ABSTRACT
Background
Sleep disorders co‐existing with cerebral small vessel disease (CSVD) are prevalent in older adults, both of which are established risk factors for cognitive decline. Here, we aimed to explore the role of glymphatic function in linking CSVD and sleep quality to cognitive decline in community‐dwelling older adults from a new perspective.
Methods
This cohort study included participants from the Shanghai Aging Study (SAS) who underwent clinical interviews, neuroimaging, and neuropsychological assessments. Diffusion tensor image analysis along the perivascular space (DTI‐ALPS) index was employed to evaluate the glymphatic function. Mediation and interaction analyses were performed to investigate the potential mediating role of DTI‐ALPS in the associations between CSVD burden/sleep quality and cognitive impairment, as well as their interaction effects.
Results
258 participants were included in cross‐sectional analysis (mean age 68.5 years, 54.7% female), with 133 followed up after a 7‐year interval. At baseline, DTI‐ALPS index simultaneously mediated the associations between CSVD score/PSQI and cognitive impairment. Interaction analyses revealed that poor sleep quality had a more significant impact on cognitive impairment among participants at high risk of CSVD. Longitudinally, while no significant mediating effect was observed, both baseline DTI‐ALPS and CSVD burden were significantly correlated with longitudinal MMSE changes.
Conclusions
Our study suggested that glymphatic function, as assessed by DTI‐ALPS, may play a crucial role in linking CSVD burden and sleep quality to cognitive decline in community‐dwelling older adults. We also emphasized the importance of individualized sleep management for individuals at high risk of CSVD.
Keywords: analysis along the perivascular space, cerebral small vessel disease (CSVD), cognitive impairment, glymphatic system, sleep
Cohort study of 258 community‐dwelling older adults to explore the role of glymphatic function in linking CSVD and sleep quality to cognitive decline using DTI‐ALPS. DTI‐ALPS simultaneously mediates the relationship between CSVD burden/sleep quality and cognitive performance. CSVD burden and sleep quality interact to exacerbate cognitive impairment.
1. Introduction
Cerebral small vessel disease (CSVD) and sleep disorders are common in older adults, and both represent independent risk factors for cognitive decline [1, 2]. CSVD is characterized by imaging markers, including white matter hyperintensities (WMH), lacunes, and cerebral microbleeds (CMBs), all of which have been consistently associated with cognitive impairment [3]. Although the relationship between enlarged perivascular spaces (EPVS) and cognition remains inconclusive, EPVS represent a key structural component of the glymphatic system and have demonstrated significant correlations with sleep efficiency [4, 5]. Sleep disturbances promote β‐amyloid (Aβ) accumulation in the brain, a well‐documented pathological process in cognitive impairment [2, 6].
Glymphatic system facilitates the exchange of cerebrospinal fluid (CSF) and interstitial fluid (ISF) within perivascular space (PVS) through Aquaporin‐4 (AQP4) located at the endfeet of astrocytes to remove metabolic waste such as Aβ [7, 8]. Previous studies have demonstrated notable overlap in mechanisms between glymphatic dysfunction and CSVD. Chronic hypertension and diabetes, recognized as established risk factors for CSVD, can disrupt glymphatic function by affecting arterial pulsation, which is a primary driver of glymphatic flow [9, 10, 11]. Additionally, the glymphatic system is proven to be active during sleep, with the clearance rate of Aβ during sleep being twice as fast as that during wakefulness [12, 13]. Notably, the frequent co‐occurrence of CSVD and sleep disturbances in older adults emphasizes the potential key role of glymphatic function in linking these factors with cognition. However, their intercurrent relationship and combined impact remain underexplored.
Therefore, in the current cohort study, we hypothesize that the glymphatic function acts as a crucial link connecting the two risk factors—sleep and CSVD—with cognitive decline in older adults. In recent studies, the diffusion tensor imaging analysis along the perivascular space (DTI‐ALPS) index has been widely used as a non‐invasive imaging marker to evaluate the function of the glymphatic system [14]. Based on this, we employed DTI‐ALPS as a preliminary proxy to explore the potential role of glymphatic dysfunction and further investigate the interactive relationship between sleep quality and CSVD burden. We here expect to provide new insights into the mechanisms underlying cognitive deterioration in older adults and help to identify potential targets for intervention.
2. Method
2.1. Study Design, Participants, Standard Protocol Approvals, and Consent
This study was based on a magnetic resonance imaging (MRI) sub‐cohort of the Shanghai Aging Study (SAS), a prospective longitudinal cohort of older adults aged ≥ 60 years in the Jingansi community, Shanghai, China [15]. From 2010 to 2011, 350 participants were consecutively enrolled from SAS who agreed to undergo MRI scanning. Exclusion criteria included: (1) cerebral large‐vessel stenosis > 50%; (2) dementia, stroke, hydrocephalus, or brain tumors; (3) contraindications to MRI. Following further exclusion criteria of this study, 258 participants were included in the cross‐sectional analysis. During the period of 2016–2017, 133 participants were followed up with repeated MRI scanning and neuropsychological assessments. Figure 1 depicts the detailed process of inclusion and exclusion.
FIGURE 1.
Flowchart of participant recruitment. DTI, diffusion tensor imaging; SAS, Shanghai Aging Study.
This study was approved by the Medical Ethics Committee of Huashan Hospital, Fudan University (no. 2009‐195). The procedures used in this study adhered to the tenets of the Declaration of Helsinki. All participants and/or their legal guardians provided written informed consent.
2.2. Clinical Assessment
Demographic and clinical data, including age, gender, education level, and vascular risk factors (VRFs), were collected using a standardized clinical research registry form by trained neurologists. A composite VRFs score was defined as the cumulative count of self‐reported or diagnosed hypertension, diabetes, hyperlipidemia, and past/current smoking, ranging from 0 to 4 [16]. The Apolipoprotein E (APOE) genotype was assessed by extracting DNA from participants' blood or saliva samples, and the presence of at least one ε4 allele was considered APOE‐ε4 positive.
