Differential impacts of job-related vs leisure-related physical activity on depressive and cognitive function among middle-aged and elderly adults

Abstract

BACKGROUND

While several studies have explored the relationship among physical activity (PA), depressive symptoms, and cognitive health, the distinct effects of PA performed for occupational vs recreational purposes remain underexplored.

AIM

To investigate the differential impacts of job-related vs leisure-related physical activities on depressive symptoms and cognitive function among adults aged 45 years and older in China.

METHODS

Data were extracted from the China Health and Retirement Longitudinal Study, encompassing 16476 participants. Depressive symptoms were assessed using the 10-item Center for Epidemiologic Studies Depression Scale, and cognitive function was evaluated using the Mini-Mental State Examination. PA was categorized by purpose (job-related vs leisure-related) and intensity (vigorous, moderate, and low), with frequency and duration measured in metabolic equivalent hours per day. An external validation using an independent clinical sample (n = 200) was conducted.

RESULTS

Vigorous-intensity PA for job demands (JVPA) was significantly associated with increased depressive symptoms [P = 0.011, odds ratio (OR) = 1.003], indicating that high-intensity job-related activities may exacerbate mental health issues. Moderate-intensity PA for entertainment or exercise (EMPA) was inversely associated with depressive symptoms (P = 0.030, OR = 0.999), suggesting that moderate-intensity leisure activities can reduce depressive symptoms. For cognitive function, the total PA for job demands was correlated with cognitive decline (P = 0.004, OR = 1.008), with the frequency of JVPA showing a positive association. However, EMPA was linked to reduced cognitive decline (P = 0.018, OR = 0.998). Clinical validation results were consistent with those obtained from the database, further supporting the reliability of the findings.

CONCLUSION

JVPA exacerbates depressive symptoms and cognitive impairment, whereas EMPA mitigates depression and supports cognitive health. Targeted interventions promoting leisure-related PA may enhance mental and cognitive well-being in older adults.

Key Words: Depressive symptoms; Cognitive function; Physical activity; Job-related activity; Leisure-related activity; Aging; China Health and Retirement Longitudinal Study

Core Tip: This study investigates the differential impacts of job-related vs leisure-related physical activities (PA) on depressive symptoms and cognitive function among middle-aged and elderly adults in China, using data from the China Health and Retirement Longitudinal Study. Key findings indicate that vigorous-intensity PA for job demands exacerbates depressive symptoms and cognitive decline, while moderate-intensity PA for entertainment reduces depressive symptoms and supports cognitive health. These results highlight the importance of promoting leisure-based PA to enhance mental well-being and cognitive outcomes in aging populations, suggesting targeted interventions could be beneficial.

INTRODUCTION

As the global population continues to age, understanding the complex effects of physical activity (PA) on mental and cognitive health in older adults has become increasingly important. This issue is particularly relevant in China, which has the largest elderly population in the world. As of 2023, over 21.1% (296.97 million) of individuals are aged 60 years and above according to the “National Ageing Development Bulletin” published by the National Health Commission of China[1,2].

Aging is often accompanied by an increased prevalence of depressive symptoms and cognitive decline, both of which significantly diminish quality of life and impose substantial economic burdens on healthcare systems[3]. Identifying modifiable lifestyle factors such as PA that can mitigate these conditions is of primary relevance for public health. PA has long been recognized as a key determinant of physical, mental, and cognitive health[4,5]. It plays a protective role against several brain age-related conditions, including cognitive decline, neurodegenerative disorders, and overall brain function deterioration[6,7]. Evidence suggests that PA can alleviate depressive symptoms through mechanisms such as improved neuroplasticity, enhanced serotonin utilization, and reductions in systemic inflammation[8]. For cognitive health, PA promotes neurogenesis, increases hippocampal volume, and reduces β-amyloid deposition, thereby mitigating the risk of cognitive impairment and dementia[9].

Despite the wealth of research on PA and its health benefits in elderly populations, much of the existing literature treats PA as a homogeneous construct, failing to differentiate between its various types and purposes. PA for job demands (JPA), often characterized by necessity and high physical demands, may not confer the same benefits as PA for entertainment or exercise (EPA), which is typically voluntary and associated with relaxation and social interaction[10,11]. JPA may even exacerbate stress and depressive symptoms due to its obligatory nature, lack of autonomy, and potential physical strain[11]. Meanwhile, EPA may positively influence mental and cognitive health through its association with psychological well-being and social engagement[12]. However, studies have not systematically explored these distinctions, leaving a significant gap in understanding the nuanced effects of different PA types on health outcomes.

This study aims to address these gaps by examining the differential impacts of job-related vs leisure-related physical activities on depressive symptoms and cognitive function among Chinese adults aged 45 and older. Using data from the nationally representative China Health and Retirement Longitudinal Study (CHARLS), we categorize PA by purpose (job vs leisure) and intensity (vigorous, moderate, low) to explore their associations with mental and cognitive health. While previous studies have focused on general physical activity levels[1,13], to our knowledge, this research is among the first to explore the distinct effects of job-related and leisure-related PA in a large aging population. Our findings highlight the importance of promoting leisure-based PAs for enhancing mental and cognitive health, providing actionable insights for policymakers and healthcare practitioners in designing effective, culturally tailored interventions.

