Left-lateralized glymphatic dysfunction is associated with MoCA-measured cognitive impairment in cerebral small vessel disease: A DTI-ALPS study

Abstract

Background

Cerebral small vessel disease (CSVD) is a major cause of vascular cognitive impairment. Glymphatic dysfunction has been implicated in cognitive decline, but its relationship with CSVD-related cognitive impairment, particularly with regard to potential hemispheric differences, remains unclear.

Methods

Forty-eight CSVD patients and 56 matched healthy controls underwent brain MRI and cognitive assessment using the Montreal Cognitive Assessment (MoCA). Bilateral ALPS indices (ALPS-L and ALPS-R) were calculated to assess glymphatic function. Pearson and partial correlation analyses were performed to examine associations between ALPS indices and MoCA scores. Logistic regression models and receiver operating characteristic (ROC) analyses were used to evaluate the associations between hemispheric ALPS indices and cognitive impairment as measured by the MoCA and assess their performance in distinguishing CSVD patients from healthy controls.

Results

Both ALPS-L and ALPS-R were significantly reduced in CSVD patients (p < 0.01). ALPS-L was positively correlated with MoCA scores (r = 0.307, p = 0.001), independent of demographic factors, whereas ALPS-R showed no significant association. ROC analyses demonstrated that incorporating ALPS indices improved the prediction of relatively lower cognitive performance measured by MoCA, with the model including ALPS-L showing higher predictive performance (AUC = 0.765) compared with the contrast model including ALPS-R (AUC = 0.737). The trend of these models in distinguishing CSVD from healthy controls is consistent.

Conclusion

CSVD patients exhibit bilateral lymphatic dysfunction. The left ALPS index showed a lateralized association with MoCA-measured cognitive impairment and may represent a potential imaging marker associated with cognitive performance in CSVD.

Highlights

• •Left ALPS, not right, predicts cognitive impairment in CSVD.
• •Glymphatic–cognition association shows hemispheric specificity.
• •ALPS-L improves prediction beyond conventional factors.

1. Introduction

Cerebral small vessel disease (CSVD) is a disorder of the small perforating arterioles, capillaries, and probably venules in brain (Wardlaw and Smith, 2019). The neuroimaging features of CSVD include white matter hyperintensities (WMHs), enlarged perivascular spaces (EPVSs), lacunes, and cerebral microbleeds (CMBs) (Shu and Zhai, 2021). CSVD is the most common pathology underlying vascular cognitive impairment and vascular dementia, accounting for approximately 20% of all dementia cases and typically associated with a poorer quality of life and increased disability risk (Sigdel and Sawant, 2026). A characteristic pattern of cognitive impairment is found with significant involvement of executive functions and slowing of processing speed (Vazquez-Marrufo and Galvao-Carmona, 2019). Traditional imaging markers such as WMHs and lacunes have been shown to correlate with cognitive impairment and possess some predictive value (Yu and Yin, 2021). However, in clinical practice, significant cognitive deficits are sometimes observed in patients with the relatively mild imaging burden, indicating that the mechanisms and early predictive biomarkers of CSVD-related cognitive impairment are not well understood (Tang and Zhang, 2022).

Recent studies indicate that the glymphatic system (GS), a brain-wide perivascular fluid transport network, may contribute to the pathology of CSVD-related cognitive impairment, as it plays a critical role in metabolic waste clearance and brain homeostasis (Nedergaard and Goldman, 2020). GS mediates cerebrospinal fluid (CSF) influx along periarterial spaces, facilitates the exchange with interstitial fluid (ISF), and enables metabolic waste clearance via perivenous and meningeal lymphatic pathways. This convective transport depends critically on arterial pulsatility, aquaporin−4-mediated astroglial water permeability, and the structural integrity of perivascular compartments (Plog and Nedergaard, 2018). Notably, accumulating evidence has demonstrated glymphatic dysfunction in CSVD (Hong and Tozer, 2025, Zhou and Zhong, 2025, Hong and Hong, 2024). The perivascular space (PVS), frequently enlarged in CSVD, serves as a principal anatomical conduit for GS-mediated ISF circulation (Mestre and Mori, 2020). Moreover, both arteriosclerosis and impaired vascular pulsatility, the characterization of CSVD, represent key driving forces of glymphatic flow. Taken together, these pathological features of CSVD disrupt both the driving mechanisms and anatomical pathways of glymphatic transport, thereby impairing perivascular fluid dynamics and solute clearance efficiency (Lee and Lee, 2024).