2.3. Measurement of Sleep Quality
Pittsburgh Sleep Quality Index (PSQI) was used to assess the sleep quality of older adults. It comprises 18 self‐rated items across 7 components: subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, use of sleep medications, and daytime dysfunction. Each component is scored from 0 to 3, with a total score ranging from 0 to 21; higher scores indicate poorer sleep quality [17]. In previous studies, a total score ≤ 5 indicated healthy sleep quality, while > 5 indicated poor sleep quality [18, 19].
2.4. Brain MRI Acquisition
The baseline MRI scanning was performed using a 1.5T GE scanner. The imaging protocols included T1‐weighted (T1WI), T2‐weighted (T2WI), fluid attenuated inversion recovery (FLAIR), diffusion tensor imaging (DTI), and T2*‐weighted gradient recall echo (T2* GRE). The parameters for the DTI sequence were as follows: TR/TE = 8000 ms/80.1 ms, slice thickness = 5 mm, flip angle = 70°, number of diffusion directions = 27, acquisition matrix = 128 × 128 mm2, and b‐values = 0 and 1000 s/mm2. Phase encoding was performed in the anterior–posterior (AP) direction, and both AP and posterior–anterior (PA) phase encoding directions were acquired to enable topup correction for susceptibility‐induced distortions. Eddy current correction was also applied to minimize distortions caused by eddy currents and head motion. The detailed parameters of other sequences can be found in our previous studies [20, 21].
2.5. Evaluation of CSVD MRI Markers
CSVD MRI markers were evaluated based on the Standards for Reporting Vascular Changes on Neuroimaging (STRIVE‐2) criteria [22] by two certified neurologists, who were blinded to clinical data. Any discrepancies were resolved by a senior neurologist. The number of lacunes was counted manually on FLAIR/T1/T2‐weighted images. WMH volume was automatically segmented and quantified using a deep learning algorithm based on U‐net detailed in our previous study [21]. To normalize WMH volumes for head size differences across participants, total intracranial volume (TIV) was calculated by summing total brain volume, sulcal volume, and ventricular CSF volume, as described before [21]. The number of CMBs was counted manually on T2*‐GRE scans according to the Microbleed Anatomical Rating Scale (MARS). EPVS were evaluated using Potter's scale, which semi‐quantitatively graded EPVS in basal ganglia (BG) and centrum semiovale (CS) from 0 to 4: grade 0 represented no EPVS, grade 1 represented 1–10 EPVS, grade 2 represented 11–20 EPVS, grade 3 represented 21–40 EPVS, and grade 4 represented ≥ 40 EPVS. The hemisphere with a higher score represented the overall EPVS burden [23].
The total CSVD score incorporated above MRI markers to assess the overall CSVD burden, ranging from 0 to 4 [24]. One point was assigned for each of the following: (1) ≥ 1 lacunes, (2) Deep WMH score ≥ 2 and/or periventricular WMH score = 3 on the Fazekas scale, (3) ≥ 1 CMBs, and (4) severe EPVS (grades 2–4) in BG.
2.6. DTI‐ALPS Evaluation
The calculation of the DTI‐ALPS index was carried out on DTI Studio (Version 3.0.3. Radiology Department, Johns Hopkins University, https://www.mristudio.org) according to previously published methods (Figure S1) [14, 25]. The ALPS index was calculated as [(Dxxproj + Dxxassoc)/(Dyyproj + Dzzassoc)] for each hemisphere, with the global DTI‐ALPS index representing the mean value of bilateral measurements. The processing and calculation were independently conducted by two trained raters. The inter‐rater reliability assessment demonstrated good agreement, with an intraclass correlation coefficient (ICC) of 0.89.
To avoid the confounding impact of DTI metrics on the interpretation of the DTI‐ALPS index, we further calculated the whole‐brain white matter median mean diffusivity (MD) and peak width of skeletonized MD (PSMD) for each participant, following the methodology established in previous research [26].
2.7. Neuropsychological Assessments
At baseline, participants underwent neuropsychological assessments within 1 week of their MRI scanning, including (1) Mini‐Mental State Examination (MMSE) for global cognition; (2) Auditory Verbal Learning Test (AVLT) or Huashan Object Memory Test (HOMT) for memory, with lower scores indicating poorer memory; (3) Trail Making Test (TMT) for executive function, with longer completion times indicating poorer executive function; (4) Stick test (ST) for visual spatial, with lower scores indicating worse visuospatial ability; (5) a modified version of Common Objects Sorting Test (COST) for language, with lower scores indicating poorer performance. Detailed methods were reported in our previous research [27].
Follow‐up assessments were conducted following the same protocols by the same team of trained neurologists as at baseline. Participants were ideally assessed by the same neurologist at both time points. Cognitive changes were quantified as the difference in scores from baseline to follow‐up (e.g., △MMSE = (follow‐up MMSE score)–(baseline MMSE score)).
2.8. Statistical Analysis
Statistical analyses were conducted on SPSS Statistics (version 27.0, IBM), R (version 4.2.1), and GraphPad Prism (version 10.0.0). Continuous variables were presented as mean ± SD (normal distribution) or median (IQR) (ordinal/non‐normal distribution), while categorical variables were reported as frequency (%). Partial correlation analyses were conducted to investigate the relationships among PSQI/CSVD burden, DTI‐ALPS, and cognition in cross‐sectional analyses, and the Spearman method was employed. Model 1 adjusted for age, gender, VRFs, APOE‐ε4 carrier status, PSMD, and education (for cognition), while Model 2 additionally adjusted for median MD.
Mediation analyses were conducted using the simple mediation model in PROCESS (version 4.2), with 5000 bootstrap resamples for confidence intervals. The linear regression model and general linear model (GLM) were employed to investigate the interaction effects between CSVD burden and sleep quality on cognition. Longitudinally, partial Spearman correlation and mediation analyses were used to explore the relationships between baseline PSQI/MRI markers and longitudinal cognitive changes, additionally adjusting for follow‐up time. A two‐sided p < 0.05 was considered statistically significant. The false discovery rate (FDR) correction method was used to correct for multiple comparisons.
