Association between subjective cognitive decline and life-space mobility in a community-based elderly adults: a moderated mediation model of depression and perceived social support

Abstract

Background

Life-space mobility (LSM) is a critical health indicator in older adults, the mechanisms underlying the relationship between subjective cognitive decline (SCD) and LSM remain unclear.

Aims

This study examined depression as a mediator between SCD and LSM, and assessed whether perceived social support (PSS) moderates the relationship between SCD and depression among Chinese community-dwelling older adults.

Methods

We seek to elucidate psychosocial mechanisms of the SCD-LSM link and inform targeted intervention strategies. Drawing on a face-to-face interview sample of 287 community-dwelling aged, this cross-sectional study utilised a moderated mediation analysis. Key constructs were evaluated by the Subjective Cognitive Decline Questionnaire-9, 15-item geriatric depression scale, Perceived Social Support Scale, the Life Space Assessment, respectively.

Results

The results showed that SCD was negatively associated with LSM (β = -0.213, p < 0.001). Mediation analysis indicated an indirect association between SCD and LSM through depression (indirect effect = -1.868, 95% CI [-2.825, -1.029]), accounting for 41.6% of the total association. Furthermore, PSS was identified as a significant moderator in the relationship between SCD and depression (β = -0.088, p < 0.05), with a stronger association observed among older adults with lower PSS levels. The interaction term contributed a unique incremental variance of ΔR² = 0.0086 to the model.

Discussion

Establishing a social support system holds promise for improving life-space mobility and alleviating depression among community-dwelling older adults with subjective cognitive decline, thereby enhancing their well-being.

Conclusion

A negative association between SCD and LSM was observed, with depression partially mediating this relationship. Importantly, PSS demonstrated a statistically significant moderating effect on the relationship between SCD and depression, although the effect size was small.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40520-025-03294-z.

Introduction

With aging, the mobility of older adults generally tends to gradually decline [1], manifested in reduced physical activity (PA) and limited social participation. While physiological function assessments like walking speed and balance tests quantify basic motor abilities, they fail to fully capture the daily activity patterns of the elderly population. Life-space mobility (LSM) integrates multidimensional factors such as physiological function, psychological state, and social support into a mobility assessment framework. This concept encompasses not only the spatial range of older adults’ physical activities, but also specific movement patterns (e.g., walking frequency, duration of physical activities) and social engagement (e.g., community socialization, family interactions) within that space. This multidimensional approach offers a new perspective for understanding older adults’ capacity for independent mobility [2, 3].

Limited LSM not only accelerates physiological decline (e.g., muscular atrophy, osteoporosis) and chronic disease risk in older adults, but also reduces their subjective well-being perception through multidimensional pathways by limiting their autonomy in daily activities (e.g., independent shopping and travelling), weakening their ability to participate in society, and exacerbating their psychological stress (e.g., loneliness and loss of self-worth) [4, 5]. To effectively delay the progression of LSM limitations in older adults, further research into the mechanisms of factors influencing LSM is needed to provide a scientific basis for developing targeted prevention strategies.

Subjective cognitive decline (SCD), also known as subjective cognitive complaints or subjective memory impairment, is defined as an early subjective experience of cognitive impairment, manifesting as self-reported deterioration in cognitive functioning [6] without meeting diagnostic criteria for mild cognitive impairment. Epidemiological surveys indicate that approximately 25% of older adults aged ≥ 60 years report SCD symptoms [7], with particularly high prevalence in Asian populations: Chinese older adults demonstrate an average SCD prevalence of 46.4%, while Japanese older adults show rates ranging from 36.1% to 68.6%—both figures significantly exceeding global averages [8]. Notably, older adults with SCD exhibit persistently low physical activity levels coupled with elevated sarcopenia risk (muscle loss) [9, 10], while diminished social engagement which accelerates cognitive-motor decline through isolation-mediated pathways [11]. Although SCD is associated with reduced PA and impaired social engagement, the mechanisms through which SCD affects LSM remain unclear. Based on the above studies, the first hypothesis of this study is as follows:

Hypothesis 1: SCD is negatively associated with LSM.

