive symptoms as mediators across distinct marital circumstances (Table 6).

Table 6.

Heterogeneity in the mediation effects across marital status subgroups.

Marital Status	Sample Size	Indirect Effect (ACME) [95% CI]	Direct Effect (ADE) [95% CI]	Total Effect [95% CI]	Proportion Mediated [95% CI]
Married (not living with spouse)	4,692	−0.172 [−0.288, −0.063]	−0.093 [−0.834, 0.623]	−0.265 [−1.023, 0.479]	65.0% [−353.4, 733.1]
Separated	4,540	−0.122 [−0.593, 0.323]	−0.452 [−3.545, 2.952]	−0.575 [−3.760, 3.006]	21.3% [−174.2, 189.1]
Divorced	4,564	−0.249 [−0.567, 0.007]	−0.171 [−1.414, 1.042]	−0.421 [−1.762, 0.803]	59.3% [−354.0, 361.3]
Widowed	5,418	−0.129 [−0.195, −0.068]	−0.341 [−0.690, −0.018]	−0.469 [−0.813, −0.142]	27.4% [12.6, 88.0]
Never married	4,568	−0.259 [−0.494, −0.052]	−1.260 [−2.600, 0.103]	−1.519 [−2.813, −0.229]	17.1% [2.2, 100.1]
Cohabited	4,534	−0.578 [−1.212, −0.026]	−1.374 [−3.914, 0.229]	−1.952 [−4.846, 0.000]	29.6% [12.7, 114.7]

ACME: average causal mediation effect; ADE: average direct effect; CI: confidence interval. The above results were obtained after adjusting for factors, including age, sex, residential area, education, BMI, ever smoked, ever alcohol use, hypertension, hyperglycemia, stroke, and dyslipidemia.

Subgroup analysis and effect modification by demographic factors

As shown in Table 7, the mediating effect of depressive symptoms was consistent across sex and residence, as indicated by non-significant interaction terms and uniformly significant indirect effects (Male: β = 0.224, 95% CI: 0.108, 0.362; Female: β = 0.217, 95% CI: 0.092, 0.336; Rural: β = 0.274, 95% CI: 0.179, 0.381; Urban: β = 0.243, 95% CI: 0.103, 0.402). In contrast, education significantly moderated the association, evidenced by a clear gradient in the total effect. This effect was strongest in illiterate participants (β = 0.704, 95% CI: 0.191, 1.123) and progressively attenuated with higher educational attainment, becoming non-significant among those with a middle school (β = 0.069, 95% CI: −0.710, 0.757) or high school or above (β = 0.115, 95% CI: −0.995, 1.323) education.

Table 7.

Robustness of the mediation effects across key demographic subgroups and tests for differences.

Variable	N (%)	Total effect	p	Direct effect	p	Mediation effect	p	Proportion mediated	p for interaction
Sex									0.8219
Male	3105 (54.75%)	0.713 (0.176, 1.195)	0.004	0.489 (−0.035, 0.986)	0.072	0.224 (0.108, 0.362)	<0.001	0.314 (0.128, 1.131)
Female	2566 (45.25%)	0.708 (0.155, 1.286)	0.016	0.491 (−0.067, 1.04)	0.068	0.217 (0.092, 0.336)	0.004	0.306 (0.104, 1.086)
Residential area									0.1197
Rural	3454 (60.91%)	1.148 (0.729, 1.586)	<0.001	0.874 (0.464, 1.339)	0.000	0.274 (0.179, 0.381)	<0.001	0.239 (0.151, 0.418)
Urban	2217 (39.09%)	0.860 (0.250, 1.471)	<0.001	0.618 (0.038, 1.226)	0.020	0.243 (0.103, 0.402)	0.004	0.282 (0.107, 0.853)
Education									0.128
Illiterate	2697 (47.56%)	0.704 (0.191, 1.123)	0.004	0.525 (0.022, 0.953)	0.036	0.179 (0.099, 0.263)	<0.001	0.254 (0.128, 0.843)
Primary school	1558 (27.47%)	0.392 (−0.163, 1.014)	0.200	0.195 (−0.329, 0.793)	0.508	0.197 (0.069, 0.333)	0.004	0.502 (−5.787, 5.528)
Middle school	951 (16.77%)	0.069 (−0.710, 0.757)	0.864	0.000 (−0.734, 0.725)	0.972	0.069 (−0.114, 0.259)	0.492	0.998 (−2.905, 2.935)
High school or above	465 (8.20%)	0.115 (−0.995, 1.323)	0.844	−0.245 (−1.348, 0.912)	0.680	0.36 (0.064, 0.823)	0.016	3.142 (−10.933, 9.176)

The above results were obtained after adjusting for factors, including age, BMI, ever smoked, ever alcohol use, hypertension, hyperglycemia, stroke, and dyslipidemia.