3. Results
3.1. Baseline Characteristics of Study Participants
At baseline, 258 participants were included (54.7% female; mean age 68.5 ± 5.9 years; mean education 11.6 years). Prevalent vascular risk factors included hypertension (53.5%), hyperlipidemia (38.0%), diabetes (14.0%), and smoking (13.6%). 39 (15.1%) carried at least one APOE‐ε4 allele, and 84 (32.6%) had poor sleep quality (PSQI score > 5). Additional baseline information is provided in Table 1.
TABLE 1.
Comparison of baseline demographic, clinical, neuroimaging, and cognitive characteristics between healthy sleeper and poor sleeper.
| Total (n = 258) | Healthy sleeper (n = 174) | Poor sleeper (n = 84) | p | |
|---|---|---|---|---|
| Demographics | ||||
| Age, mean (SD) | 68.5 (5.9) | 68.0 (5.6) | 69.5 (6.4) | 0.329 |
| Female, n (%) | 141 (54.7) | 90 (51.7) | 51 (60.7) | 0.174 |
| Education (years), median (IQR) | 12 (6) | 12 (6) | 12 (4) | 0.078 |
| APOE‐ε4 carrier, n (%) | 39 (15.1) | 25 (14.4) | 14 (16.7) | 0.629 |
| Vascular risk factors | ||||
| Smoker, n (%) | 35 (13.6) | 20 (11.5) | 15 (17.9) | 0.162 |
| Hypertension, n (%) | 138 (53.5) | 88 (50.6) | 50 (59.5) | 0.177 |
| Diabetes, n (%) | 36 (14) | 25 (14.4) | 11 (13.1) | 0.782 |
| Hyperlipidemia, n (%) | 98 (38) | 65 (37.4) | 33 (39.3) | 0.765 |
| VRFs, median (IQR) | 1 (1) | 1 (2) | 1 (1) | 0.213 |
| Imaging characteristics | ||||
| DTI‐ALPS, mean (SD) | 1.20 (0.13) | 1.22 (0.13) | 1.16 (0.13) | < 0.001*** |
| CSVD burden, median (IQR) | 0 (1) | 0 (1) | 0 (1) | 0.558 |
| Lacunes, median (IQR) | 0 (0) | 0 (0) | 0 (0) | 0.664 |
| WMH volume, median (IQR) | 5.7 (7.8) | 5.4 (7.1) | 6.0 (7.9) | 0.249 |
| CMBs, median (IQR) | 0 (0) | 0 (0) | 0 (0) | 0.055 |
| EPVS, median (IQR) | 2 (0) | 2 (1) | 2 (0) | 0.45 |
| Cognitive performance | ||||
| MMSE, median (IQR) | 29 (2) | 29 (2) | 28 (2) | < 0.001*** |
| Memory, median (IQR) | 9.8 (3.4) | 9.9 (3.5) | 9.3 (3.9) | 0.07 |
| Visual–spatial, median (IQR) | 5.7 (1.7) | 6 (2) | 5.7 (1.9) | 0.015* |
| Executive function, median (IQR) | 94 (40.6) | 90.5 (37.2) | 99 (54.9) | 0.023* |
| Language, median (IQR) | 29.3 (2) | 29.3 (2) | 28.7 (2.3) | 0.062 |
Note: ***p < 0.001; **p < 0.01; * p < 0.05.
Abbreviations: APOE‐ε4, the ε4 allele of the apolipoprotein E gene; CMBs, cerebral microbleeds; CSVD, cerebral small vessel disease; DTI‐ALPS, diffusion tensor image analysis along the perivascular space; EPVS, enlarged perivascular spaces; IQR, interquartile range; MMSE, Mini‐Mental State Examination; SD, standard deviation; VRFs, vascular risk factors; WMH, white matter hyperintensities.
3.2. Correlations Between PSQI/CSVD Score and DTI‐ALPS at Baseline
As shown in Figure 2A, higher DTI‐ALPS was significantly associated with lower PSQI (r = −0.17, p = 0.005) and lower CSVD burden (r = −0.12, p = 0.049) after adjusting for age, gender, VRFs, APOE‐ε4, and PSMD (Model 1). Among CSVD MRI markers, higher DTI‐ALPS was significantly associated with lower WMH volume (r = −0.15, FDR‐p = 0.019, Figure 2B). After further adjustment for median MD (Model 2), the association between DTI‐ALPS and PSQI remained significant (r = −0.15, p = 0.02), while other correlations were not statistically significant (Table S1). Besides, no significant correlations were observed between DTI‐ALPS and lacunes/CMBs/EPVS.
FIGURE 2.
The relationship between sleep quality, CSVD, glymphatic function, and cognitive performance at baseline. Model adjusted for age, gender, VRFs, APOE‐ε4 carrier status, PSMD, education (for cognitive performance), and TIV (for WMH volume). The false discovery rate (FDR) correction method was used to correct for multiple comparisons. CMBs, cerebral microbleeds; CSVD, cerebral small vessel disease; DTI‐ALPS, diffusion tensor image analysis along the perivascular space; EF, executive function; EPVS, enlarged perivascular spaces; MMSE, Mini‐Mental State Examination; PSQI, Pittsburgh Sleep Quality Index; VS, visual spatial; WMH, white matter hyperintensities.
3.3. Correlations Between DTI‐ALPS and Cognitive Performance at Baseline
Partial correlation analyses revealed that higher DTI‐ALPS was significantly associated with better memory (r = 0.15, FDR‐p = 0.025), visual–spatial abilities (r = 0.16, FDR‐p = 0.025), and executive function (r = −0.25, FDR‐p < 0.001) in Model 1 (Figure 2C). While in Model 2, only the association with executive function remained significant (r = −0.22, FDR‐p = 0.005, Table S2). No significant correlations were found between DTI‐ALPS and MMSE/language.