MATERIALS AND METHODS

Data source and study design

The data for this study were obtained from the 2018 CHARLS, which is a longitudinal survey designed to be representative of individuals aged 45 years and older residing in mainland China, covering the period of 2011-2018. This comprehensive dataset encompasses a wide array of information, including socioeconomic status and health conditions, thereby supporting scientific research focused on middle-aged and elderly populations.

All data collected during CHARLS are housed within the CHARLS database at Peking University, China. CHARLS was conducted in accordance with the Declaration of Helsinki and received approval from the Biomedical Ethics Review Committee of Peking University (No. IRB00001052-11015). Informed consent was obtained from all participants prior to their inclusion in the study. The CHARLS team removed direct identifiers prior to data release to ensure participant confidentiality, retaining only anonymized participant IDs for cross-period tracking and matching, thereby preventing any potential tracing back to individuals. Further details regarding CHARLS can be accessed on the official website: http://charls.pku.edu.cn/en/.

In alignment with the objectives of this study, specific exclusion criteria for participant selection were established, which included the absence of data from the 10-item Center for Epidemiologic Studies Depression Scale (CESD-10) questionnaire and cognitive-related assessments. Consequently, a total of 16476 subjects were selected from an initial pool of 19816 participants. Figure 1 presents the flowchart detailing the data selection process.

Figure 1 Flowchart of sample selection. CHARLS: China Health and Retirement Longitudinal Study; MMSE: Mini Mental State Examination; CESD-10: 10-item Center for Epidemiologic Studies Depression Scale.

Variables and measurements

Depressive symptoms: Depressive symptoms were measured by the summary score from the validated Chinese version of CESD-10 (score: 0-30). A threshold of 12 indicated possible clinical depression[14].

Cognitive function: Cognitive functions were measured by the summary score from the Mini Mental State Examination (MMSE; score: 0-30), including simple tasks in several areas: The test of time and place, the repeating lists of words, arithmetic such as serial subtractions of seven, language use and comprehension, and basic motor skills[15]. A threshold of 20 indicated possible moderate or severe cognitive decline.

PA: The level of PA was assessed by measuring different parameters.

Purpose: Participants were asked about the purpose of PA. The answers were classified into two types: (1) JPA; and (2) EPA.

Intensity: PA was classified into three intensity levels: (1) Vigorous-intensity activity (VPA) refers to activities that can cause shortness of breath (e.g., carrying heavy stuff, digging, hoeing, aerobic workout, bicycling at a fast speed, and riding a cargo bike/motorcycle); (2) Moderate-intensity activity (MPA) refers to activities that can make participants’ breathe faster than usual (e.g., carrying light stuff, bicycling at a normal speed, mopping, Tai-Chi, and speed walking); and (3) Low-intensity activity (LPA) refers to activities such as walking. Specifically, JPA was classified into vigorous-intensity JPAs (JVPA), moderate-intensity JPAs (JMPA), and low-intensity intensity JPAs (JLPA). Likewise, EPA was classified into vigorous-intensity EPA (EVPA), moderate-intensity EPA (EMPA) and low-intensity intensity EPA (ELPA).

Frequency: Participants were asked how many days a week they participated in VPA/MPA/LPA for at least 10 minutes. The responses ranged from 0 day to 7 days per week.

Duration Participants were asked how long they do VPA/MPA/LPA every day. The responses included no activity over 10 minutes/day, 10-29 minutes/days, 30-119 minutes/day, 120-239 minutes/day, and ≥ 240 minutes/day. The duration was recorded as 5 minutes/day, 20 minutes/day, 75 minutes/day, 180 minutes/days, and 240 minutes/day separately to calculate the scores of total JPAs (JTPA) and total EPA (ETPA)[16].

JTPA and ETPA scores: The amount of PA was calculated by multiplying frequency by the duration of each PA intensity. Instead of hours per day, the JTPA and ETPA scores were measured in metabolic equivalent (MET) hours per day (hour/day) by using a structured validated questionnaire. According to previous studies[17], 1 MET refers to oxygen consumption at rest, VPA can be expressed as 8 METs, MPA can be expressed as 4 METs, and LPA can be expressed as 3.3 METs. The JVPA score was calculated as the product of the frequency of JVPA, the duration of JVPA and 8 METs, and the JTPA score was calculated as the sum of the JVPA, JMPA, and JLPA scores.

Covariates: The covariates were screened based on basic individual information, including gender, age, residence education, marital status, smoking status, alcohol consumption, and sleep duration.

Retrospective clinical cases

A retrospective clinical case review of 200 adults aged 45 years and older who visited the Department of Geriatrics at our hospital between January 2020 and December 2024 was conducted to further validate the findings from the CHARLS dataset. The inclusion criteria for these cases included the following: (1) Age ≥ 45 years; (2) Availability of complete medical records, including depressive symptoms and cognitive function assessments; and (3) Documented PA levels categorized as job-related or leisure-related.