Typically, cognitive impairment in CSVD occurs due to chronic cerebral hypoperfusion and blood–brain barrier disruption (Tang and Zhang, 2022). Glymphatic dysfunction may result in impaired clearance of amyloid-β (Aβ) and other metabolites in CSVD. Consistently, multiple studies have found that Aβ and other neurotoxic metabolites accumulate in patients with CSVD, potentially leading to neuronal dysfunction and loss, thus contributing to cognitive decline (Ye and Seo, 2015). Notably, studies in Alzheimer’s disease (AD) and relevant animal models have demonstrated that glymphatic dysfunction promotes Aβ accumulation and cognitive impairment (Koronyo and Rentsendorj, 2023, Reeves and Karimy, 2020). Thereby, impaired glymphatic clearance has also been hypothesized to represent a possible mechanism underlying CSVD-related cognitive impairment.

Imaging the glymphatic system is essential for understanding its effects on cerebral physiology and pathology. Compared with other traditional ex vivo or contrast-enhanced methods, magnetic resonance imaging (MRI) overcomes the invasiveness and static limitations, and assesses glymphatic flow function (Botta and Hutuca, 2025). For example, diffusion tensor imaging analysis along the perivascular space (ALPS index), quantifies water diffusivity along perivascular trajectories beside the lateral ventricular body and is increasingly used to investigate correlations between glymphatic function, clinical symptoms, and neuroimaging biomarkers (Jia and Chen, 2025). The ALPS index has been found to be associated with cognitive function in neurodegenerative diseases, and can also be used to indicate the development of dementia such as AD and Parkinson's disease (Bao and Luo, 2025). Moreover, recent studies suggest that it is also associated with cognitive decline in CSVD patients (Liu and Maillard, 2025). However, the relative impacts of bilateral ALPS indices on CSVD progression and associated cognitive impairment are still unknown.

Therefore, this study aimed to (1) assess changes in bilateral lymphatic function in CSVD patients relative to healthy individuals using the DTI-ALPS method and to investigate its association with cognitive impairment; (2) compare the predictive performance of different models incorporating demographic factors and bilateral ALPS indices to determine whether the ALPS index could serve as an imaging biomarker for identifying CSVD.

2. Methods

2.1. Study population

A total of 48 patients with CSVD and 56 healthy controls (HCs) were recruited prospectively between January 2023 and December 2025 at our hospital. CSVD patients were identified according to the Standards for Reporting Vascular Changes on Neuroimaging (STRIVE, version 2). The inclusion criteria of CSVD patients were: (1) age between 45 and 75 years old; (2) right-handedness; (3) at least 3 years of education; (4) a Fazekas grade of ≥ 3 for WMHs. The exclusion criteria were: (1) any history of other neurological or psychiatric disorders (e.g., Parkinson's disease, Alzheimer's disease); (2) MRI showed any other preexisting structural brain lesions (e.g., hemorrhage, cerebral trauma, cerebrovascular malformation, brain tumors); (3) any MRI contraindications; (4) severe MRI artifacts or incomplete completion of neuropsychological scales. HCs were recruited from the community during the same study period, and matched to the CSVD group in terms of age, sex, and years of education.

All demographic information, vascular risk information, and history of diseases or medicine were recorded. All participants provided informed consent documents.