Discussion

This study revealed the complex correlation between marital status and cognition among older adults in China. The results suggested a significant positive relationship between marital status and cognitive function scores. This association was partially mediated by depressive symptoms (accounting for 27.39% of the total effect), while a substantial direct effect was also observed.

These results were consistent with previous research on the marital status-cognitive function relationship. This alignment not only corroborated the universality of marriage's protective effect but also added evidence from China's elderly population. The marriage protection effect hypothesis suggests that married individuals generally experience fewer physical health issues, less severe depressive symptoms and life stress, and lower morbidity and mortality rates compared to unmarried adults.^26–28 An American study reported higher risks of dementia and cognitive decline among divorced or widowed older adults. Unmarried elderly people face an increased risk of memory and orientation impairment.²⁹ A Japanese survey found that unmarried people were significantly more likely to develop dementia than their married counterparts, a trend that persisted even after adjusting for socioeconomic and lifestyle factors.³⁰ This marriage protection effect may operate through several mechanisms. First, marriage establishes a unique health maintenance system. Couples often demonstrate significant convergence in health habits such as smoking cessation, regular exercise, and adherence to medical advice.^31–33 These shared habits may help reduce the risk of cognitive decline. Second, spousal communication serves as an important form of social interaction that can activate the brain's reward system.^34,35 Moreover, emotional intimacy between spouses can also significantly reduce cortisol levels in stress responses,³⁶ while also enhancing well-being and life satisfaction, reducing the incidence of social isolation and depressive symptoms, and thereby helping to preserve cognitive function.

Our cross-sectional analysis revealed a statistically significant mediating role of depressive symptoms in the relationship between marital status and cognitive function, with the mediation proportion estimated at approximately 27.39%. Additionally, a substantial direct association between marital status and cognitive function was observed. Few studies have examined the effects of past depressive symptoms on marriage and cognition. A survey of older adults in 12 European countries found that marital status may function as a mediating factor, with marriage appearing to attenuate the negative correlation between depressive symptoms and cognitive performance, particularly in verbal fluency.³⁷ Zhang et al. reported that depression and informal social support partially mediated the relationship between cognitive impairment and marital status,³⁸ a finding that aligned with our results. Our study specifically examined depressive symptoms as a mediator while controlling for other known risk factors for cognitive impairment, thereby enhancing the rigor of the findings. Nekehia et al. investigated how gender-based marital power imbalance in Mexico correlates with cognitive ability, with depression serving as a mediating factor.³⁹ Their results indicated that higher marital power of husbands was negatively associated with wives’ cognitive ability, and that wives’ marital power imbalance indirectly influenced cognitive ability through depressive symptoms. This result parallels our observation that depressive symptoms are involved in the association between marriage and cognitive function. A key distinction, however, is that their study specifically examined these dynamics among married individuals, focusing on marital power imbalance. Mental health emerges as a key element connecting marital status with cognitive outcomes. Individuals in stable marriages often benefit from daily emotional support provided by their spouses, which appears to promote psychological well-being and may consequently help reduce the risk of cognitive decline.^40,41 As individuals age, those who are unmarried often experience substantial psychological stress due to social expectations and lack of support. Furthermore, adverse marital statuses such as widowhood, separation, and divorce increase mental pressure and negative emotions, and can even lead to psychological problems such as depressive symptoms, thereby affecting cognitive abilities.⁴² The underlying neurobiological mechanisms may be associated with hypothalamic-pituitary-adrenal axis (HPA axis) dysfunction, neuroinflammation, and hippocampal abnormalities. Patients with depressive symptoms often exhibit HPA axis hyperactivity,^43,44 leading to elevated cortisol levels. Chronic elevation of cortisol is believed to cause damage to brain structures such as the hippocampus, thereby impairing cognitive function.^45,46 Furthermore, depressive symptoms promote microglial activation,^47,48 and increase the secretion of pro-inflammatory cytokines,^49,50 including IL-6 and TNF-α,resulting in neuronal damage and impaired neuroplasticity, which in turn contribute to cognitive decline.^51,52 Structural abnormalities in the hippocampus and disrupted hippocampal functional connectivity have been observed in elderly individuals with depressive symptoms.^53,54 As the hippocampus is a critical region for episodic memory, weakened functional connectivity in this area is likely to contribute to cognitive impairment and increase the risk of dementia.^55–57