Data were collected through a structured review of electronic medical records. Depressive symptoms were assessed using the CESD-10 questionnaire, and cognitive function was evaluated using MMSE. PA data were extracted from patient histories, detailing the purpose (job-related vs leisure-related), intensity (vigorous, moderate, low), frequency, and duration of activities. For each participant, the JTPA and ETPA scores were calculated using the same MET-based formula described in the “Covariates” section. Covariates, such as gender, age, educational level, marital status, smoking status, alcohol consumption, and sleep duration, were recorded. This additional dataset aims to provide clinical validation and support the generalizability of the findings from the CHARLS database. The study received approval from the Institutional Review Board and Ethics Committee of our hospital. Given the use of anonymized patient data with no potential for harm, informed consent was waived in accordance with the ethical guidelines for retrospective research.

Statistical analysis

Continuous variables were reported as mean ± SD and categorical variables as the frequency and percentage. Multinomial logistic regression models were used to estimate the association between physical activities for different purposes and depressive symptoms and cognitive function, considering the potential effects of the confounding variables. Two models were established for the main analysis: A multivariate adjustment model adjusted for gender, age, residential status, educational level, marital status, smoking, alcohol consumption, and sleeping duration (model 1) and then adjusted for nap duration (model 2). Nap duration was included in a separate model to isolate its potential confounding effect and assess whether controlling for nap duration significantly alters the relationship between PA and mental health outcomes, ensuring a thorough evaluation of all relevant factors. Some selected covariates were found to have missing values during data processing, but the amount of missing data remained within the acceptable range for statistical analysis. The issue of missing data was addressed during the data processing phase by employing the multiple imputation by chained equations method using a random forest algorithm. This approach allows for the generation of plausible values for missing data, thereby enhancing the robustness and validity of the analysis.

RESULTS

Characteristics of participants

The cross-sectional study included 16476 participants. Among them, 48.1% were male, 55.9% were under 60 years old, 78.7% resided in urban areas, 62.9% had at most a primary school education, 87.2% were married, 42.4% had a smoking history, and 37.5% had a drinking history. The sleep duration was 6.227 ± 1.912, and the nap duration was 40.273 ± 48.302 minutes (Table 1).

Table 1 Characteristics of participants, n (%)/mean ± SD.

Variables	Value (n = 16476)
Male	7929 (48.1)
Age < 60 years	9210 (55.9)
Rural residence	12962 (78.7)
Education level
Primary school or below	10366 (62.9)
Middle school or above	6110 (37.1)
Married	14363 (87.2)
Smoking	6989 (42.4)
Drinking	4539 (37.5)
Sleep duration, hours	6.227 ± 1.912
Nap duration, minutes	40.273 ± 48.302

Healthy status and subjective assessments in study participants

Among the participants, 4.7% experienced adverse health events, and 42.8% reported having no comorbidities. Most participants (98.3%) were independent in their daily activities, and 88.2% reported that their health did not influence their daily life. Self-perceptions varied, with 55.3% rating their health as good, and financial perceptions showed that 78.5% considered their income moderate (Table 2).

Table 2 Healthy status and subjective assessments in study participants, n (%).

Variables	Value (n = 16476)
Adverse health events	774 (4.7)
Reported comorbidity
0	7068 (42.8)
1	6179 (37.5)
2	2323 (14.2)
≥ 3	906 (5.5)
Activities of daily living
Independent	16196 (98.3)
Dependent	280 (1.7)
Whether health influence life
Yes	1927 (11.8)
No	14549 (88.2)
Self-perceived health
Poor	1154 (7.0)
Moderate	6211 (37.7)
Good	9111 (55.3)
Self-perceived income
Poor	1944 (11.8)
Moderate	12933 (78.5)
Rich	1599 (9.7)

Descriptive statistics of depression symptoms in CESD-10

The average score for “everything was an effort” was notably higher at 1.95 ± 0.20, indicating the prevalence of this symptom, whereas restless sleep and trouble concentrating scored 1.55 ± 0.41 and 1.63 ± 0.39, respectively. Participants reported feeling bothered by things (1.59 ± 0.38) and depressed (1.61 ± 0.42) with moderate frequency. Lower frequencies were noted for feeling fearful (1.27 ± 0.36) and hopeless about the future (0.98 ± 0.32), with the lowest score observed for the inability to “get going” (0.95 ± 0.21). Notably, participants reported feeling lonely, with an average score of 1.45 ± 0.58, but maintained a moderate level of happiness (1.48 ± 0.25) (Table 3).

Table 3 Descriptive statistics of depression symptoms in 10-item Center for Epidemiologic Studies Depression Scale, mean ± SD.

Variables	Value (n = 16476)
Everything was an effort	1.95 ± 0.20
Restless sleep	1.55 ± 0.41
Trouble concentrating	1.63 ± 0.39
Bothered by things	1.59 ± 0.38
Depressed	1.61 ± 0.42
Fearful	1.27 ± 0.36
Lonely	1.45 ± 0.58
Happy	1.48 ± 0.25
Could not “get going”	0.95 ± 0.21
Hopeful about future	0.98 ± 0.32

Descriptive statistics of cognitive function scores in MMSE

The mean total cognitive function score was 25.41 ± 2.39, indicating a generally preserved cognitive status among participants. The scores across specific domains revealed varied cognitive abilities, with orientation (to time and place) scoring 4.46 ± 0.55 and registration at 2.31 ± 0.49. The performance in attention and calculation was moderate, with a mean score of 3.28 ± 0.95, and recall ability scored at 2.32 ± 0.89. Language competence was reflected by a score of 1.89 ± 0.29. Repetition and complex commands had lower mean scores of 0.95 ± 0.46 and 1.55 ± 1.35, respectively, suggesting these areas may involve higher cognitive demands (Table 4).