2.2. Clinical data

The demographic information was collected, including age, sex, and years of education. All participants underwent a standardized cognitive assessment. The Montreal Cognitive Assessment (MoCA) is a brief cognitive screening test used widely by professionals in clinical settings (Nasreddine and Phillips, 2005). MoCA was utilized to assess global cognitive function, as a compound measure of cognition in CSVD.

2.3. MRI acquisition

All participants were scanned on a 3.0-T MRI scanner (Magnetom Vida, Siemens AG), with a 32-channel head coil. Patients wore noise-canceling headphones to reduce scanner noise and limit head movement with flexible foam. The imaging protocol included axial T1-weighted images, T2-weighted images, T2 fluid-attenuated inversion recovery (FLAIR), diffusion-weighted imaging (DWI), susceptibility-weight imaging (SWI) and diffusion tensor imaging (DTI), and excluded possible lesions as specified in the exclusion criteria. Diffusion tensor imaging (DTI) was performed using a single-shot echo-planar imaging sequence with the following parameters: repetition time (TR) = 5100 ms; echo time (TE) = 84 ms; b-values = 0 and 1000 s/mm²; diffusion gradient directions = 64; field of view (FOV) = 25.6 × 25.6 cm²; matrix size = 128 × 128; slice thickness = 2 mm; and number of slices = 72.

2.4. DTI-ALPS analysis

The ALPS index was calculated as described by Taoka et al. (Taoka and Masutani, 2017). The raw DTI data was preprocessed using the FMRIB software library (FSL) diffusion toolbox (FDT), including correction for eddy current and head motion. The whole brain diffusion index of each voxel was calculated, and diffusivity values of Dx, Dy, and Dz along the x-, y-, and z-axes were generated respectively.

Regions of interest (ROIs) were placed at the level of the lateral ventricular body, where the PVS was horizontal from left to right (x-axis) and perpendicular to the ventricular walls. In this position, the projection fibers of the cortex were in the craniocaudal direction (z-axis), and the association fibers were in the anteroposterior direction (y-axis). For each hemisphere, diffusivity along the x-axis was measured in both projection fiber regions (Dxproj) and the association fiber regions (Dxass). Diffusivity along the y-axis was measured in the projection fiber area (Dyproj), and diffusivity along the z-axis was measured in the association fiber area (Dzass).

The ALPS index was calculated using the following formula:

ALPS index = mean(Dx_proj, Dx_ass) / mean(Dy_proj, Dz_ass)

The bilateral ALPS indices (ALPS-L and ALPS-R) were calculated separately, and the mean ALPS index was obtained by averaging the left and right values.

2.5. Statistical analysis

All statistical analyses were performed using IBM SPSS version 26.0 and R version 4.2.0. Statistical tests were two-sided, with p < 0.05 considered statistically significant. For categorical variables, the differences between two groups were assessed by the chi-square test or Fisher's exact test. For continuous variables, Shapiro-Wilk test was applied to evaluate the normality of data distribution, and the intergroup differences were compared by the two-sample t-test and Mann-Whitney U-test as appropriate. The intragroup differences in bilateral ALPS indices were evaluated by paired-sample t-test.

Pearson correlation test was applied to assess the correlations between the bilateral ALPS indices and demographic variables and MoCA scores. Partial correlation was used to investigate the relationship between ALPS indices and cognitive function, adjusting for age, sex, and years of education. For additional ROC analyses of cognitive performance, participants were stratified into relatively lower- and higher-cognition groups according to the median MoCA score (24 points). Multivariate logistic regression models were used to evaluate the predictive value for two purposes: (1) to distinguish participants with relatively lower versus higher cognitive performance and (2) to distinguish CSVD patients from healthy controls. The area under the receiver operating characteristic curve (ROC) was used to evaluate the ability of these models, and pairwise comparisons of ROC curves were performed using the DeLong test.