In the subgroup analysis, we observed a significant negative total effect of widowhood on cognitive function, a finding consistent with a substantial body of existing literature that confirms bereavement as a critical risk factor for cognitive health in later life. More critically, our analysis revealed that depressive symptoms play a pivotal mediating role in this relationship. This indicates that the detrimental effect of widowhood on cognitive function is not solely attributable to a single pathway but is partially mediated through elevated levels of depressive symptoms. Substantial evidence suggests that widowhood may trigger mental health issues such as depressive symptoms, which can subsequently impair cognitive function.^58–60

This mediating effect remains stable across different sexes and residential areas. However, educational attainment emerges as a significant effect modifier. A clear gradient is observed: the total effect is most pronounced in individuals with only a primary school education and progressively attenuates with higher educational levels, becoming non-significant in the high school and above education groups. This pattern is highly consistent with the Cognitive Reserve Hypothesis.⁶¹ Theoretically, individuals with higher educational attainment possess greater neural resilience or more efficient cognitive networks, which provide a buffer against the negative impacts of psychological distress and its neurobiological consequences.^62–66 This reserve capacity enables them to compensate for cognitive impairment induced by marital adversity, thereby attenuating the observed strength of the association.⁶⁷

Therefore, in developing intervention strategies for older adults with cognitive impairment, we recommend establishing a multidimensional assessment framework. This framework should emphasize social support factors, such as marital status, and integrate early screening for depressive symptoms into routine cognitive assessment processes. For individuals presenting with depressive symptoms and lacking marital support, a personalized comprehensive intervention plan combining marital counseling, antidepressant treatment, and cognitive training is recommended. This multi-targeted intervention strategy may be more effective than a single cognitive intervention in delaying cognitive decline.

Limitations and outlook

First, the cross-sectional design cannot rule out the possibility of reverse causality or bidirectional relationships between marital status and cognitive function. Second, although the cognitive assessment tool used in CHARLS covers the core dimensions similar to those in the MMSE, the absence of a standardized total score may limit direct comparability with other studies. Furthermore, the classification of marital status does not distinguish between the quality of marriage. Marriage is not universally beneficial. Not all married people have positive, supportive partners. There are still negative, unsatisfactory marital states. In such cases, ending an unhealthy marriage through divorce could potentially have a positive influence on cognitive health. Thus, future research should more fully consider the role of marital quality in cognitive outcomes.⁶⁸ Additionally, while our secondary analysis employing detailed marital categories revealed important heterogeneity, the cross-sectional nature of the data still limits causal inference regarding these subgroup-specific pathways. Future longitudinal studies with sufficient power to examine these distinct marital transitions over time are warranted. Finally, despite adjusting for major confounding factors, unmeasured variables may still exist and exert residual confounding effects on the outcome.

Future studies should adopt longitudinal designs with multiple data collection points to clarify the causal connection between marital status, depressive symptoms, and cognitive function. Incorporating assessments of marital quality such as intimacy levels, conflict frequency, and marital satisfaction would help refine the characterization of marriage-related protective factors. Further investigation into multidimensional mechanisms including biological factors, behavioral patterns, and social networks is also recommended to advance this field of research.

Conclusion

Based on our findings, depressive symptoms mediate a substantial proportion of the observed association between marital status and cognitive function in older adults. These results highlight the importance of considering marital status and implementing early screening and intervention of depressive symptoms in clinical strategies aimed at supporting cognitive health in this population.

Footnotes

Acknowledgements

We thank the participants of CHARLS.

ORCID iD

Baohui Jia

Ethical considerations

The China Health and Retirement Longitudinal Study (CHARLS) was approved by the Institutional Review Board at Peking University (Approval Number: IRB00001052-11015). This secondary analysis of the de-identified CHARLS public data was exempt from further ethical review.

Consent to participate

Written informed consent was obtained from all participants by the CHARLS team prior to their inclusion in the original study.

Consent for publication

Not applicable

Author contribution(s)

Guanhua Zong: Conceptualization; Methodology; Writing – original draft.