Cumulative evidence indicates that individuals with SCD demonstrate heightened susceptibility to depression, with SCD functioning as an independent predictor of depression progression [12, 13]. The life-space “cone” model identifies psychosocial factors as crucial factor of LSM, where mood disorders directly constrain individuals’ spatial engagement through reduced behavioral motivation (e.g., willingness to venture outdoors, exploratory confidence). Longitudinal studies demonstrate cognitive decline exacerbates depression [14], which subsequently mediate LSM declines, revealing mood disorders’ critical mediating role in linking cognitive and functional deterioration. Therefore, based on the above studies, the second hypothesis of this study is put forward:

Hypothesis 2: Depression played a significant mediating role between SCD and LSM.

Perceived Social Support (PSS) refers to an individual’s subjective evaluation of emotional care, instrumental assistance, and value recognition available through their social network, functioning as a protective resilience factor that buffers against life stressors and adverse events [15, 16]. Research demonstrates significant correlations between social support levels and SCD [17, 18]. With increased frequency of interpersonal contact, individuals show significantly reduced risks of developing depression [19]. Notably, PSS exhibits prominent moderating effects in the association between psychological stress and health outcomes. For example, Qin et al. found high PSS levels buffer the negative impacts of socioeconomic comparisons on quality of life at the individual level [20]. Similarly, in older populations, PSS demonstrates protective effects against social network reduction caused by social anxiety, improving psychosocial functioning [21]. Combined with previous studies, the third hypothesis of this study is formulated as follows.

Hypothesis 3: PSS exerted a moderating effect on the SCD-depression relationship.

This study aimed to explore the moderated mediation effect of depression and PSS in the relationship between SCD and LSM among community-dwelling older adults. Specifically, we sought to: (a) assess whether SCD could be significantly associated with LSM, (b) test whether depression mediates the SCD-LSM relationship, and(c) examine whether PSS moderates the relationship between SCD and depression. To our knowledge, this is the first study to simultaneously examine the interplay between subjective cognitive decline, depression, and perceived social support in relation to life-space mobility within a comprehensive moderated mediation framework among Chinese community-dwelling older adults. This approach allows for a more nuanced understanding of the psychosocial mechanisms underlying mobility limitation and can inform the development of targeted, multi-component interventions.

Methods

Study design and participants

This cross-sectional study recruited community-dwelling older adults from three communities in China, between July 2024 and March 2025, using a convenience sampling method. All selected communities featured well-developed and comprehensive community facilities. Data were collected through structured questionnaires administered via face-to-face interviews. To mitigate the potential for common method bias, several procedural remedies were implemented during the survey design stage, including: (1) optimizing the item sequence by mixing measures of different constructs (e.g., independent, mediating, and dependent variables) to disrupt surface-level logical connections and reduce participants’ speculation about variable relationships; (2) assuring participants that all data would be used solely for aggregate statistical analysis to minimize social desirability bias; (3) inserting brief rest reminders and attention-check items between measurement sections to facilitate psychological separation and reduce automated responding. All participants provided written informed consent after receiving full disclosure of the study procedures. Inclusion criteria required participants to be: (1) age ≧ 60 years with permanent community residency, (2) capable of normal communication (verbal/written), and (3) willing to provide voluntary consent. Exclusion criteria included: (1) severe sensory impairments or communication barriers, (2) diagnosed psychiatric disorders or cognitive impairment, as determined by a two-stage assessment conducted by a uniform community physician team during the screening period (June to August 2024), which involved cognitive function evaluation using the Chinese version of the MMSE with education-adjusted cutoffs (illiterate ≤ 17, primary school ≤ 20, secondary school or above ≤ 24) and comprehensive psychiatric assessment based on ICD-10 criteria, medical records, and clinical interview, (3) non-permanent residency (< 6 month). A total of 309 questionnaires were distributed, with 287 valid responses retained after excluding incomplete or inconsistent entries, resulting in a 92.9% validity rate. The study protocol was adhered to the principles of the Declaration of Helsinki and approved by the Ethics Committee of Harbin Medical University Daqing Campus (No. HMUDQ 20250521001). This study followed the STROBE guidelines for observational studies.