3. Results

3.1. Baseline characteristics

Table 1 shows the demographic and MoCA scores for CSVD patients as well as the HC. This study included 48 patients with CSVD (20 men and 28 women; mean age, 62.58 ± 5.65 years) and 56 HCs (32 men and 24 women; mean age, 60.82 ± 8.30 years). There were no statistically significant differences in age, sex, and years of education between the two groups. The MoCA scores were significantly lower in the CSVD group compared with the HC group (P = 0.006), indicating poorer cognitive function.

Table 1.

Baseline demographic and clinical features of the study sample.

Variable	CSVD Group (n = 48)	HC Group (n = 56)	Pvalue
Age at imaging	62.58 ± 5.65	60.82 ± 8.30	0.204
Sex, n(%)
Male	20 (41.67%)	32 (57.14%)	0.168
Female	28 (58.33%)	24 (42.86%)
MoCA score	22.10 ± 2.85	23.89 ± 3.50	0.006
Education index	8.88 ± 2.75	9.39 ± 4.21	0.454

Note: Unless otherwise specified, data are presented as mean ± Standard Deviation (SD).

Abbreviations: CSVD = cerebral small vessel disease; HC = healthy controls; MoCA = Montreal Cognitive Assessment.

3.2. Comparison of diffusivities and ALPS indices

Table 2 shows the differences in diffusivity values and ALPS indices between the CSVD and HC groups. There was no significant difference in most diffusivity values between the two groups except Dyproj-L (p = 0.036). Both ALPS-L (p = 0.001) and ALPS-R (p = 0.005) were significantly lower in the CSVD group compared with the HC group. Interestingly, the skewness of the right ALPS index differed significantly (Fig. 1). However, the two-sample paired t-test revealed no significant differences in bilateral ALPS index in both HC and CSVD patients (Table 3).

Table 2.

Comparison of the diffusivities and ALPS indices among the study groups.

Variable	CSVD Group (n = 48)	HC Group (n = 56)	Pvalue
Dxproj-L	0.627 (0.565–0.650)	0.619 (0.581–0.644)	0.661
Dxass-L	0.640 (0.599–0.682)	0.689 (0.610–0.755)	0.057
Dyproj-L	0.454 (0.421–0.496)	0.431 (0.394–0.462)	0.036
Dzass-L	0.377 (0.340–0.430)	0.377 (0.339–0.418)	0.429
Dxproj-R	0.624 (0.558–0.673)	0.623 (0.579–0.655)	0.951
Dxass-R	0.663 (0.593–0.728)	0.690 (0.608–0.743)	0.123
Dyproj-R	0.450 (0.409–0.513)	0.443 (0.398–0.489)	0.120
Dzass-R	0.401 (0.348–0.463)	0.399 (0.335–0.450)	0.533
ALPS-L	1.519 (1.374–1.620)	1.595 (1.509–1.749)	0.001
ALPS-R	1.497 (1.340–1.620)	1.586 (1.435–1.707)	0.005
ALPS-mean	1.506 (1.415–1.606)	1.619 (1.496–1.712)	0.000

Note: The data are presented as medians with interquartile ranges in parentheses. Diffusivities are reported as apparent diffusion coefficients (×10⁻³ mm²/s). The statistical significance threshold was set to p < 0.05 (*p < 0.05,**p < 0.01,***p < 0.001).

Abbreviations: CSVD = cerebral small vessel disease; HC = healthy controls; ALPS = analysis along the perivascular space; ALPS-L = left ALPS; ALPS-R = right ALPS; Dxproj = diffusivity along x-axis in projection fiber area; Dxass = diffusivity along x-axis in association fiber area; Dyproj = diffusivity along y-axis in projection fiber area; Dzass = diffusivity along z-axis in association fiber area; L = left; R = right.

Fig. 1.