3.4. DTI‐ALPS as a Significant Mediator Between Sleep/CSVD Burden and Cognitive Performance at Baseline
As shown in Figure 3, DTI‐ALPS significantly mediated the association between CSVD burden and executive function, with a mediation effect of 18.7%. DTI‐ALPS also significantly mediated the relationship between PSQI and executive function, with a mediation effect of 26.1%. While further adjusting for median MD, the statistical significance of mediation effects was attenuated (Tables S3 and S4). Additionally, DTI‐ALPS significantly mediated the relationship between global WMH volume and executive function, with a mediation effect of 23.1% (Figure S2 and Table S5). These findings indicated that DTI‐ALPS simultaneously mediated the associations between sleep quality/CSVD burden and cognitive performance.
FIGURE 3.
Path diagram of the mediation analyses at baseline. (A) The mediating effect of DTI‐ALPS in the relationship between CSVD burden and executive function; (B) The mediating effect of DTI‐ALPS in the relationship between PSQI and executive function. Model adjusted for age, gender, VRFs, education, APOE‐ε4, and PSMD. ***p < 0.001; **p < 0.01; *, p < 0.05. CSVD, cerebral small vessel disease; DTI‐ALPS, diffusion tensor image analysis along the perivascular space; PSQI, Pittsburgh Sleep Quality Index.
3.5. Interaction Analyses of Sleep and CSVD Burden on Cognitive Performance at Baseline
The linear regression model revealed a significant interaction effect between CSVD burden and sleep quality (β = −0.16, p = 0.002, Table S6). To further investigate the interaction effect, participants were categorized into healthy sleepers (PSQI score ≤ 5) and poor sleepers (PSQI score > 5). Additionally, participants were grouped by CSVD scores: the high‐risk group (CSVD score ≥ 1) and the low‐risk group (CSVD score < 1) [28]. The GLM revealed a significant interaction effect between CSVD risk level and sleep status (F = 5.69, p = 0.018), indicating that the impact of sleep quality on MMSE is modulated by CSVD risk level. Specifically, post hoc pairwise comparisons indicated that the impact of poor sleep quality on lower MMSE is more pronounced among participants at high risk of CSVD (p = 0.001). Additionally, although no significant interaction effect between sleep status and CSVD risk level was observed in other cognitive domains, pairwise comparisons still revealed that poor sleep quality was associated with worse performance in memory, executive function, and visual–spatial ability in participants at high risk of CSVD (Figure 4).
FIGURE 4.
The impact of sleep status on cognition in participants with different levels of CSVD risk. Post hoc pairwise comparisons showed that (A) among participants at high risk of CSVD, there was a significant difference in MMSE scores between healthy sleepers and poor sleepers (p = 0.001), whereas no statistically significant difference was observed in the low risk group (p = 0.072); (B) in the high risk group, there was a significant difference in memory between healthy sleepers and poor sleepers (p = 0.023), whereas no statistically significant difference was observed in the low risk group (p = 0.435); (C) there was a significant difference in executive function between healthy sleepers and poor sleepers in the high risk group (p = 0.008), whereas no statistically significant difference was observed in the low risk group (p = 0.182); (D) among participants at low risk of CSVD, there was a significant difference in visual spatial scores between healthy sleepers and poor sleepers (p = 0.01), whereas no statistically significant difference was observed in the high risk group (p = 0.195); (E) there was no difference in language between healthy sleepers and poor sleepers in both groups. CSVD, cerebral small vessel disease; EF, executive function; MMSE, Mini‐Mental State Examination; VS, visual spatial.
3.6. Correlations Between Baseline Sleep/MRI Markers and Longitudinal Cognitive Changes
The longitudinal analyses included 133 participants with an average follow‐up interval of 6.8 ± 0.6 years. Comparison of baseline characteristics between participants who completed follow‐up and those lost to follow‐up revealed significant differences: the latter group was older and demonstrated a lower DTI‐ALPS index, higher CSVD burden, and worse cognitive performance (all p < 0.05; Table S7). These differences suggested the presence of attrition bias, the potential impact of which is carefully considered in our discussion regarding result interpretation.
After adjusting for baseline age, gender, VRFs, education, APOE‐ε4, PSMD, and follow‐up interval, baseline DTI‐ALPS was significantly associated with longitudinal changes in MMSE scores (r = 0.24, p = 0.007; Figure 5A). Baseline CSVD burden also showed a significant correlation with MMSE changes (r = −0.17, p = 0.047; Figure 5B), while the association between PSQI and MMSE changes did not reach statistical significance (Figure 5C). Among MRI markers, baseline WMH volumes were significantly associated with longitudinal MMSE changes (r = −0.19, p = 0.039; Figure 5D). Detailed relationships between baseline PSQI/MRI markers and longitudinal cognitive changes were presented in Figure S3. No significant mediating effect of baseline DTI‐ALPS was observed in longitudinal analyses.
FIGURE 5.
The relationship between baseline sleep quality, CSVD, glymphatic function, and longitudinal changes in cognitive performance. Model adjusted for baseline age, gender, VRFs, education, APOE‐ε4, PSMD (for DTI‐ALPS), TIV (for WMH volume), and follow‐up interval. CSVD, cerebral small vessel disease; DTI‐ALPS, diffusion tensor image analysis along the perivascular space; MMSE, Mini‐Mental State Examination; PSQI, Pittsburgh Sleep Quality Index; WMH, white matter hyperintensities. △MMSE = (follow‐up MMSE)—(baseline MMSE).
4. Discussion
Our study reached three main conclusions: (1) Both CSVD burden and sleep quality were independently associated with DTI‐ALPS, which was further related to multiple cognitive domains and longitudinal decline in MMSE; (2) DTI‐ALPS simultaneously mediated the associations between CSVD burden and sleep quality and cognitive performance; (3) CSVD burden and sleep quality interacted to exacerbate cognitive impairment.