Table 4 Descriptive statistics of cognitive function scores in Mini Mental State Examination, mean ± SD.

Variables	Value (n = 16476)
Total score	25.41 ± 2.39
Orientation (to time, to place)	4.46 ± 0.55
Registration	2.31 ± 0.49
Attention and calculation	3.28 ± 0.95
Recall	2.32 ± 0.89
Language	1.89 ± 0.29
Repetition	0.95 ± 0.46
Complex commands	1.55 ± 1.35

Associations between depressive symptoms and PA frequency and duration for different purposes

Multivariate regression analysis (model 1) revealed that JTPA was significantly associated with depressive symptoms [P = 0.011, odds ratio (OR) = 1.003, 95% confidence interval (CI): 1.001-1.005]. Specifically, the frequency and duration per session of JVPA showed significant associations with depressive symptoms (frequency: P = 0.022, OR = 1.017, 95%CI: 1.002-1.031; duration: P = 0.005, OR = 1.001, 95%CI: 1.000-1.001) in Table 5. ETPA was significantly associated with depressive symptoms (P = 0.011, OR = 1.003, 95%CI: 1.001-1.005), with the duration per session of EMPA showing a significant association (P = 0.030, OR = 0.999, 95%CI: 0.998-1.000). Other PA variables, including frequency and duration of JMPA, JLPA, ELPA, and EVPA and frequency of EMPA, did not demonstrate statistically significant associations with depressive symptoms.

Table 5 Multivariate regression results for depressive symptoms (model 1).

Variables	P value	OR	95%CI
JTPA score	0.011a	1.003	1.001-1.005
JVPA frequency (day/week)	0.022a	1.017	1.002-1.031
JMPA frequency (day/week)	0.619	1.003	0.990-1.017
JLPA frequency (day/week)	0.768	1.002	0.990-1.014
JVPA duration per time (minutes)	0.005b	1.001	1.000-1.001
JMPA duration per time (minutes)	0.164	1.000	1.000-1.001
JLPA duration per time (minutes)	0.782	1.000	0.999-1.000
ETPA score	0.011a	1.003	1.001-1.005
EVPA frequency (day/week)	0.797	1.005	0.970-1.040
EMPA frequency (day/week)	0.109	0.964	0.921,1.008
ELPA frequency (day/week)	0.718	10.992	0.951-1.035
EVPA duration per time (minutes)	0.291	1.001	0.999-1.002
EMPA duration per time (minutes)	0.030a	0.999	0.998-1.000
ELPA duration per time (minutes)	0.064	0.999	0.999-1.000

^aP < 0.05.

^bP < 0.01.

OR: Odds ratio; CI: Confidence interval; JTPA: Total physical activity for job demands; JVPA: Physical activity for job demands in vigorous-intensity; JMPA: Physical activity for job demands in moderate-intensity; JLPA: Physical activity for job demands in low-intensity; ETPA: Total physical activity for entertainments or exercise demands; EVPA: Physical activity for entertainments or exercise demands in vigorous-intensity; EMPA: Physical activity for entertainments or exercise demands in moderate-intensity; ELPA: Physical activity for entertainments or exercise demands in low-intensity.

In model 2, JTPA was significantly associated with depressive symptoms (P = 0.012, OR = 1.003, 95%CI: 1.001-1.005), as was the frequency of JVPA, showing a significant positive association (P = 0.012, OR = 1.003, 95%CI: 1.001-1.005) in Table 6. Additionally, the duration per session of JVPA was significantly related to depressive symptoms (P = 0.005, OR = 1.001, 95%CI: 1.000-1.001). ETPA was inversely associated with depressive symptoms (P = 0.022, OR = 0.992, 95%CI: 0.985-0.999), with EMPA displaying a significant inverse correlation (P = 0.031, OR = 0.999, 95%CI: 0.998-1.000). Other variables, including the frequency and duration of JMPA, JLPA, EVPA, and ELPA and frequency of EMPA, did not show statistically significant associations, suggesting these activities may not be independently influential on depressive symptoms in this cohort.

Table 6 Multivariate regression results for depression symptoms (model 2).

Variables	P value	OR	95%CI
JTPA score	0.012a	1.003	1.001-1.005
JVPA frequency (day/week)	0.012a	1.003	1.001-1.005
JMPA frequency (day/week)	0.629	1.003	0.990-1.017
JLPA frequency (day/week)	0.782	1.002	0.990-1.014
JVPA duration per time (minutes)	0.005b	1.001	1.000-1.001
JMPA duration per time (minutes)	0.168	1.000	1.000-1.001
JLPA duration per time (minutes)	0.778	1.000	0.999-1.000
ETPA score	0.022a	0.992	0.985-0.999
EVPA frequency (day/week)	0.795	1.005	0.970-1.040
EMPA frequency (day/week)	0.109	0.964	0.921-0.008
ELPA frequency (day/week)	0.721	0.992	0.951-1.035
EVPA duration per time (minutes)	0.290	1.001	0.999-1.002
EMPA duration per time (minutes)	0.031a	0.999	0.998-1.000
ELPA duration per time (minutes)	0.670	0.999	0.999-1.000

^aP < 0.05.