Ran Li: Methodology; Software; Writing – review & editing.

Yuhang Jiang: Data curation; Formal analysis; Writing – original draft.

Zehao Chen: Investigation; Writing – original draft.

Shanshan Yan: Investigation; Writing – original draft.

Zongxi Yi: Investigation; Validation; Visualization; Writing – original draft.

Xinyu Ren: Investigation; Validation; Visualization; Writing – original draft.

Baohui Jia: Formal analysis; Project administration; Supervision; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the National key research and development program on the Modernization of Traditional Chinese Medicine during the 14th Five-Year Plan Period (Grant No. 2023YFC3503704).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

This study reviewed publicly available datasets. For more information, please visit .

References

Sachdev

Blacker

Blazer

, et al. Classifying neurocognitive disorders: the DSM-5 approach. Nat Rev Neurol 2014; 10: 634–642.

Albert

DeKosky

Dickson

, et al. The diagnosis of mild cognitive impairment due to Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 2011; 7: 270–279.

Aranda

Kremer

Hinton

, et al. Impact of dementia: health disparities, population trends, care interventions, and economic costs. J Am Geriatr Soc 2021; 69: 1774–1783.

Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: an analysis for the Global Burden of Disease Study 2019. Lancet Public Health 2022; 7: e105–e125.

Aljezawi

Kofahi

Abu Khait

, et al. The burden in family caregivers of people living with dementia: prevalence and predictors. Psychogeriatrics 2024; 24: 1132–1138.

Altintas

Luyat

Gallouj

, et al. Burden of family caregivers in Alzheimer’s disease: the role of caregivers’ perception of cognitive impairment. Appl Neuropsychol Adult 2024; 13: 1–7.

Skaria

. The economic and societal burden of Alzheimer disease: managed care considerations. Am J Manag Care 2022; 28: S188–S196.

Penning

. Marital status, childlessness, and social support among older Canadians. Can J Aging 2014; 33: 426–447.

Wyke

Ford

. Competing explanations for associations between marital status and health. Soc Sci Med 1992; 34: 523–532.

10.

Kaplan

Kronick

. Marital status and longevity in the United States population. J Epidemiol Community Health 2006; 60: 760–765.

11.

Bookwala

Marshall

Manning

. Who needs a friend? Marital status transitions and physical health outcomes in later life. Health Psychol 2014; 33: 505–515.

12.

Robards

Evandrou

Falkingham

, et al. Marital status, health and mortality. Maturitas 2012; 73: 295–299.

13.

Lee

Huang

Hilal

. Association between marital status and cognitive impairment in a multi-ethnic Asian population. Neuroepidemiology 2024; 58: 326–334.

14.

Kim

. Gender differences in the link between marital quality and cognitive decline among older adults in Korea. Psychiatry Investig 2021; 18: 1091–1099.

15.

Nerobkova

Park

Shin

, et al. Marital transition and cognitive function among older adults: the Korean Longitudinal Study of Aging (2006-2020). BMC Geriatr 2022; 22: 1003.

16.

Chen

Z-C

Wang

X-D

, et al. Association between marital status and cognitive impairment based on a cross-sectional study in China. Int J Geriatr Psychiatry 2022; 37: 1–9.

17.

Jennings

Farrell

Liu

, et al. Associations between cognitive function and marital status in the United States, South Africa. Mexico, and China. SSM Popul Health 2022; 20: 101288.

18.

Brown

Lin

I-F

Vielee

, et al. Midlife marital dissolution and the onset of cognitive impairment. Gerontologist 2020; 61: 1085–1094.

19.

Yan

X-Y

Huang

S-M

Huang

C-Q

, et al. Marital status and risk for late life depression: a meta-analysis of the published literature. J Int Med Res 2011; 39: 1142–1154.

20.

Semkovska

Quinlivan

O’Grady

, et al. Cognitive function following a major depressive episode: a systematic review and meta-analysis. Lancet Psychiatry 2019; 6: 851–861.

21.

Rock

Roiser

Riedel

, et al. Cognitive impairment in depression: a systematic review and meta-analysis. Psychol Med 2014; 44: 2029–2040.

22.

Christman

Bermudez

Hao

, et al. Accelerated brain aging predicts impaired cognitive performance and greater disability in geriatric but not midlife adult depression. Transl Psychiatry 2020; 10: 317.

23.