Accumulating evidence has demonstrated the impairment of glymphatic function in CSVD patients, establishing its role as a key mechanism in cognitive decline. Cross‐sectional findings consistently showed that glymphatic dysfunction, as assessed by DTI‐ALPS, was significantly correlated with CSVD imaging markers and cognitive impairment [25, 29, 30, 31]. However, longitudinal results remained inconsistent. Hong et al. did not observe a significant relationship between DTI‐ALPS and the progression of CSVD imaging markers during a 3‐year follow‐up period [29]. Nevertheless, baseline DTI‐ALPS effectively predicted long‐term cognitive decline and dementia risk, aligning with our results. Our study extended these findings by demonstrating that glymphatic function mediated the relationship between CSVD burden and cognition in a more general population. Notably, the causal relationships between glymphatic dysfunction and CSVD markers remain incompletely understood, as they may interact complexly with multiple pathological mechanisms including blood–brain barrier (BBB) disruption, chronic hypoperfusion, and neuroinflammation [32], highlighting the necessity for longitudinal studies. Besides overall CSVD burden, we particularly emphasized WMH due to its high prevalence in older adults and strong association with dementia risk [33]. Our findings demonstrated the significant impact of WMH on cognitive impairment in both cross‐sectional and longitudinal analyses.
It is widely accepted that sleep disturbances contribute to cognitive decline by impairing glymphatic clearance efficiency, leading to the accumulation of metabolic waste in the brain [12, 34, 35]. This mechanism has been implicated in various neurodegenerative diseases [34], with similar effects observed among cognitively healthy adults [6, 36, 37]. However, most evidence derives from rodent studies [12, 38]. Human research linking poor sleep to increased deposition of brain waste has largely relied on positron emission tomography (PET) imaging [6, 36, 37]. Recently, non‐invasive methods such as DTI‐ALPS have enabled the assessment of glymphatic function in large‐scale population investigations. A cross‐sectional study among community older adults revealed a close interaction among DTI‐ALPS, sleep, and cognition [39]. A recent study demonstrated that DTI‐ALPS mediated the association between sleep quality and multimodal brain network connectivity, which was further associated with memory in older adults [40]. Our study corroborated and extended these findings in a larger community‐based cohort, identifying glymphatic dysfunction, assessed by DTI‐ALPS, as a mediator between sleep quality and cognitive impairment.
Poor sleep and CSVD are common comorbidities among older adults. Growing evidence highlights the critical role of sleep disruption and impaired brain waste clearance in CSVD pathogenesis [34]. Previous studies have observed associations between sleep indicators and CSVD burden in community‐dwelling adults and obstructive sleep apnea (OSA) patients [41, 42]. However, their interaction impact on cognition has been rarely explored [39, 41, 43]. Our findings demonstrated that the combined effect of CSVD and poor sleep exacerbated cognitive impairment more than either factor alone. This underscores the importance of integrated management strategies and highlights sleep improvement as a promising therapeutic target, particularly for older adults at elevated risk of CSVD [34, 44]. Furthermore, we identified glymphatic dysfunction as a shared pathway linking sleep quality and CSVD to cognition. This is consistent with recent literature suggesting that glymphatic failure may represent a final common pathway to dementia in both vascular and non‐vascular aging populations [45]. The development of therapeutic strategies aimed at enhancing glymphatic function, including physical and pharmacological treatments, may offer clinical benefits in mitigating cognitive decline associated with CSVD.
Notably, our longitudinal analyses failed to replicate the mediation effects of glymphatic dysfunction observed in cross‐sectional analyses. Additionally, while prior evidence demonstrated the long‐term impact of sleep duration on cognitive decline [46], we found no significant association between baseline PSQI and longitudinal cognitive changes. This may be attributed to several factors. First, the relatively small sample size and high loss to follow‐up rate could reduce the statistical power. Second, attrition bias potentially leads to an underestimation of the true associations, as individuals with more severe clinical profiles were disproportionately lost to follow‐up. Third, selection bias at baseline‐involving relatively healthy community‐dwelling older adults and excluding those with dementia or severe cognitive impairment‐resulted in minimal cognitive progression during follow‐up, which reduced the ability to detect longitudinal differences. Finally, PSQI may inadequately reflect long‐term sleep patterns influencing cognition. Nonetheless, our findings showed that baseline DTI‐ALPS and CSVD burden remained significantly associated with longitudinal cognitive changes, underscoring their critical role in cognitive progression.
Recent studies have raised concerns about whether DTI‐ALPS provides independent information beyond conventional DTI metrics such as MD and PSMD, which have been shown to be closely associated with cognition [26, 29]. In our study, the observed associations and mediation effects involving DTI‐ALPS remained significant after adjusting for PSMD but weakened after additional adjustment for median MD, suggesting that DTI‐ALPS may partly represent white matter microstructural changes captured by MD. This highlights the necessity of carefully interpreting DTI‐ALPS findings in the context of broader diffusivity changes. Notably, the relationship between sleep quality, DTI‐ALPS, and executive function remained significant after adjustment for PSMD and MD, suggesting that DTI‐ALPS may retain sensitivity to sleep‐related glymphatic function and its impact on cognitive impairment.
Our study had several limitations. First, although ALPS‐index significantly correlates with intrathecal contrast‐based glymphatic evaluation, it may not fully characterize the dynamic complexity of the glymphatic system and requires further pathophysiological validation [25, 47]. PVS volume fluctuates across different sleep stages [48]; our static DTI‐ALPS measurement during wakefulness cannot fully reflect the change of glymphatic activity during sleep. Additionally, the specificity of DTI‐ALPS may be confounded by underlying white matter microstructural changes. Nevertheless, we employed DTI‐ALPS as a practical preliminary imaging marker to investigate glymphatic dysfunction in community‐dwelling older adults, given its accessibility and widespread adoption in existing research, which facilitated cross‐study comparisons. Furthermore, we did not investigate AQP4 gene variants, which may influence glymphatic function and cognitive decline [49]. Future studies should incorporate additional glymphatic markers, such as perivascular space volume, choroid plexus volume, or CSF‐BOLD coupling, and comprehensively account for factors influencing glymphatic function (e.g., AQP4 gene variants and sleep state) to facilitate a more robust evaluation. Second, PSQI‐based sleep quality measures may be biased by subjective reporting. Objective sleep measures, such as polysomnography, would strengthen future studies. Third, the relatively small sample size, combined with potential selection bias at baseline and attrition bias during follow‐up, may limit the generalizability of our findings and warrant caution when extrapolating these findings to broader populations.