Differences in the left ALPS (A) and right ALPS (B) between CSVD and HC group. The statistical significance threshold was set to p < 0.05 (*p < 0.05,**p < 0.01,***p < 0.001). ALPS-L = left ALPS; ALPS-R = right ALPS; CSVD = cerebral small vessel disease; HC = healthy controls; SK = Skewness.

Table 3.

Comparison of the bilateral diffusivities and ALPS indices within each study group.

	CSVD Group (n = 48)		P	HC Group (n = 56)		P
	L	R		L	R
Dxproj (mean±SD)	0.615 ± 0.066	0.629 ± 0.125	0.427	0.622 ± 0.093	0.630 ± 0.091	0.301
Dxass (mean±SD)	0.648 ± 0.089	0.656 ± 0.091	0.616	0.689 ± 0.126	0.686 ± 0.105	0.815
Dyproj (mean±SD)	0.455 ± 0.062	0.472 ± 0.126	0.348	0.432 ± 0.050	0.441 ± 0.070	0.201
Dzass (mean±SD)	0.387 ± 0.094	0.408 ± 0.076	0.168	0.374 ± 0.071	0.398 ± 0.075	0.028
ALPS (mean±SD)	1.517 ± 0.174	1.483 ± 0.178	0.199	1.631 ± 0.166	1.585 ± 0.184	0.073

Note: Diffusivities are reported as apparent diffusion coefficients (×10⁻³ mm²/s). Abbreviations: CSVD = cerebral small vessel disease; HC = healthy controls; ALPS = analysis along the perivascular space; ALPS-L = left ALPS; ALPS-R = right ALPS; Dxproj = diffusivity along x-axis in projection fiber area; Dxass = diffusivity along x-axis in association fiber area; Dyproj = diffusivity along y-axis in projection fiber area; Dzass = diffusivity along z-axis in association fiber area; SD = Standard Deviation.

3.3. Correlation between ALPS Indices and MoCA Scores

The Pearson's correlation was used to further explore the relationship between ALPS indices and MoCA-measured cognitive performance in patients with CSVD. Our results showed that ALPS-L (r = 0.307, p = 0.001) was positively correlated with MoCA scores in CSVD patients. The correlation between the ALPS-mean and MoCA scores showed a positive trend, although it did not reach statistical significance (r = 0.186, p = 0.058). However, ALPS-R (r = 0.021, p = 0.829) showed no significant correlation with MoCA scores (Fig. 2, Fig. 3). After adjusting for age and education using partial correlation analysis, the correlations between different ALPS indices remained the same as before (ALPS-L, r = 0.279, p = 0.004; ALPS-mean, r = 0.149, p = 0.130; ALPS-R, r = −0.012, p = 0.905) (Fig. 3).

Fig. 2.

Correlation matrix between demographic variables, MoCA score, and ALPS index. Pearson correlation coefficients (r) are indicated within boxes, and color bars represent the magnitude and direction of the correlation coefficients. The statistical significance threshold was set to p < 0.05 (*** p < 0.001, ** p < 0.01, * p < 0.05). ALPS = analysis along the perivascular space; ALPS-L = left ALPS; ALPS-R = right ALPS; MoCA = Montreal Cognitive Assessment.

Fig. 3.

Correlation between ALPS indices and MoCA scores. Scatter plots of Pearson correlations between MoCA scores and (A) the mean ALPS index, (B) the left ALPS index, and (C) the right ALPS index. Scatter plots showing the correlations between MoCA residuals and the residuals of the (D) mean, (E) left, and (F) right ALPS indices after controlling for age, sex, and education level. Orange dots represent the CSVD group, and blue dots represent the HC group. The black solid line indicates the linear regression fit, with the corresponding Pearson correlation coefficient (r) and p value displayed in the upper left corner. ALPS = analysis along the perivascular space; ALPS-L = left ALPS; ALPS-R = right ALPS; MoCA = Montreal Cognitive Assessment.