Fan

, et al. A chain mediation model reveals the association between depression and cognitive function in the elderly. Sci Rep 2024; 14: 31375.

24.

Andresen

Malmgren

Carter

, et al. Screening for depression in well older adults: evaluation of a short form of the CES-D (Center for Epidemiologic Studies Depression Scale). Am J Prev Med 1994; 10: 77–84.

25.

Hendriks

Ranson

Peetoom

, et al. Risk factors for young-onset dementia in the UK Biobank. JAMA Neurol 2024; 81: 134–142.

26.

Waldron

Hughes

Brooks

. Marriage protection and marriage selection–prospective evidence for reciprocal effects of marital status and health. Soc Sci Med 1996; 43: 113–123.

27.

Rendall

Weden

Favreault

, et al. The protective effect of marriage for survival: a review and update. Demography 2011; 48: 481–506.

28.

Goldman

Korenman

Weinstein

. Marital status and health among the elderly. Soc Sci Med 1995; 40: 1717–1730.

29.

Liu

Zhang

Burgard

, et al. Marital status and cognitive impairment in the United States: evidence from the National Health and Aging Trends Study. Ann Epidemiol 2019; 38: 28–34. e2.

30.

Nakahori

Sekine

Yamada

, et al. Association between marital status and cognitive function in Japan: results from the Toyama Dementia Survey. Psychogeriatrics 2021; 21: 627–635.

31.

Liu

Umberson

. The times they are a changin’: marital status and health differentials from 1972 to 2003. J Health Soc Behav 2008; 49: 239–253.

32.

Phillips

. Lifestyle modulators of neuroplasticity: how physical activity, mental engagement, and diet promote cognitive health during aging. Neural Plast 2017; 2017: 3589271.

33.

Amanollahi

Anjomshoa

, et al. Mitigating the negative impacts of aging on cognitive function; modifiable factors associated with increasing cognitive reserve. Eur J Neurosci 2021; 53: 3109–3124.

34.

Evans

IEM

Martyr

Collins

, et al. Social isolation and cognitive function in later life: a systematic review and meta-analysis. J Alzheimers Dis 2019; 70: S119–S144.

35.

Kawamichi

Sugawara

Hamano

, et al. Increased frequency of social interaction is associated with enjoyment enhancement and reward system activation. Sci Rep 2016; 6: 24561.

36.

Ditzen

Germann

Meuwly

, et al. Intimacy as related to cortisol reactivity and recovery in couples undergoing psychosocial stress. Psychosom Med 2019; 81: 16–25.

37.

Ciziceno

Maggio

. Cognitive functioning and depressive symptoms in the aging population: the buffering role of marital status. Soc Sci Med 2025; 383: 118425.

38.

Zhang

Zheng

. The relationship between marital status and cognitive impairment in Chinese older adults: the multiple mediating effects of social support and depression. BMC Geriatr 2024; 24: 367.

39.

Quashie

Saenz

Sheehan

, et al. Gendered marital power, depression, and cognition among older adults in Mexico. J Gerontol B Psychol Sci Soc Sci 2025; 80: gbae193.

40.

Lee

Back

Kim

, et al. Systematic review of health behavioral risks and cognitive health in older adults. Int Psychogeriatr 2010; 22: 174–187.

41.

Seeman

Lusignolo

Albert

, et al. Social relationships, social support, and patterns of cognitive aging in healthy, high-functioning older adults: macArthur studies of successful aging. Health Psychol 2001; 20: 243–255.

42.

Brown

Lin

I-F

Vielee

, et al. Midlife marital dissolution and the onset of cognitive impairment. Gerontologist 2020; 61: 1085–1094.

43.

Keller

Gomez

Williams

, et al. HPA Axis in major depression: cortisol, clinical symptomatology and genetic variation predict cognition. Mol Psychiatry 2017; 22: 527–536.

44.

Schlosser

Wolf

Wingenfeld

. Cognitive correlates of hypothalamic-pituitary-adrenal axis in major depression. Expert Rev Endocrinol Metab 2011; 6: 109–126.

45.

Rabl

Bartova

Sezen

, et al. HPA Axis in psychotic and non-psychotic major depression: cortisol plasma levels and hippocampal volume. J Affect Disord 2025; 377: 14–22.

46.

Gerritsen

Comijs

van der Graaf

, et al. Depression, hypothalamic pituitary adrenal axis, and hippocampal and entorhinal cortex volumes–the SMART Medea study. Biol Psychiatry 2011; 70: 373–380.