Measures

Subjective cognitive decline

The 9-item Subjective Cognitive Decline Questionnaire (SCD-Q9) was used to assess SCD symptoms [22]. This scale was developed and validated in a Chinese context by Hao et al., and has good reliability and validity [23]. This scale consists of two dimensions with a total of 9 items. Total scores range from 0 to 9, where higher scores indicate more severe SCD symptoms. A score of ≥ 5 is classified as SCD, while scores < 5 are classified as non-SCD. In the present study, the Cronbach’s α was 0.718.

Depression

Depression was assessed using the 15-item Geriatric Depression Scale (GDS-15) [24]. The GDS-15 scores range from 0 to 15, with higher scores indicating more severe depression; a threshold of ≥ 5 was used to define clinically significant depression. The Cronbach’s α coefficient for this scale was 0.707 in the study.

Perceived social support

Perceived social support (PSS) was assessed using the 12-item Perceived Social Support Scale (PSSS), and was a widely used and well-validated instrument [15, 25]. The scale comprises three dimensions: family support, friend support, and other support. Each item is rated on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree), a total score between 12 and 84, where higher scores reflect greater perceived social support. In this study, the Cronbach’s α for the scale was 0.943.

Life-space mobility

The Life Space Assessment (LSA-C) is an instrument that measures LSM by assessing the distance individuals report moving within the past 4 weeks [26]. Its conceptual framework incorporates three dimensions of mobility in older adults: spatial extent, frequency, and independence. Total scores range from 0 to 120, with higher values reflecting greater life-space mobility. Scores below 60 indicate limited life-space mobility. In the current study, the Cronbach’s α for the scale was 0.853.

Other variables

The sociodemographic characteristics were collected through resident interview. Included age, sex, education, marital status, income/month, smoking/month, drinking/month, sleep time(h)/day, medications, falls/six months, mobility, exercise tolerance, comorbid conditions, objective cognitive functioning (MMSE score).

Data analysis

We used a moderated mediation model to test the mediating role of depression between SCD and LSM and the moderating role of PSS on the relationship between SCD and depression. All statistical analyses were performed in SPSS v 26.0 software for Windows (IBM Corp), and the SPSS PROCESS macro was used to test the mediation, moderation, and moderated mediation hypotheses [27]. Prior to conducting the primary inferential analyses, all continuous variables used in the regression and PROCESS models were standardized (converted to z-scores). This preprocessing step was undertaken to reduce multicollinearity, particularly for the model containing the interaction term, and to facilitate the interpretation of coefficients. Firstly, for descriptive statistics, quantitative data were presented as median (lower quartile, upper quartile). Spearman’s rank correlation test was used to evaluate correlations among SCD, depression, PSS, and LSM. Secondly, the effect of SCD on LSM was assessed using multivariable regression models adjusted for demographic confounders. Finally, to further explore the SCD-LSM relationship, we applied Model 4 (simple mediation) and Model 7 (moderated mediation) from the PROCESS macro. Depression was specified as the mediator, and PSS as the moderator. All mediation and moderated mediation models were adjusted for the following covariates: age, sex, education, marital status, income/month, smoking/month, drinking/month, sleep time(h)/day, medications, falls/six months, mobility, exercise tolerance, comorbid conditions, and MMSE score. Regression coefficients were tested using bias-corrected bootstrap 95% confidence intervals (CIs) with 5,000 resamples. Effects were considered statistically significant if the 95% CI excluded zero. Simple slope analyses were conducted to further characterize the moderation patterns. A Johnson-Neyman (J-N) plot was generated to visualize the region of significance for the moderation effect. Sensitivity analyses were also performed to evaluate the robustness of the mediation and moderation effects. All hypothesis tests were two-sided, with p < 0.05 set as the significance threshold.

Results

Robustness tests

To examine the robustness of our core findings, we conducted a comprehensive sensitivity analysis using hierarchical regression within the same moderated mediation framework. Models were sequentially adjusted for demographic variables in the first step (age, sex), socioeconomic and social support factors in the second step (education, marital status, income/month), health behaviors and medication use in the third step (smoking/month, drinking/month, sleep time(h)/day, medications), physical health and functional status factors in the fourth step (falls/six months, mobility, exercise tolerance, comorbid conditions), and finally objective cognitive functioning based on MMSE score in the full model. The point estimates of the key pathway coefficients remained stable in both magnitude and direction across all successive layers of adjustment. Formal significance testing confirmed that the core findings were robust to the inclusion of these multiple potential confounders.