3.4. ROC analyses for MoCA-measured cognitive performance

To further investigate the relationship between ALPS indices and cognitive performance, participants were stratified into relatively lower- and higher-cognition groups according to the median MoCA score (24 points). ROC analyses were performed using different logistic regression models (Table 4 and Fig. 4). The base model including age, sex, and years of education yielded an AUC of 0.739 (95% CI: 0.644–0.834). Incorporation of ALPS-L increased the AUC to 0.765 (95% CI: 0.674–0.856), whereas incorporation of ALPS-R did not improve model performance (AUC = 0.737, 95% CI: 0.642–0.833). The model including both ALPS-L and ALPS-R achieved the highest AUC of 0.779 (95% CI: 0.690–0.868). Pairwise comparisons using the DeLong test showed no statistically significant differences between ROC curves (Supplementary Figure S1 A).

Table 4.

The performance of different models distinguishing different MoCA-measured cognition.

Model	AUC	P	95%CI
Base (Age + Edu + Sex)	0.739	＜0.001	0.644–0.834
Base + ALPS-mean	0.748	＜0.001	0.654–0.841
Base + ALPS-L	0.765	＜0.001	0.674–0.856
Base + ALPS-R	0.737	＜0.001	0.642–0.833
Base + ALPS-L + ALPS-R	0.779	＜0.001	0.690–0.868

Abbreviations: Base = Age + Education index + Sex; ALPS = analysis along the perivascular space; ALPS-L = left ALPS; ALPS-R = right ALPS; AUC = area under the curve; CI = Confidence interval.

Fig. 4.

(A, B) Receiver operating characteristic (ROC) curves of different models. The ROC curves for different models distinguishing relatively lower- and higher-cognition groups based on MoCA score. Base = Age + Education index + Sex; ALPS = analysis along the perivascular space; ALPS-L = left ALPS; ALPS-R = right ALPS.

3.5. ROC analyses for discrimination of CSVD and healthy control

Furthermore, as shown in Table 5 and Fig. 5, ROC curve analyses were performed for different models and the AUCs were calculated (Fig. 5). The base model including age, sex, and years of education showed limited predictive performance (AUC = 0.608, 95% CI: 0.498–0.717). The model predictive performance improved after incorporating different ALPS indices. Compared with ALPS-R (AUC = 0.676, 95% CI: 0.573–0.778), the models including ALPS-L and ALPS-mean showed the higher AUC of 0.712 (95% CI: 0.613–0.811) and 0.714 (95% CI: 0.616–0.812). When ALPS-L and ALPS-R were included simultaneously, the model showed the best predictive performance, achieving the highest AUC of 0.722 (95% CI: 0.625–0.820). These results suggest that incorporating ALPS indices, particularly ALPS-L, enhances the ability to identify and improve discrimination between CSVD patients and healthy controls. However, DeLong test showed no statistically significant differences between ROC curves, except for the Base and Base + ALPS-L groups, as well as the Base and Base + ALPS-L + ALPS-R groups (Supplementary Figure S1 B).

Table 5.

The performance of different models distinguishing CSVD patients from healthy controls.

Model	AUC	P	95%CI
Base (Age + Edu + Sex)	0.608	0.0591	0.498–0.717
Base + ALPS-mean	0.714	0.0002	0.616–0.812
Base + ALPS-L	0.712	0.0002	0.613–0.811
Base + ALPS-R	0.676	0.0021	0.573–0.778
Base + ALPS-L + ALPS-R	0.722	0.0001	0.625–0.820

Abbreviations: Base = Age + Education index + Sex; ALPS = analysis along the perivascular space; ALPS-L = left ALPS; ALPS-R = right ALPS; AUC = area under the curve; CI = Confidence interval.

Fig. 5.

(A, B) Receiver operating characteristic (ROC) curves of different models. The ROC curves for different models distinguishing CSVD patients from healthy controls. Base = Age + Education index + Sex; ALPS = analysis along the perivascular space; ALPS-L = left ALPS; ALPS-R = right ALPS.