^bP < 0.01.

OR: Odds ratio; CI: Confidence interval; JTPA: Total physical activity for job demands; JVPA: Physical activity for job demands in vigorous-intensity; JMPA: Physical activity for job demands in moderate-intensity; JLPA: Physical activity for job demands in low-intensity; ETPA: Total physical activity for entertainments or exercise demands; EVPA: Physical activity for entertainments or exercise demands in vigorous-intensity; EMPA: Physical activity for entertainments or exercise demands in moderate-intensity; ELPA: Physical activity for entertainments or exercise demands in low-intensity.

Associations between cognitive function and PA frequency and duration for different purposes

In this investigation of cognitive function and its association with PA among middle-aged and elderly adults in CHARLS, the multivariate regression analyses (models 1 and 2) highlighted key findings (Tables 7 and 8). JTPA was consistently correlated with cognitive decline (model 1: P = 0.004, OR = 1.008, 95%CI: 1.002-1.007; model 2: P = 0.003, OR = 1.007, 95%CI: 1.002-1.011). The frequency of JVPA showed significant positive associations with cognitive decline in both models (model 1: P = 0.001, OR = 1.046, 95%CI: 1.018-1.075; model 2: P = 0.023, OR = 1.017, 95%CI: 1.002-1.031). Though the duration of JVPA was not statistically significant, the EMPA duration per time illustrated a consistent inverse relationship with cognitive decline in both models (model 1: P = 0.017, OR = 0.998, 95%CI: 0.996-1.000; model 2: P = 0.018, OR = 0.998, 95%CI: 0.996-1.000). Other PA variables, including the frequency and duration of JMPA, JLPA, EVPA, and ELPA and frequency of EMPA, did not demonstrate statistically significant associations, suggesting that these activities may not be as influential on cognitive outcomes in this cohort.

Table 7 Multivariate regression results for cognitive function (model 1).

Variables	P value	OR	95%CI
JTPA score	0.004b	1.008	1.002-1.007
JVPA frequency (day/week)	0.001b	1.046	1.018-1.075
JMPA frequency (day/week)	0.210	0.984	0.959-1.009
JLPA frequency (day/week)	0.990	1.000	0.978-1.022
JVPA duration per time (minutes)	0.107	1.001	1.000-1.001
JMPA duration per time (minutes)	0.693	1.000	0.999-1.001
JLPA duration per time (minutes)	0.315	1.000	1.000-1.001
ETPA score	0.358	0.996	0.984-1.006
EVPA frequency (day/week)	0.397	1.025	0.968-1.085
EMPA frequency (day/week)	0.158	0.949	0.882-1.021
ELPA frequency (day/week)	0.870	0.994	0.927-1.067
EVPA duration per time (minutes)	0.071	1.002	1.000-1.004
EMPA duration per time (minutes)	0.017a	0.998	0.996-1.000
ELPA duration per time (minutes)	0.732	1.000	0.999-1.001

^aP < 0.05.

^bP < 0.01.

OR: Odds ratio; CI: Confidence interval; JTPA: Total physical activity for job demands; JVPA: Physical activity for job demands in vigorous-intensity; JMPA: Physical activity for job demands in moderate-intensity; JLPA: Physical activity for job demands in low-intensity; ETPA: Total physical activity for entertainments or exercise demands; EVPA: Physical activity for entertainments or exercise demands in vigorous-intensity; EMPA: Physical activity for entertainments or exercise demands in moderate-intensity; ELPA: Physical activity for entertainments or exercise demands in low-intensity.

Table 8 Multivariate regression results for cognitive function (model 2).

Variables	P value	OR	95%CI
JTPA score	0.003b	1.007	1.002-1.011
JVPA frequency (day/week)	0.023a	1.017	1.002-1.031
JMPA frequency (day/week)	0.629	1.003	0.990-1.017
JLPA frequency (day/week)	0.782	1.002	0.990-1.014
JVPA duration per time (minutes)	0.105	1.001	1.000-1.001
JMPA duration per time (minutes)	0.712	1.000	0.999-1.001
JLPA duration per time (minutes)	0.315	1.000	1.000-1.001
ETPA score	0.357	0.995	0.984-1.006
EVPA frequency (day/week)	0.795	1.005	0.970-1.040
EMPA frequency (day/week)	0.109	0.964	0.921-0.008
ELPA frequency (day/week)	0.721	0.992	0.951-1.035
EVPA duration per time (minutes)	0.068	1.002	1.000-1.004
EMPA duration per time (minutes)	0.018a	0.998	0.996-1.000
ELPA duration per time (minutes)	0.709	1.000	0.999-1.001

^aP < 0.05.

^bP < 0.01.