Common method bias test and multicollinearity test

In this study, Harman’s single-factor test was employed to assess common method bias. The results revealed 17 factors with eigenvalues > 1 in the unrotated exploratory factor analysis. The variance explained by the first factor was 22.37%, below the critical threshold of 40%, suggesting no substantial common method bias. To evaluate multicollinearity, all predictor variables were tested using multiple linear regression, with variance inflation factor (VIF) values < 5, indicating no significant multicollinearity issues.

Descriptive statistics

A total of 287 participants were enrolled in the study, with a median age of 70 years (IQR: 67–75) (Supplementary Table 1). Among them, 44.3% were female. Sociodemographic characteristics revealed that 72.5% had attained secondary education or higher, 85.4% were married, and 48.4% reported a monthly income between 3001 and 5000 RMB. Lifestyle factors showed that 12.2% were current smokers and 9.8% consumed alcohol. Regarding health status, 34.1% reported sleeping less than 6 h per day, and 10.8% were on medications. In terms of physical function, 14.3% reported a fall history within the preceding six months, 12.9% had unsatisfactory mobility and 18.5% had unsatisfactory exercise tolerance, and 31.7% had five or more comorbid conditions. Additionally, the median MMSE score was 26 (IQR: 24–27).

Correlation analysis

According to Spearman’s rank correlation test and univariate linear regression analysis, LSM was significantly correlated with all study variables (Table 1 and Supplementary Table 2). SCD showed a strong positive correlation with depression (ρ = 0.525, p < 0.001), and negative correlations with LSM (ρ= −0.602, p < 0.001) and PSS (ρ= −0.350, p < 0.01). Depression demonstrated strong negative correlations with LSM (ρ= −0.737, p < 0.001) and PSS (ρ= −0.588, p < 0.001). A positive correlation was observed between LSM and PSS (ρ = 0.473, p < 0.001). To further examine the association between SCD and LSM, we performed multivariate linear regression analyses adjusted for demographic confounders (Table 2 Model 1). After adjustment, SCD remained significantly negatively associated with LSM (β = −0.365, B=−4.490, p < 0.001, 95%CI [−5.594, −3.386]).

Table 1.

Descriptive statistics and correlation analysis results of the study variables

Variables	M (Q1, Q3)	SCD	Depression	LSM	PSS
SCD	4.5 (3.5, 6.5)	1
Depression	4.0 (2.0, 6.0)	0.525‡ (0.428, 0.617)	1
LSM	61.5 (38.0, 82.0)	−0.602‡ (−0.676, −0.514)	−0.737‡ (−0.802, −0.660)	1
PSS	56.0 (47.0, 64.0)	−0.350† (−0.460, −0.232)	−0.588‡ (−0.673, −0.487)	0.473‡ (0.363, 0.573)	1

M, median; Q1, lower quartile; Q3, upper quartile; ^*P < 0.05; ^†P < 0.01; ^‡P < 0.001

Table 2.

Analyses of the moderated mediation model

Predictors		Model 1 (LSM)	Model 2 (Depression)	Model 3 (LSM)	Model 4 (depression))
SCD	β	−0.365	0.355	−0.213	0.241
	B [95% CI]	−4.490[−5.594, −3.386]	0.478[0.324, 0.631]	−2.622[−3.612, −1.631]	0.668[0.381, 0.955]
	SE	0.561	0.078	0.5031	0.146
	t value	−8.007	6.124	−5.2118	4.578
	p value	< 0.001	< 0.001	< 0.001	< 0.001
PSS	β				− 0.431
	B [95% CI]				−1.194[−1.451, −0.936]
	SE				0.131
	t value				−9.156
	p value				< 0.001
Depression	β			−0.428
	B [95% CI]			−3.910[−4.634, −3.186]
	SE			0.368
	t value			−10.634
	p value			< 0.001
Int	β				−0.088
	B [95% CI]				−0.243[−0.458, −0.028]
	SE				0.109
	t value				−2.228
	p value				< 0.05
Model Fit
R²		0.621	0.387	0.733	0.535
ΔR²		-	-	-	0.0086
F		27.623‡	10.633‡	43.443‡	17.120‡