47.

Tang

C-Z

Zhang

D-F

Yang

J-T

, et al. Overexpression of microRNA-301b accelerates hippocampal microglia activation and cognitive impairment in mice with depressive-like behavior through the NF-κB signaling pathway. Cell Death Dis 2019; 10: 316.

48.

Liu

C-P

Zhong

Sun

J-X

, et al. Mir-146a reduces depressive behavior by inhibiting microglial activation. Mol Med Rep 2021; 23: 463.

49.

Guo

Rao

Mao

, et al. Common cellular and molecular mechanisms and interactions between microglial activation and aberrant neuroplasticity in depression. Neuropharmacology 2020; 181: 108336.

50.

Yang

, et al. Stress induced microglial activation contributes to depression. Pharmacol Res 2022; 179: 106145.

51.

Zheng

, et al. Minocycline ameliorates depressive-like behavior and demyelination induced by transient global cerebral ischemia by inhibiting microglial activation. Front Pharmacol 2019; 10: 1247.

52.

Fan

Cao

Fang

, et al. Pro-inflammatory microglia exacerbate high-altitude-induced cognitive impairment by driving lipid droplet accumulation in astrocytes. Antioxidants (Basel) 2025; 14: 918.

53.

Zhou

, et al. Smaller gray matter volume of hippocampus/parahippocampus in elderly people with subthreshold depression: a cross-sectional study. BMC Psychiatry 2016; 16: 219.

54.

Xie

Chen

, et al. Late-life depression, mild cognitive impairment and hippocampal functional network architecture. Neuroimage Clin 2013; 3: 311–320.

55.

Bai

Xie

Watson

, et al. Aberrant hippocampal subregion networks associated with the classifications of aMCI subjects: a longitudinal resting-state study. PLoS One 2011; 6: e29288.

56.

Allen

Barnard

McColl

, et al. Reduced hippocampal functional connectivity in Alzheimer disease. Arch Neurol 2007; 64: 1482–1487.

57.

Alexopoulos

Hoptman

Kanellopoulos

, et al. Functional connectivity in the cognitive control network and the default mode network in late-life depression. J Affect Disord 2012; 139: 56–65.

58.

Wu-Chung

Kennedy

Medina

, et al. Cortical thickness and low-grade inflammation moderate the association between depressive symptoms and cognitive function in early widowhood: a preliminary study. Brain Behav Immun 2025; 129: 620–633.

59.

Biddle

Jacobs

HIL

d’Oleire Uquillas

, et al. Associations of widowhood and β-amyloid with cognitive decline in cognitively unimpaired older adults. JAMA Netw Open 2020; 3: e200121.

60.

Singham

Bell

Saunders

, et al. Widowhood and cognitive decline in adults aged 50 and over: a systematic review and meta-analysis. Ageing Res Rev 2021; 71: 101461.

61.

Roe

Xiong

Miller

, et al. Education and Alzheimer disease without dementia: support for the cognitive reserve hypothesis. Neurology 2007; 68: 223–228.

62.

Sumowski

Chiaravalloti

Wylie

, et al. Cognitive reserve moderates the negative effect of brain atrophy on cognitive efficiency in multiple sclerosis. J Int Neuropsychol Soc 2009; 15: 606–612.

63.

Bradson

Cadden

Riegler

, et al. Cognitive reserve moderates the effects of fatigue and depressive symptoms in multiple sclerosis. Arch Clin Neuropsychol 2023; 38: 1597–1609.

64.

Porricelli

Tecilla

Pucci

, et al. Cognitive reserve modulates mental health in adulthood. Aging Clin Exp Res 2024; 36: 139.

65.

Weiler

Casseb

de Campos

, et al. Cognitive reserve relates to functional network efficiency in Alzheimer’s disease. Front Aging Neurosci 2018; 10: 255.

66.

Meng

D’Arcy

. Education and dementia in the context of the cognitive reserve hypothesis: a systematic review with meta-analyses and qualitative analyses. PLoS One 2012; 7: e38268.

67.

Stern

. How can cognitive reserve promote cognitive and neurobehavioral health? Arch Clin Neuropsychol 2021; 36: 1291–1295.

68.

Haghighi

Littler

EAL

Mauer-Vakil

, et al. Exploring the relationship between marital quality and cognitive function: a systematic review. Soc Sci Med 2024; 355: 117120.