OR: Odds ratio; CI: Confidence interval; JTPA: Total physical activity for job demands; JVPA: Physical activity for job demands in vigorous-intensity; JMPA: Physical activity for job demands in moderate-intensity; JLPA: Physical activity for job demands in low-intensity; ETPA: Total physical activity for entertainments or exercise demands; EVPA: Physical activity for entertainments or exercise demands in vigorous-intensity; EMPA: Physical activity for entertainments or exercise demands in moderate-intensity; ELPA: Physical activity for entertainments or exercise demands in low-intensity.

External validation

The retrospective clinical study included 200 clinical participants. Among them, 49.0% were male, 52.5% were under 60 years old, 78.5% resided in rural areas, 63.0% had at most a primary school education, 87.0% were married, 42.5% had a smoking history, and 37.5% had a drinking history. The sleep duration was 6.251 ± 1.873 hours, and the nap duration was 40.306 ± 44.532 minutes (Table 9). Among the clinical participants, 4.5% experienced adverse health events, and 45.0% reported having no comorbidities. Most participants (98.5%) were independent in their daily activities, and 88.0% reported that their health did not influence their daily life. Self-perceptions varied, with 56.0% rating their health as good, and financial perceptions showed that 78.5% considered their income moderate (Table 10).

Table 9 Characteristics of clinical participants, n (%)/mean ± SD.

Variables	Value (n = 200)
Male	98 (49.0)
Age < 60 years	105 (52.5)
Rural residence	157 (78.5)
Education level
Primary school or below	126 (63.0)
Middle school or above	74 (37.0)
Married	174 (87.0)
Smoking	85 (42.5)
Drinking	75 (37.5)
Sleep duration, hour	6.251 ± 1.873
Nap duration, minute	40.306 ± 44.532

Table 10 Healthy status and subjective assessments in clinical participants, n (%).

Variables	Value (n = 200)
Adverse health events	9 (4.5)
Reported comorbidity
0	90 (45.0)
1	75 (37.5)
2	25 (12.5)
≥ 3	10 (5.0)
Activities of daily living
Independent	197 (98.5)
Dependent	3 (1.5)
Whether health influence life
Yes	24 (12.0)
No	176 (88.0)
Self-perceived health
Poor	18 (9.0)
Moderate	70 (35.0)
Good	112 (56.0)
Self-perceived income
Poor	24 (12.0)
Moderate	157 (78.5)
Rich	19 (9.5)

The average score for “everything was an effort” was notably higher at 1.91 ± 0.25, indicating the prevalence of this symptom, whereas restless sleep and trouble concentrating scored 1.54 ± 0.33 and 1.65 ± 0.35, respectively. Participants reported feeling bothered by things (1.54 ± 0.41) and depressed (1.67 ± 0.46) with moderate frequency. Lower frequencies were noted for feeling fearful (1.23 ± 0.34) and hopeless about the future (0.94 ± 0.29), with the lowest score observed for the inability to ‘get going’ (0.97 ± 0.27). Notably, participants reported feeling lonely, with an average score of 1.40 ± 0.52, but maintained a moderate level of happiness (1.45 ± 0.31) (Table 11).

Table 11 Depression symptoms in 10-item Center for Epidemiologic Studies Depression Scale of clinical participants, mean ± SD.

Variables	Value (n = 200)
Everything was an effort	1.91 ± 0.25
Restless sleep	1.54 ± 0.33
Trouble concentrating	1.65 ± 0.35
Bothered by things	1.54 ± 0.41
Depressed	1.67 ± 0.46
Fearful	1.23 ± 0.34
Lonely	1.40 ± 0.52
Happy	1.45 ± 0.31
Could not “get going”	0.97 ± 0.27
Hopeful about future	0.94 ± 0.29

The mean total cognitive function score was 25.47 ± 2.11, indicating a generally preserved cognitive status among participants. The scores across specific domains revealed varied cognitive abilities, with orientation (to time and place) scoring 4.26 ± 0.56 and registration at 2.23 ± 0.43. The performance in attention and calculation was moderate, with a mean score of 3.19 ± 0.74, and recall ability scored at 2.24 ± 0.79. Language competence was reflected by a score of 1.96 ± 0.32. Repetition and complex commands had lower mean scores of 0.92 ± 0.45 and 1.68 ± 1.56, respectively, suggesting that these areas may involve higher cognitive demands (Table 12).

Table 12 Cognitive function scores in Mini Mental State Examination of clinical participants, mean ± SD.

Variables	Value (n = 200)
Total score	25.47 ± 2.11
Orientation (to time, to place)	4.26 ± 0.56
Registration	2.23 ± 0.43
Attention and calculation	3.19 ± 0.74
Recall	2.24 ± 0.79
Language	1.96 ± 0.32
Repetition	0.92 ± 0.45
Complex commands	1.68 ± 1.56

Multivariate regression analysis (external validation) revealed that JTPA was significantly associated with depressive symptoms (P = 0.020, OR = 1.004, 95%CI: 1.001-1.007). Specifically, the frequency and duration per session of JVPA showed significant associations with depressive symptoms (frequency: P = 0.030, OR = 1.015, 95%CI: 1.001-1.029; duration: P = 0.008, OR = 1.001, 95%CI: 1.000-1.002). ETPA was significantly associated with depressive symptoms (P = 0.015, OR = 1.004, 95%CI: 1.001-1.007), with the duration per session of EMPA showing a significant association (P = 0.040, OR = 0.999, 95%CI: 0.998-1.000). Other PA variables, including frequency and duration of JMPA, JLPA, ELPA, and EVPA and frequency of EMPA, did not demonstrate statistically significant associations with depressive symptoms (Table 13).