Int, SCD*PSS; ΔR², the unique variance in depression explained by the Int; F, the overall significance of each regression model; B, unstandardized coefficient; β, standardized coefficient; CI: Confidence interval

Mediation effect analysis

In our study, we applied Model 4 of the PROCESS macro to assess the mediating role of depression between SCD and LSM (Table 2 model 3). After the total depression score was included in the model, the direct effect of SCD on LSM remained statistically significant (B=−2.622, p < 0.001, 95%CI [−3.612, −1.631]), depression demonstrated a significant negative association with LSM (B=−3.910, p < 0.001, 95%CI [−4.634, −3.186]). Meanwhile, the positive effect between SCD and depression was also significant (B = 0.478, p < 0.001, 95%CI [0.324, 0.631]). The indirect effect value of SCD on LSM through depression was − 1.868 (95%CI [−2.825, −1.029]) (Table 3). All 95% CI for indirect pathways excluded zero, confirming a significant mediation effect. The proportion of the total effect mediated by depression was 41.6%. Therefore, depression had a partial mediating effect on the relationship between SCD and LSM.

Table 3.

Analysis of the mediating effect of depression

	β	B	S.E.	Boot LLCI	Boot ULCI	Relative mediation effect %
Total effect	−0.365	−4.490	0.561	−5.594	−3.386
Direct effect	−0.213	−2.622	0.503	−3.612	−1.631	58.4%
Indirect effect	−0.152	−1.868	0.456	−2.825	−1.029	41.6%

LLCI: lower limit of the confidence interval; ULCI: upper limit of the confidence interval; B, unstandardized coefficient; β, standardized coefficient

Moderated mediation effect analysis

To examine the moderating role of perceived social support (PSS), we incorporated the total PSS score into the model and applied PROCESS macro’s Model 7 to analyze the interaction between SCD and PSS on depression (Table 3 model 4 and Fig. 1). SCD had a significant positive direct effect on depression (β = 0.241, p < 0.001), PSS had a significant negative effect on depression (β = −0.431, p < 0.001). The SCD × PSS interaction term showed a significant negative effect on depression (β = −0.088, p < 0.05), indicating that PSS moderated the SCD-depression relationship. The full statistical output for this moderated mediation analysis (PROCESS Model 7), including coefficients for all predictor variables and covariates, is available in Supplementary Table 4.

Fig. 1.

The Moderated Mediation Model of Subjective Cognitive Decline (SCD) on Life-Space Mobility (LSM). Note: The model depicts depression as a mediator and perceived social support (PSS) as a moderator. Path coefficients are standardized estimates (β). Solid lines represent significant paths. ^*P < 0.05; ^†P < 0.01; ^‡P < 0.001.Path SCD → Depression, β = 0.241, 95%CI [0.138, 0.345]; Path Depression → LSM, β = −0.428, 95%CI [−0.507, −0.349]; Path SCD → LSM (direct effect), β = −0.213, 95%CI [−0.294, −0.133]; Moderation: SCD × PSS → Depression, β = −0.088, 95%CI [−0.165, −0.010]

We examined the conditional effects of SCD on depression by stratifying participants into low (mean − 1 SD) and high (mean + 1 SD) PSS groups. Simple slopes analysis demonstrated a significant but attenuated positive effect of SCD on depression in the high PSS group (effect = 0.420, p < 0.05, 95%CI [0.027, 0.813]), whereas the effect intensified in the low PSS group (effect = 0.940, p < 0.001, 95%CI [0.625, 1.278]) (Supplementary Table 3). Further moderated mediation analysis confirmed that the indirect effect of SCD on LSM through depression was statistically significant across all levels of PSS (Table 4). Specifically, the conditional indirect effects were significant for individuals with low, average, and high levels of PSS, as all 95% bootstrap confidence intervals excluded zero. However, the pairwise contrasts between these effects were not statistically significant, indicating that while the moderation persists at all levels, the differences in effect strength between groups lack statistical support. The point estimate of the index of moderated mediation was positive (index = 0.951), and the effects were numerically stronger in the low PSS group, providing a pattern suggestive of a buffering role (Fig. 2). However, the bootstrap 95% CI for this index included zero [−0.043, 1.926], indicating that the difference in the strength of mediation across PSS levels was not statistically significant.