4. Discussion

In this study, we quantified bilateral and mean ALPS indices to assess glymphatic function and examined their associations with MoCA-measured cognitive impairment in patients with CSVD. We found that (1) patients with CSVD exhibited significantly reduced ALPS indices, indicating impaired glymphatic transport; (2) lower ALPS indices were associated with worse global cognition measured by MoCA, independent of demographic factors, and although bilateral ALPS indices are reduced in CSVD, only the left-sided ALPS index was independently associated with global cognition. (3) In the additional ROC analyses, models incorporating ALPS-L consistently demonstrated numerically higher AUC values than models only incorporating ALPS-R. Taken together, our results demonstrate that glymphatic dysfunction is present in CSVD and contributes to cognitive impairment, extending current understanding of glymphatic involvement in CSVD by suggesting that its impact on cognition is not spatially uniform but may reflect hemisphere-dependent vulnerability.

The ALPS index has been identified as an indicator of several conditions associated with the glymphatic system and is considered a potential reflection of its function (Buccellato and D'Anca, 2022). In our study, the reduced ALPS indices observed in CSVD patients were consistent with previous MRI studies and support the hypothesis of glymphatic system dysfunction. It is generally recognized that impaired brain fluid transport via the glymphatic system plays a critical role in the initiation and progression of CSVD (Taoka and Ito, 2024). This may be due to stagnation of interstitial fluid (ISF) transport, which disrupts cerebral fluid homeostasis and subsequently leads to transient white matter edema, perivascular space dilatation, and eventually demyelination (Benveniste and Nedergaard, 2022, Markus and de Leeuw, 2023).

We found a positive association between ALPS indices and MoCA, which remained significant after adjusting for demographic covariates. This suggests that the ALPS index, particularly ALPS-L, may be a sensitive and independent marker for early detection of CSVD-related cognitive deficits. This aligns with a previous study by Tang et al. (Tang and Zhang, 2022). Recent CSVD cohorts have shown that lower ALPS indices correlate with deficits in executive function, processing speed, and global cognition, and may mediate the relationship between PVS burden and cognitive impairment (Ai and Li, 2026). Similarly, mediation analyses in cerebral amyloid angiopathy indicate that the ALPS index partly explains the link between small vessel injury and cognitive dysfunction (Xu and Su, 2022, Ke and Mo, 2022). Mechanistically, impaired glymphatic clearance may promote the accumulation of neurotoxic metabolites, including Aβ and inflammatory by-products, which accelerate neuronal dysfunction and network disintegration, contributing to neuronal damage and cognitive decline (Chen and Li, 2025, Yang and Zhang, 2020).

On the one hand, our results suggest that bilateral ALPS indices are decreased without significant hemispheric asymmetry. On the other hand, in our study, the ALPS-L was correlated with MoCA-defined CSVD-related cognitive impairment and exhibits predictive value for cognitive impairment in ROC analysis, whereas ALPS-R was not. This apparent asymmetry likely reflects domain-specific vulnerability rather than true unilateral pathology.