In conclusion, our study identified glymphatic function as a shared pathway linking CSVD burden and poor sleep quality to cognitive impairment, highlighting its potential role as a target for early detection and intervention. The CSVD‐sleep interaction further emphasized the importance of sleep improvement in high‐risk populations. Large‐scale longitudinal studies integrating neurology, sleep medicine, and advanced neuroimaging techniques are warranted to elucidate underlying mechanisms.
Author Contributions
Rong Zhou: writing – original draft preparation, data curation, formal analysis, methodology, visualization, investigation. Weiyi Zhong: conceptualization, data curation, investigation, writing – review and editing. Yiwei Xia: data curation, investigation. Yunqing Ying: data curation, investigation. Yi Wang: data curation, investigation. Lumeng Yang: data curation, investigation. Jiajie Xu: data curation, software. Jinyi Cao: data curation, investigation. Zonghui Liang: investigation, data curation. Xiaoxiao Wang: data curation, software. Qiang Dong: project administration, supervision. Ding Ding: methodology, resources, supervision, writing – review and editing. Xin Cheng: conceptualization, methodology, resources, supervision, writing – review and editing, funding acquisition.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Data S1: ene70384‐sup‐0001‐Supinfo.docx.
Acknowledgements
We express our sincere gratitude to all the staff, enrolled participants, and their families.
Zhou R., Zhong W., Xia Y., et al., “The Relationship Between Cerebral Small Vessel Disease, Sleep Quality, and Cognitive Impairment Among Community‐Dwelling Older Adults: Exploring the Role of Glymphatic Function,” European Journal of Neurology 32, no. 10 (2025): e70384, 10.1111/ene.70384.
Funding: This work was supported by Noncommunicable Chronic Diseases—National Science and Technology Major Project (2023ZD0504903) and Shanghai Municipal Health Commission (2022XD022).
Rong Zhou and Weiyi Zhong contributed equally as first authors.
Xin Cheng and Ding Ding are the co‐corresponding authors.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Ohlmeier L., Nannoni S., Pallucca C., et al., “Prevalence of, and Risk Factors for, Cognitive Impairment in Lacunar Stroke,” International Journal of Stroke 18, no. 1 (2023): 62–69, 10.1177/17474930211064965. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Yaffe K., Falvey C. M., and Hoang T., “Connections Between Sleep and Cognition in Older Adults,” Lancet Neurology 13, no. 10 (2014): 1017–1028, 10.1016/S1474-4422(14)70172-3. [DOI] [PubMed] [Google Scholar]
- 3. Markus H. S. and de Leeuw F. E., “Cerebral Small Vessel Disease: Recent Advances and Future Directions,” International Journal of Stroke 18, no. 1 (2023): 4–14, 10.1177/17474930221144911. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Hilal S., Tan C. S., Adams H. H. H., et al., “Enlarged Perivascular Spaces and Cognition: A Meta‐Analysis of 5 Population‐Based Studies,” Neurology 91, no. 9 (2018): e832–e842, 10.1212/WNL.0000000000006079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Del Brutto O. H., Mera R. M., Del Brutto V. J., and Castillo P. R., “Enlarged Basal Ganglia Perivascular Spaces and Sleep Parameters. A Population‐Based Study,” Clinical Neurology and Neurosurgery 182 (2019): 53–57, 10.1016/j.clineuro.2019.05.002. [DOI] [PubMed] [Google Scholar]
- 6. Spira A. P., Gamaldo A. A., An Y., et al., “Self‐Reported Sleep and β‐Amyloid Deposition in Community‐Dwelling Older Adults,” JAMA Neurology 70, no. 12 (2013): 1537–1543, 10.1001/jamaneurol.2013.4258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Iliff J. J., Wang M., Liao Y., et al., “A Paravascular Pathway Facilitates CSF Flow Through the Brain Parenchyma and the Clearance of Interstitial Solutes, Including Amyloid β,” Science Translational Medicine 4, no. 147 (2012): 147ra111, 10.1126/scitranslmed.3003748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Wardlaw J. M., Benveniste H., Nedergaard M., et al., “Perivascular Spaces in the Brain: Anatomy, Physiology and Pathology,” Nature Reviews Neurology 16, no. 3 (2020): 137–153, 10.1038/s41582-020-0312-z. [DOI] [PubMed] [Google Scholar]
- 9. Mestre H., Tithof J., Du T., et al., “Flow of Cerebrospinal Fluid Is Driven by Arterial Pulsations and Is Reduced in Hypertension,” Nature Communications 9, no. 1 (2018): 4878, 10.1038/s41467-018-07318-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Katz P. S., Trask A. J., Souza‐Smith F. M., et al., “Coronary Arterioles in Type 2 Diabetic (Db/Db) Mice Undergo a Distinct Pattern of Remodeling Associated With Decreased Vessel Stiffness,” Basic Research in Cardiology 106, no. 6 (2011): 1123–1134, 10.1007/s00395-011-0201-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Iliff J. J., Wang M., Zeppenfeld D. M., et al., “Cerebral Arterial Pulsation Drives Paravascular CSF‐Interstitial Fluid Exchange in the Murine Brain,” Journal of Neuroscience 33, no. 46 (2013): 18190–18199, 10.1523/JNEUROSCI.1592-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Xie L., Kang H., Xu Q., et al., “Sleep Drives Metabolite Clearance From the Adult Brain,” Science 342, no. 6165 (2013): 373–377, 10.1126/science.1241224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Fultz N. E., Bonmassar G., Setsompop K., et al., “Coupled Electrophysiological, Hemodynamic, and Cerebrospinal Fluid Oscillations in Human Sleep,” Science 366, no. 6465 (2019): 628–631, 10.1126/science.aax5440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Taoka T., Masutani Y., Kawai H., et al., “Evaluation of Glymphatic System Activity With the Diffusion MR Technique: Diffusion Tensor Image Analysis Along the Perivascular Space (DTI‐ALPS) in Alzheimer's Disease Cases,” Japanese Journal of Radiology 35, no. 4 (2017): 172–178, 10.1007/s11604-017-0617-z. [DOI] [PubMed] [Google Scholar]