Table 4.

Conditional indirect effect of PSS on depression at specific levels of PSS

Effect/Contrast	Level of moderator (PSS)	Effect/Contrast value	S.E.	Boot LLCI	Boot ULCI
Conditional indirect	Low(M - SD)	−3.676	1.038	−5.768	−1.699
	Mean	−2.538	0.705	−3.984	−1.205
	High(M + SD)	−1.643	0.726	−3.140	−0.297
Contrast	Mean vs. Low	1.139	0.589	−0.0514	2.307
	High vs. Low	1.033	1.052	−0.092	4.119
	High vs. Mean	0.895	0.463	−0.040	1.812
Index of moderated mediation		0.951	0.492	−0.043	1.926

The index of moderated mediation is the effect of the moderator on the indirect effect; LLCI: lower limit of the confidence interval; ULCI: upper limit of the confidence interval

Fig. 2.

The moderating role of Perceived Social Support (PSS) in the Association Between Subjective Cognitive Decline (SCD) and Depression. Notes: Panel (a) presents the Johnson-Neyman (J-N) plot, which illustrates the region of significance for the moderating effect of PSS. The solid line represents the conditional effect of SCD on depression across a continuum of PSS levels (z-scores). The shaded area represents the 95% confidence band. The vertical dashed lines indicate the specific PSS values at which the effect of SCD on depression transitions between statistical significance and non-significance

Panel (b) shows the simple slopes of the association between SCD and depression at low (Mean − 1 SD), average (Mean), and high (Mean + 1 SD) levels of PSS. The association was stronger at low levels of PSS (simple slope = 0.940, p < 0.001) and weaker at high levels of PSS (simple slope = 0.420, p < 0.05). Detailed statistics for the simple slopes are presented in Supplementary Table 3.

Notably, the interaction effect was statistically significant yet small in effect size (observed ΔR² = 0.0086). While this moderating effect of PSS was statistically significant but small in magnitude. A post-hoc sensitivity power analysis was conducted using G*Power for a linear multiple regression test (R² increase) with the following parameters: sample size (N) = 287, α = 0.05, power = 0.80, number of tested predictors = 1 (for the SCD×PSS interaction term), and total number of predictors = 15. This analysis indicated a minimum detectable effect size of ΔR² ≈ 0.027. This discrepancy underscores two important methodological considerations: (1) our study was underpowered to detect very small interaction effects, representing a limitation for interpreting the null contrast findings; and (2) the fact that this interaction nevertheless reached statistical significance suggests that the true effect may be stronger than observed, or that our model specification effectively reduced error variance, thereby enhancing detection efficiency. This lends credibility to the reliability of the reported interaction effect.

These results indicate that depression mediates the relationship between SCD and LSM regardless of social support levels. While the observed trend is theoretically interesting, the current analysis cannot conclusively determine that PSS moderates (i.e., significantly alters the strength of) this mediating pathway. Future research with greater power is needed to investigate this potential moderation.

Discussion

This study explored the relationship between SCD and LSM among community-dwelling older adults, as well as the effect of depression and PSS in both. The results showed that SCD was negatively associated with LSM. Depression partially explained the association between SCD and LSM. PSS moderated the relationship between SCD and depression, specifically demonstrating that the association of SCD on depression varied in intensity depending on individuals’ PSS levels. Our results elucidate potential mechanisms behind LSM limitations while highlighting psychological factors’ critical role. There is an urgent need to enhance mental health monitoring for community-dwelling older adults, particularly emphasizing high-risk subgroups exhibiting depression and low social support levels. Due to the cross-sectional design, causal inferences cannot be made from these associations. Notably, 47.7% had limited life-space mobility, which is slightly higher than that reported in previous studies of general older adult populations [28, 29], which may suggest that community-dwelling older adults reporting SCD face relatively greater challenges in maintaining their LSM. This observation warrants increased attention in both clinical practice and community health management.