The following principles may explain why this phenomenon occurs. First, brain regions are often activated asymmetrically, which is associated with cognitive function (Tsintzou and Poirier, 2025). Generally, the left hemisphere serves a crucial role in language, executive function, and logical reasoning. In contrast, the right hemisphere is relatively more involved in visuospatial processing and nonverbal reasoning (Hartwigsen and Bengio, 2021). Vascular cognitive impairment affects specific cognitive domains selectively rather than all domains (Tang and Zhang, 2022); a number of studies have demonstrated that executive function, attention, and memory are the main cognitive domains impaired in CSVD (Schroeter, 2022, Hamilton and Backhouse, 2021, Yang and Deng, 2021). Among these domains, both executive function and working memory function mainly depend on frontal–subcortical circuits of the left hemisphere except attention (Jahn, 2013, Hou and Hou, 2022). Second, several cognitive domains assessed by the MoCA, particularly language and certain executive functions, are more strongly associated with the dominant hemisphere (Chan and Altendorff, 2017). MoCA assesses several cognitive domains, including visuospatial abilities, language, attention, working memory, executive functions, and long-term verbal memory (Yuan and Liu, 2015), some of which exhibit varying degrees of hemispheric specialization. Notably, although attention networks are generally right-lateralized, the right hemisphere focuses more on spatial attention (Hartwigsen and Bengio, 2021, D'Esposito and Aguirre, 1998), which is less emphasized in MoCA. Third, the ALPS index is based only on the measurements beside the lateral ventricular body, which could only serve as an approximate measure of the regional area rather than global brain (Cao and Huang, 2024). The cognitive impairment may occur in other right brain regions, which is beyond the detection range of the ALPS-R. Collectively, ALPS-L is more sensitive to cognitive impairment measured by MoCA, and it does not mean that the right hemisphere is irrelevant to CSVD-related cognitive impairment.

Some limitations should be acknowledged. First, the glymphatic system occur throughout the brain, while the ALPS index was measured only beside the lateral ventricle body. Therefore, DTI-ALPS could not assess global glymphatic dynamics. Second, cognition was assessed only using the MoCA, which provides a global measure of cognitive performance and does not allow detailed characterization of specific cognitive domains. Therefore, although ALPS-L was associated with MoCA-measured cognition, the present study could not determine whether this association was driven by specific lateralized cognitive functions. Future studies incorporating domain-specific neuropsychological assessments may help clarify whether left- and right-sided ALPS indices show distinct associations with hemisphere-specific cognitive functions. Third, the present study is a single-center study with a relatively small sample size. DeLong testing did not demonstrate significant differences between ROC curves, the relatively modest sample size may have limited the statistical power to detect small differences in predictive performance. Given that the significance of the DeLong test is influenced by both the magnitude of AUC differences and their associated standard errors, larger studies are required to validate the observed numerical differences between models. Finally, those models in cross-sectional study could not infer temporal associations and causal associations between ALPS decrease and cognitive impairment in CSVD patients; comprehensive longitudinal studies are required.

5. Conclusion

In conclusion, our study has demonstrated that the only left ALPS index is independently associated with MoCA-measured cognitive performance, and models incorporating ALPS-L showed numerically higher performance than models incorporating ALPS-R.

Ethics Approval and Consent to Participate

All participants provided written informed consent prior to inclusion in the study. This study was conducted in accordance with the principles of the Declaration of Helsinki. The study protocol was reviewed and approved by the Ethics Committee of the Second Affiliated Hospital of Anhui Medical University (approval number: YX2026–120).

Consent for Publication

Not applicable.

Funding

This work was supported by the National Natural Science Foundation of China (32300869).

CRediT authorship contribution statement

Liangping Ni: Writing – review & editing, Writing – original draft, Software, Resources, Investigation, Data curation, Conceptualization. Yang Li: Writing – original draft, Supervision, Methodology, Formal analysis, Data curation. Mengyu Tian: Writing – review & editing, Writing – original draft, Visualization, Methodology, Formal analysis, Conceptualization. Dai Zhang: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization. Jun Zhang: Validation, Supervision, Data curation. Longsheng Wang: Writing – review & editing, Supervision, Project administration, Conceptualization. Kunpeng Cheng: Methodology, Formal analysis. Rong Wang: Methodology, Formal analysis. Bangyue Wang: Software.

Declaration of Generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT (OpenAI) for grammar and style editing under author supervision. No scientific content (e.g., study design, analyses, results, interpretations) was generated by AI. The authors reviewed and edited the output as needed and take full responsibility for the content of the published article.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

Supplementary material

Data availability

Anonymized data are available upon reasonable request from any qualified investigator.