- 15. Ding D., Zhao Q., Guo Q., et al., “The Shanghai Aging Study: Study Design, Baseline Characteristics, and Prevalence of Dementia,” Neuroepidemiology 43, no. 2 (2014): 114–122, 10.1159/000366163. [DOI] [PubMed] [Google Scholar]
- 16. Zhang R., Lin M., Cho J., et al., “Oxygen Extraction Fraction in Small Vessel Disease: Relationship to Disease Burden and Progression,” Brain 148, no. 6 (2025): 1950–1962, 10.1093/brain/awae383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Luo J., Zhu G., Zhao Q., et al., “Prevalence and Risk Factors of Poor Sleep Quality Among Chinese Elderly in an Urban Community: Results From the Shanghai Aging Study,” PLoS One 8, no. 11 (2013): e81261, 10.1371/journal.pone.0081261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Tsai P. S., Wang S. Y., Wang M. Y., et al., “Psychometric Evaluation of the Chinese Version of the Pittsburgh Sleep Quality Index (CPSQI) in Primary Insomnia and Control Subjects,” Quality of Life Research 14, no. 8 (2005): 1943–1952, 10.1007/s11136-005-4346-x. [DOI] [PubMed] [Google Scholar]
- 19. Saito Y., Hayakawa Y., Kamagata K., et al., “Glymphatic System Impairment in Sleep Disruption: Diffusion Tensor Image Analysis Along the Perivascular Space (DTI‐ALPS),” Japanese Journal of Radiology 41, no. 12 (2023): 1335–1343, 10.1007/s11604-023-01463-6. [DOI] [PubMed] [Google Scholar]
- 20. Mok V., Srikanth V., Xiong Y., et al., “Race‐Ethnicity and Cerebral Small Vessel Disease—Comparison Between Chinese and White Populations,” International Journal of Stroke 9 Suppl A100 (2014): 36–42, 10.1111/ijs.12270. [DOI] [PubMed] [Google Scholar]
- 21. Xia Y., Shen Y., Wang Y., et al., “White Matter Hyperintensities Associated With Progression of Cerebral Small Vessel Disease: A 7‐Year Chinese Urban Community Study,” Aging (Albany NY) 12, no. 9 (2020): 8506–8522, 10.18632/aging.103154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Duering M., Biessels G. J., Brodtmann A., et al., “Neuroimaging Standards for Research Into Small Vessel Disease‐Advances Since 2013,” Lancet Neurology 22, no. 7 (2023): 602–618, 10.1016/S1474-4422(23)00131-X. [DOI] [PubMed] [Google Scholar]
- 23. Potter G. M., Chappell F. M., Morris Z., and Wardlaw J. M., “Cerebral Perivascular Spaces Visible on Magnetic Resonance Imaging: Development of a Qualitative Rating Scale and Its Observer Reliability,” Cerebrovascular Diseases 39, no. 3–4 (2015): 224–231, 10.1159/000375153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Klarenbeek P., van Oostenbrugge R. J., Rouhl R. P., Knottnerus I. L., and Staals J., “Ambulatory Blood Pressure in Patients With Lacunar Stroke: Association With Total MRI Burden of Cerebral Small Vessel Disease,” Stroke 44, no. 11 (2013): 2995–2999, 10.1161/STROKEAHA.113.002545. [DOI] [PubMed] [Google Scholar]
- 25. Zhang W., Zhou Y., Wang J., et al., “Glymphatic Clearance Function in Patients With Cerebral Small Vessel Disease,” NeuroImage 238 (2021): 118257, 10.1016/j.neuroimage.2021.118257. [DOI] [PubMed] [Google Scholar]
- 26. Baykara E., Gesierich B., Adam R., et al., “A Novel Imaging Marker for Small Vessel Disease Based on Skeletonization of White Matter Tracts and Diffusion Histograms,” Annals of Neurology 80, no. 4 (2016): 581–592, 10.1002/ana.24758. [DOI] [PubMed] [Google Scholar]
- 27. Ding D., Zhao Q., Guo Q., et al., “Prevalence of Mild Cognitive Impairment in an Urban Community in China: A Cross‐Sectional Analysis of the Shanghai Aging Study,” Alzheimer's & Dementia 11, no. 3 (2015): 300–309.e302, 10.1016/j.jalz.2013.11.002. [DOI] [PubMed] [Google Scholar]
- 28. Kitagawa K., Toi S., Hosoya M., et al., “Small Vessel Disease Burden Predicts Incident Stroke and All‐Cause Death, but Not Acute Coronary Event,” Hypertension Research 47, no. 11 (2024): 3001–3009, 10.1038/s41440-024-01797-2. [DOI] [PubMed] [Google Scholar]
- 29. Hong H., Tozer D. J., and Markus H. S., “Relationship of Perivascular Space Markers With Incident Dementia in Cerebral Small Vessel Disease,” Stroke 55, no. 4 (2024): 1032–1040, 10.1161/STROKEAHA.123.045857. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Yang J., Liu Y., Ma Y., et al., “Association of Glymphatic Clearance Function With Imaging Markers and Risk Factors of Cerebral Small Vessel Disease,” Journal of Stroke and Cerebrovascular Diseases 34, no. 2 (2025): 108187, 10.1016/j.jstrokecerebrovasdis.2024.108187. [DOI] [PubMed] [Google Scholar]