Our study found that SCD was negatively associated with LSM, which is consistent with our Hypothesis 1. Previous studies have established associations between LSM and cognitive impairment [30, 31], while our research provides novel findings showing an association between SCD and LSM even before overt objective cognitive decline. This finding aligns with Rotenberg et al.’s [32] observation of altered activity participation patterns in older adults with SCD, specifically manifesting as significantly lower performance in instrumental activities of daily living (IADL), leisure activity engagement, and social interaction frequency compared to healthy controls. This suggests that SCD may restrict daily behavioral patterns through functional activity restrictions while accelerating social withdrawal processes. Concurrently, neurobehavioral studies suggest that LSM limitations in older adults with SCD may be linked to alterations in brain regions critical for spatial navigation and decision-making (e.g., the hippocampus and basal forebrain) [33]. While our data cannot directly test this neural mechanism, it represents a plausible hypothesis for future investigation. SCD further diminishes individuals’ adaptive efficiency to environmental demands, promoting “safety behavior” strategies that reinforce an unhealthy cycle of LSM restriction.

The present study showed that depression played a partial mediating role in the relationship between SCD and LSM, which is consistent with our Hypothesis 2. It should be noted that this represents statistical mediation rather than causal mediation. Among community-dwelling older adults, SCD correlates with subclinical depression [34, 35]. Lee et al. [36] found that migraine patients with SCD reported higher depression severity compared to those without SCD, demonstrating depression’s association with elevated SCD risk. From a neurocognitive perspective, depression may diminish planning/decision-making capacities and reduce environmental exploration motivation by impairing goal-directed executive functions [37]. We hypothesize that this could involve dysfunction in prefrontal cortical regions, but this mechanistic pathway requires direct validation in future studies combining neuroimaging and behavioral measures. According to Beck’s cognitive theory of depression, depressive states involve negative perceptions of self, environment, and future. These neural deficits exacerbate individuals’ maladaptive interpretations of self-capacity (e.g., “I always forget where I put things”), environmental challenges (e.g., “Complex roads might make me get lost”), future expectations (e.g., “Going out won’t benefit me”). Such cognitive biases further trigger self-restriction behaviors (actively narrowing activity ranges to home surroundings), indicating that community-dwelling older adults’ distorted self-appraisal of cognitive capacity intensifies social withdrawal patterns.

This study revealed that the level of PSS moderated the association between SCD and depression, which is consistent with our Hypothesis 3. Santini’s longitudinal study demonstrates that lack of social support exerts significant effects on depression among community-dwelling older adults [38].The moderating role of PSS may involve emotional regulation, potentially enhancing cognitive-emotional resilience through stress-processing neural pathways, though this requires specific validation in SCD populations [39]. Critically, as a subjective experience, PSS reflects feelings of loneliness and perceived lack of support, impacting mental health beyond objective social metrics. Neurostructurally, higher PSS is associated with white matter integrity in tracts supporting emotion regulation [40], which may underpin the tendency for high-PSS individuals to employ proactive coping strategies against SCD. Behaviorally, higher PSS correlates with broader life-space mobility, as seen in dementia caregivers who report more outings and greater activity ranges [41].

A nuanced finding was the non-significant difference in mediation strength between high and low PSS groups. This challenges a linear buffering model and instead supports the optimal matching theory [42], which emphasizes the fit between support type and recipient needs. Empirical work confirms that support which fails to match personal preferences (such as in chronic pain populations) correlates with higher depression, highlighting the importance of contextual relevance over mere availability [43].

We hypothesize that the attenuated buffering in high-PSS individuals may stem from such a mismatch. Those with high PSS likely expect support that aligns with their specific preferences. If received support is perceived as generic, overprotective, or limiting autonomy, it may undermine self-efficacy and respect, potentially leading to activity restriction to avoid being a social burden [44]. Thus, mismatched high PSS may not enhance resilience and could even increase mobility fragility by prompting maladaptive coping. Future research should therefore move beyond global support measures to examine the qualitative alignment between support provided and desired, which is crucial for developing personalized mobility interventions for older adults with SCD.