Table 13 Multivariate regression results for depressive symptoms (external validation).

Variables	P value	OR	95%CI
JTPA score	0.020a	1.004	1.001-1.007
JVPA frequency (day/week)	0.030a	1.015	1.001-1.029
JMPA frequency (day/week)	0.650	1.004	0.989-1.019
JLPA frequency (day/week)	0.800	1.001	0.989-1.013
JVPA duration per time (minutes)	0.008b	1.001	1.000-1.002
JMPA duration per time (minutes)	0.180	1.000	1.000-1.001
JLPA duration per time (minutes)	0.800	1.000	0.999-1.001
ETPA score	0.015a	1.004	1.001-1.007
EVPA frequency (day/week)	0.800	1.006	0.968-1.045
EMPA frequency (day/week)	0.120	0.960	0.915-1.007
ELPA frequency (day/week)	0.750	1.000	0.955-1.047
EVPA duration per time (minutes)	0.300	1.001	0.999-1.003
EMPA duration per time (minutes)	0.040a	0.999	0.998-1.000
ELPA duration per time (minutes)	0.080	0.999	0.999-1.000

^aP < 0.05.

^bP < 0.01.

OR: Odds ratio; CI: Confidence interval; JTPA: Total physical activity for job demands; JVPA: Physical activity for job demands in vigorous-intensity; JMPA: Physical activity for job demands in moderate-intensity; JLPA: Physical activity for job demands in low-intensity; ETPA: Total physical activity for entertainments or exercise demands; EVPA: Physical activity for entertainments or exercise demands in vigorous-intensity; EMPA: Physical activity for entertainments or exercise demands in moderate-intensity; ELPA: Physical activity for entertainments or exercise demands in low-intensity.

JTPA was consistently correlated with cognitive decline (P = 0.005, OR = 1.007, 95%CI: 1.002-1.012). The frequency of JVPA showed significant positive associations with cognitive decline (P = 0.002, OR = 1.040, 95%CI: 1.015-1.066). Though the duration of JVPA was not statistically significant, the EMPA duration per time illustrated a consistent inverse relationship with cognitive decline (P = 0.020, OR = 0.998, 95%CI: 0.996-1.000). Other PA variables, including frequency and duration of JMPA, JLPA, EVPA, and ELPA and frequency of EMPA, did not demonstrate statistically significant associations, suggesting that these activities may not be as influential on cognitive outcomes in this cohort (Table 14).

Table 14 Multivariate regression results for cognitive decline (external validation).

Variables	P value	OR	95%CI
JTPA score	0.005b	1.007	1.002-1.012
JVPA frequency (day/week)	0.002b	1.040	1.015-1.066
JMPA frequency (day/week)	0.250	0.988	0.963-1.014
JLPA frequency (day/week)	0.995	1.000	0.977-1.023
JVPA duration per time (minutes)	0.150	1.001	1.000-1.002
JMPA duration per time (minutes)	0.720	1.000	0.999-1.001
JLPA duration per time (minutes)	0.350	1.000	0.999-1.001
ETPA score	0.400	0.997	0.985-1.009
EVPA frequency (day/week)	0.450	1.020	0.960-1.084
EMPA frequency (day/week)	0.200	0.950	0.880-1.025
ELPA frequency (day/week)	0.900	0.995	0.925-1.070
EVPA duration per time (minutes)	0.100	1.002	1.000-1.004
EMPA duration per time (minutes)	0.020a	0.998	0.996-1.000
ELPA duration per time (minutes)	0.750	1.000	0.999-1.001

^aP < 0.05.

^bP < 0.01.

OR: Odds ratio; CI: Confidence interval; JTPA: Total physical activity for job demands; JVPA: Physical activity for job demands in vigorous-intensity; JMPA: Physical activity for job demands in moderate-intensity; JLPA: Physical activity for job demands in low-intensity; ETPA: Total physical activity for entertainments or exercise demands; EVPA: Physical activity for entertainments or exercise demands in vigorous-intensity; EMPA: Physical activity for entertainments or exercise demands in moderate-intensity; ELPA: Physical activity for entertainments or exercise demands in low-intensity.

DISCUSSION

In analyzing the effects of PA on depressive symptoms and cognitive function among middle-aged and elderly adults using data from CHARLS, this study provides novel insights that enhance the understanding of the nuanced roles that different types and intensities of PA play in mental health and cognitive resilience. The association between depressive symptoms and PA, particularly the differentiation between JPA and EPA, is a compelling aspect of the results. The positive correlation identified between JVPA and depressive symptoms contrasts with much of the existing literature that generally supports a buffering or alleviative effect of PA on depression[18,19]. This discrepancy could be attributed to the psychological and physical stress associated with PA performed out of necessity rather than choice[20]. Vigorous work-related activities may be perceived as less enjoyable and more physically draining, potentially exacerbating stress and fatigue, which could compound depressive symptoms, particularly in middle-aged and elderly adults who may have limited physical capacity to recover from high-intensity demands[21,22].