- 31. Tang J., Zhang M., Liu N., et al., “The Association Between Glymphatic System Dysfunction and Cognitive Impairment in Cerebral Small Vessel Disease,” Frontiers in Aging Neuroscience 14 (2022): 916633, 10.3389/fnagi.2022.916633. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Tian Y., Zhao M., Chen Y., Yang M., and Wang Y., “The Underlying Role of the Glymphatic System and Meningeal Lymphatic Vessels in Cerebral Small Vessel Disease,” Biomolecules 12, no. 6 (2022): 748, 10.3390/biom12060748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Hong H., Hong L., Luo X., et al., “The Relationship Between Amyloid Pathology, Cerebral Small Vessel Disease, Glymphatic Dysfunction, and Cognition: A Study Based on Alzheimer's Disease Continuum Participants,” Alzheimer's Research & Therapy 16, no. 1 (2024): 43, 10.1186/s13195-024-01407-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Hein Z. M., Che Mohd Nassir C. M. N., Che Ramli M. D., et al., “Cerebral Small Vessel Disease: The Impact of Glymphopathy and Sleep Disorders,” Journal of Cerebral Blood Flow and Metabolism 45 (2025): 3933, 10.1177/0271678X251333933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Qin S., Leong R. L. F., Ong J. L., and Chee M. W. L., “Associations Between Objectively Measured Sleep Parameters and Cognition in Healthy Older Adults: A Meta‐Analysis,” Sleep Medicine Reviews 67 (2023): 101734, 10.1016/j.smrv.2022.101734. [DOI] [PubMed] [Google Scholar]
- 36. Shokri‐Kojori E., Wang G.‐J., Wiers C. E., et al., “β‐Amyloid Accumulation in the Human Brain After One Night of Sleep Deprivation,” Proceedings of the National Academy of Sciences of the United States of America 115, no. 17 (2018): 4483–4488, 10.1073/pnas.1721694115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Sharma R. A., Varga A. W., Bubu O. M., et al., “Obstructive Sleep Apnea Severity Affects Amyloid Burden in Cognitively Normal Elderly. A Longitudinal Study,” American Journal of Respiratory and Critical Care Medicine 197, no. 7 (2018): 933–943, 10.1164/rccm.201704-0704OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Deng S., Hu Y., Chen S., et al., “Chronic Sleep Fragmentation Impairs Brain Interstitial Clearance in Young Wildtype Mice,” Journal of Cerebral Blood Flow and Metabolism 44, no. 9 (2024): 1515–1531, 10.1177/0271678X241230188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Siow T. Y., Toh C. H., Hsu J. L., et al., “Association of Sleep, Neuropsychological Performance, and Gray Matter Volume With Glymphatic Function in Community‐Dwelling Older Adults,” Neurology 98, no. 8 (2022): e829–e838, 10.1212/WNL.0000000000013215. [DOI] [PubMed] [Google Scholar]
- 40. Ma J., Chen M., Liu G.‐H., et al., “Effects of Sleep on the Glymphatic Functioning and Multimodal Human Brain Network Affecting Memory in Older Adults,” Molecular Psychiatry 30, no. 5 (2025): 1717–1729, 10.1038/s41380-024-02778-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Baril A.‐A., Pinheiro A. A., Himali J. J., et al., “Lighter Sleep Is Associated With Higher Enlarged Perivascular Spaces Burden in Middle‐Aged and Elderly Individuals,” Sleep Medicine 100 (2022): 558–564, 10.1016/j.sleep.2022.10.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Song T.‐J., Park J.‐H., Choi K. H., et al., “Moderate‐To‐Severe Obstructive Sleep Apnea Is Associated With Cerebral Small Vessel Disease,” Sleep Medicine 30 (2017): 36–42, 10.1016/j.sleep.2016.03.006. [DOI] [PubMed] [Google Scholar]
- 43. Wang J., Chen X., Liao J., et al., “The Influence of Non‐Breathing‐Related Sleep Fragmentation on Cognitive Function in Patients With Cerebral Small Vessel Disease,” Neuropsychiatric Disease and Treatment 15 (2019): 1009–1014, 10.2147/NDT.S193869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Yang Q., Wei X., Deng B., et al., “Cerebral Small Vessel Disease Alters Neurovascular Unit Regulation of Microcirculation Integrity Involved in Vascular Cognitive Impairment,” Neurobiology of Disease 170 (2022): 105750, 10.1016/j.nbd.2022.105750. [DOI] [PubMed] [Google Scholar]
- 45. Nedergaard M. and Goldman S. A., “Glymphatic Failure as a Final Common Pathway to Dementia,” Science 370, no. 6512 (2020): 50–56, 10.1126/science.abb8739. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Ma Y., Liang L., Zheng F., Shi L., Zhong B., and Xie W., “Association Between Sleep Duration and Cognitive Decline,” JAMA Network Open 3, no. 9 (2020): e2013573, 10.1001/jamanetworkopen.2020.13573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Taoka T., Ito R., Nakamichi R., Nakane T., Kawai H., and Naganawa S., “Diffusion Tensor Image Analysis ALong the Perivascular Space (DTI‐ALPS): Revisiting the Meaning and Significance of the Method,” Magnetic Resonance in Medical Sciences 23, no. 3 (2024): 268–290, 10.2463/mrms.rev.2023-0175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Bojarskaite L., Vallet A., Bjørnstad D. M., et al., “Sleep Cycle‐Dependent Vascular Dynamics in Male Mice and the Predicted Effects on Perivascular Cerebrospinal Fluid Flow and Solute Transport,” Nature Communications 14, no. 1 (2023): 953, 10.1038/s41467-023-36643-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Liu L., Zeng Q., Luo X., et al., “Association Between Single Nucleotide Polymorphisms in the Aquaporin‐4 Gene and Longitudinal Changes in White Matter Free Water and Cognitive Function in Non‐Demented Older Adults,” Human Brain Mapping 46, no. 4 (2025): e70171, 10.1002/hbm.70171. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1: ene70384‐sup‐0001‐Supinfo.docx.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available due to privacy or ethical restrictions.