The strength of this study lies in exploring the impact of SCD on LSM and constructing a moderated mediation model to explain their relationship. Several limitations of this study also should be noted. First, the cross-sectional design precludes causal inferences, and the modest sample size may limit the generalizability of the findings and reduce power to detect small effects. Future longitudinal studies with larger samples are needed to establish temporal precedence and improve the robustness of the observed associations. Second, the use of a convenience sample from three urban communities in Daqing, Heilongjiang Province, limits the generalizability of our findings. Although the selected communities represented different urban functional areas (traditional old town, new residential-commercial area, and industrial residential zone), participants exhibited relative homogeneity in socioeconomic status, with 72.5% possessing junior high school education or higher and generally higher economic status. This may limit the applicability of our results to populations from geographically diverse regions, thus requiring validation in more heterogeneous urban-rural cohorts. Third, we handled missing data using listwise deletion. While this approach is justified by the high data completeness (92.9% validity rate) and invokes the missing completely at random (MCAR) assumption, it may nonetheless introduce potential selection bias. Finally, although the current study examined key psychological and cognitive factors, the limited inclusion of behavioral, functional, social, and environmental variables constrains a more comprehensive understanding of the mechanisms underlying LSM limitations. Specifically, the absence of objectively measured physical activity (e.g., accelerometer data), detailed assessment of pain intensity, standardized evaluation of IADLs, as well as factors such as anxiety, social isolation and sensory impairments may have omitted important confounding variables and mediating pathways. It is important to note that this study primarily focused on the depression pathway and did not examine other important mechanisms such as anxiety, pain, and social isolation, which represents a key direction for future research to address.

Furthermore, it must be acknowledged that LSM is inherently shaped by contextual factors, particularly characteristics of the built environment such as neighborhood safety, transportation accessibility, and urban design. The current study’s framework, while focusing on psychological and individual-level predictors, did not incorporate objective measures of these environmental determinants (e.g., GIS-based walkability indexes, systematic audits of neighborhood infrastructure). This omission narrows the conceptual scope of our analysis. Future research would greatly benefit from adopting a multi-level framework that concurrently examines individual-level functional and psychosocial measures alongside neighborhood-level environmental data. Based on the present findings and limitations, future research should prioritize: (1) longitudinal or intervention designs to establish temporal precedence and causality; (2) the incorporation of objective behavioral (e.g., accelerometry), physiological, and environmental measures (e.g., neighborhood walkability) to complement self-report data; and (3) the investigation of other potential mediators (e.g., anxiety, pain, social isolation) to build a more comprehensive theoretical model of life-space mobility restriction.

Given the significant association between LSM and health outcomes in aging populations, systematically investigating the dynamic trajectories of LSM changes and their regulatory mechanisms should be prioritized as a key focus area for geriatric public health interventions. Our findings highlight the practical importance of early screening for subjective cognitive decline and enhancing perceived social support in community health strategies for older adults. Primary prevention could involve annual SCD screening using tools such as the SCD-Q9, combined with community programs designed to bolster social support, such as mindfulness-based group training or social facilitation initiatives. For individuals identified with SCD and low PSS, secondary interventions such as social prescribing, which connects them to tailored community activities, could be beneficial. Tertiary strategies could monitor LSM trajectories using wearable technology to trigger early interventions. Implementing such a multi-tiered prevention framework could help maintain mobility and well-being, delaying functional decline through early identification and personalized support.

Conclusions

This study reveals a significant independent association between SCD and LSM restriction in community-dwelling older adults, while revealing the psychopathological pathway through which SCD exacerbates LSM restriction via the indirect association of depression. We further validate PSS as a modifiable buffer in the relationship between SCD and psychological distress, thereby providing preliminary empirical support for the interconnected “SCD-psychological-function” framework.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

YL and LM led the overall study conceptualization. HB, QX, and HK performed data collection. YL and SZ conducted data analysis and interpretation. YL, SZ, and QX drafted the initial manuscript. LM and HB critically revised the manuscript for important intellectual content. All authors read and approved the final version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (Grant numbers 72204066) and the Philosophy and Social Science Research in Heilongjiang Province (Grant numbers 21RKC212).

Data availability

The datasets generated or analyzed during the current study are not publicly available owing to decisions of the Ethics Committee but are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

The study protocol was adhered to the principles of the Declaration of Helsinki and approved by the Ethics Committee of Harbin Medical University Daqing Campus (No. HMUDQ 20250521001). This study followed the STROBE guidelines for observational studies.

Informed consent

All participants provided written informed consent after receiving full disclosure of the study procedures.