Meanwhile, EMPA was inversely associated with depressive symptoms, suggesting a potentially protective effect. This is aligned with the broader literature indicating that voluntary, leisure-based PA is more consistently associated with mental well-being[23]. EPA, especially when engaging in social elements, likely provides not only a physical outlet but also emotional and cognitive enrichments such as socialization, fulfillment, and stress reduction[24]. These findings support the notion of exercise as a multi-dimensional intervention in mental health, where the context and nature of the activity may importantly determine the outcome.

The analysis of the relationship between PA and cognitive function revealed that total PA and the frequency of JVPAs were positively associated with cognitive decline. These findings suggest a complex relationship where, while structured, regular vigorous PA can stimulate neurogenesis and enhance synaptic plasticity, the increased stress and fatigue associated with high-intensity job-related PA, particularly when performed under pressure or without sufficient recovery, could counteract these cognitive benefits and contribute to cognitive decline over time[25].

While EMPA showed associations suggesting reduced cognitive decline, the absence of statistically significant associations for other levels and types of PA warrants further discussion. Some forms of PA may only exert subtle influence over cognitive domains, insufficient to manifest in measurable improvements over short durations or in certain contexts[26,27]. In addition, individual differences in health status, cognitive baseline, and lifestyle factors like diet and education could modulate how, and to what extent, physical activities impact cognitive health, possibly masking or interacting with the observed effects within this cohort[28].

The role of lifestyle covariates, such as smoking, alcohol consumption, and sleep, further complicates these relationships[29]. Each of these factors is independently known to affect mental and cognitive health, potentially confounding observed associations and necessitating careful control in analyses. The benefits of PA are possibly more easily realized or magnified in conjunction with healthier lifestyles, whereas adverse habits may diminish or negate potential gains[29].

Mechanistically, the role of biological pathways is worth considering, such as the anti-inflammatory effects of exercise, the modulation of neurotransmitter systems (e.g., increased release of endorphins, serotonin, and dopamine), and improvements in cardiovascular health that reduce the risk of comorbidities affecting mood and cognition[30]. Specifically, EMPA, which was inversely associated with depressive symptoms, may confer greater systemic benefits than high-intensity JVPA, which positively correlated with depressive symptoms. For example, the systemic benefits from moderate-intensity, sustained exercises could be more robust than those of intermittent, high-intensity activities that may instead increase stress markers when performed under pressure or without sufficient recovery[31]. The results of the present study indicated that job-related PA was associated with cognitive decline, potentially due to increased stress and fatigue.

Voluntary leisure-based PA offers additional psychological benefits, such as improved mood and reduced cognitive fatigue. The autonomy in choosing leisure activities enhances their holistic benefits, contributing to mental resilience beyond mere physiological exertion[32]. Given these insights, promoting leisure-based PA can serve as an effective intervention strategy for mental health and cognitive preservation in aging populations. Policymakers and health practitioners should consider individual preferences and contexts when designing PA programs, ensuring that they are accessible, enjoyable, and suitable for all participants.

While this study provides valuable insights into the relationship between different types of PA and their effects on depressive symptoms and cognitive function among middle-aged and elderly adults, several limitations must be acknowledged. The cross-sectional nature of the study restricts the ability to establish causal relationships because only associations at a single time point were observed. The observed associations do not necessarily imply causality. The reliance on self-reported data for PA and other lifestyle factors introduces the potential for recall bias and inaccuracies, which could affect the robustness of the findings. Cultural and socioeconomic factors unique to the Chinese population, such as traditional health beliefs and economic disparities, may influence PA levels and cognitive outcomes, potentially affecting the generalizability of the findings. While the results demonstrated statistical significance, the ORs (e.g., 1.003 and 0.999) suggest very small effect sizes, which may limit the clinical significance of these findings. The absence of data on the intensity or frequency of certain cognitive and lifestyle factors, which were not captured or controlled for, may confound the results. Furthermore, while the study adjusted for several covariates, other unmeasured confounding variables, such as dietary habits and genetic predispositions, may exist, which could influence the outcomes. Future studies could benefit from longitudinal designs with specific measures, such as repeated assessments over time and more comprehensive data collection, to better elucidate these complex relationships.

CONCLUSION

This study suggests that PA associated with job demands, particularly VPAs, may be linked to an increased risk of depression and moderate to severe cognitive decline in middle-aged and elderly adults. Conversely, moderate-intensity PA undertaken for recreational or exercise purposes appears to potentially confer protective effects against depression and cognitive decline in middle-aged and older populations. These findings suggest that distinguishing between the types of PA is important because job-related physical exertion may not serve as a substitute for engaging in leisure or exercise-oriented activities. Future research should explore whether these associations hold over time and investigate the mechanisms underlying these potential relationships. Additionally, targeted interventions aimed at promoting mental health through appropriate PA modalities could be explored.

ACKNOWLEDGEMENTS

The authors extend their heartfelt thanks to all participants of the China Health and Retirement Longitudinal Study team for their dedication to the project. We are deeply appreciative of Peking University for their invaluable support in collecting, organizing, and providing access to the China Health and Retirement Longitudinal Study database, as well as the involvement of all individuals associated with this effort.