Imaging study of white matter microstructural damage and glymphatic system dysfunction in patients with primary insomnia: a combined analysis based on PSMD, DTI-ALPS, and EPVS

Introduction

Insomnia has become a major public health issue worldwide because of its high prevalence and broad impact on physical and mental health. Epidemiological studies have shown that sleep disorders are highly prevalent across populations. A meta-analysis covering 49 countries reported that the global prevalence of sleep disorders from 2019 to 2021 was approximately 40.5% (1,2). Similarly, a 2022 survey in China showed that the prevalence of insomnia symptoms was about 46.8%, suggesting that sleep difficulties are also common in the Chinese population (3). According to the International Classification of Sleep Disorders, Third Edition (ICSD-3), insomnia is characterized by persistent difficulty initiating or maintaining sleep, accompanied by dissatisfaction with sleep and impaired daytime functioning (4). These daytime impairments may include fatigue, emotional instability, physical discomfort, and cognitive decline, indicating that insomnia is not merely a nighttime sleep complaint but a disorder involving multiple physiological and psychological systems. Long-term insomnia may further affect brain structure and function, interfere with attention, concentration, and memory, and increase the risk of anxiety and depression (5,6). Therefore, insomnia may contribute to a vicious cycle involving sleep disturbance, mood symptoms, and cognitive dysfunction.

Recent neuroscience research suggests that sleep, particularly the non-rapid eye movement (NREM) stage, is indispensable for maintaining the normal functioning of the brain’s glymphatic system (7). This system serves as a central pathway for clearing metabolic waste products, such as β-amyloid protein, from the central nervous system, and its function may depend on intact and regular sleep architecture (8). Compared with wakefulness, glymphatic waste clearance may be enhanced during deep sleep (9). Conversely, chronic insomnia or disrupted sleep architecture may be associated with impaired glymphatic function (10). Clinical observations indicate that insomnia is often accompanied by a state of chronic hyperarousal, which may interfere with the normal rhythm of norepinephrine, affect cerebrospinal fluid dynamics, and subsequently hinder glymphatic circulation (11). These alterations may promote the gradual accumulation of neurotoxic metabolites in brain tissue, ultimately accelerating brain aging and increasing the risk of neurodegenerative diseases.

The orexin/hypocretin system is a key regulator of wakefulness and sleep-wake stability (12). Through its projections to arousal-promoting regions such as the locus coeruleus, tuberomammillary nucleus, and raphe nuclei, dysregulated orexin signaling may contribute to chronic hyperarousal and disrupted sleep architecture in insomnia. Given that glymphatic clearance is closely related to consolidated sleep, particularly NREM sleep, orexin-related hyperarousal may also be relevant to glymphatic dysfunction. In neuroimaging studies, conventional magnetic resonance imaging (MRI) sequences such as T2-weighted fluid-attenuated inversion recovery (T2-FLAIR) can detect overt white matter hyperintensities but are less sensitive to subtle abnormalities within normal-appearing white matter (13). Visual scoring of enlarged perivascular spaces (EPVS) is a clinically accessible imaging marker that may reflect glymphatic dysfunction and cerebral small vessel disease burden (14). The clinical significance of EPVS may vary by brain region, as basal ganglia EPVS and centrum semiovale EPVS may be associated with different pathophysiological processes (15). The diffusion tensor imaging analysis along the perivascular space (DTI-ALPS) index has been proposed as an indirect imaging marker of glymphatic activity by measuring water diffusivity along perivascular spaces (16). Previous studies have linked reduced DTI-ALPS values to glymphatic dysfunction and cognitive decline in neurodegenerative diseases such as Alzheimer’s disease (17-19). More recently, Xiong et al. used DTI-ALPS to evaluate glymphatic system dysfunction in patients with insomnia and found that the left, right, and average DTI-ALPS indices were significantly lower in patients with insomnia than in healthy controls, suggesting impaired glymphatic circulation in insomnia (20). However, recent evidence suggests that DTI-ALPS alone may capture only part of glymphatic alterations (21,22). Therefore, a comprehensive assessment combining multiple imaging methods is needed to evaluate the overall status of this system.

Peak width of skeletonized mean diffusivity (PSMD) is a fully automated diffusion tensor imaging-derived metric that is increasingly used to quantify global white matter microstructural heterogeneity. It is calculated as the difference between the 95th and 5th percentiles of mean diffusivity (MD) values within the white matter skeleton and reflects diffuse alterations in white matter integrity rather than disease-specific pathology (23,24). Compared with conventional mean diffusion parameters, PSMD may provide greater sensitivity and stability in detecting subtle white matter tract abnormalities (25,26). Previous studies have shown that PSMD is associated with cognitive performance, particularly processing speed, in neurological conditions such as cerebral small vessel disease and Alzheimer’s disease (27,28).

At present, few studies systematically investigating alterations in brain white matter and the glymphatic system in insomnia patients by jointly applying metrics such as PSMD, DTI-ALPS, and EPVS. Based on the aforementioned background, this study aims to compare the differences in these imaging indicators between insomnia patients and healthy individuals, analyze the associations between these indicators and insomnia severity as well as cognitive decline, and further explore the application value of PSMD and DTI-ALPS in research on the neural mechanisms of insomnia. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0804/rc).

Methods

Study participants

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of the 900th Hospital of PLA Joint Logistic Support Force (approval No. 2026-007). The requirement for written informed consent was waived because of the retrospective nature of the study and the use of anonymized data.

This was a retrospective cross-sectional study. Data were collected from patients with primary insomnia who visited the Department of Neurology at the 900th Hospital of PLA Joint Logistic Support Force between February 2023 and February 2026. Screening was performed through clinical interviews by two senior neurologists. A total of 32 patients with primary insomnia were ultimately included, comprising 14 men and 18 women.

Participants in the primary insomnia group must meet the following inclusion criteria: (I) diagnosed with primary insomnia according to the International Classification of Sleep Disorders (Third Edition); (II) aged 18 years or older; (III) right-handed; (IV) insomnia duration exceeding 3 months; (V) Insomnia Severity Index (ISI) score ≥8; (VI) no use of psychoactive drugs for at least 2 weeks prior to enrollment; (VII) absence of major physical illnesses. Exclusion criteria: (I) cranial MRI indicating organic lesions of the central nervous system; (II) comorbid with other types of sleep disorders or mental illnesses; (III) history of drug abuse; (IV) contraindications to MRI or substandard image quality.

A total of 32 healthy participants matched for sex, age, and education level were recruited as the control group [healthy control (HC), 19 men and 13 women]. Exclusion criteria for the healthy control group were the same as for the insomnia group. Clinical data including age, sex, and years of education were collected for all participants.

Clinical and neuropsychological assessments

All participants underwent the following assessments: the ISI and Pittsburgh Sleep Quality Index (PSQI) for sleep status; the Hamilton Depression Rating Scale (HAMD) and Hamilton Anxiety Rating Scale (HAMA) for emotional state; and the Montreal Cognitive Assessment (MoCA) and Mini-Mental State Examination (MMSE) for global cognitive function.

MRI data acquisition

All scans were performed using a Siemens 3.0-T MRI scanner. Participants were placed in the supine position, the anterior-posterior commissure line was used as the scanning baseline.

The sequence parameters were as follows:

• T2-weighted imaging (T2WI): axial, repetition time (TR) =2,600 ms, echo time (TE) =98.0 ms, number of excitations (NEX) =2, matrix =384×307, field of view (FOV) =250 mm × 203 mm, flip angle (FA) =120°, slices =30, slice thickness =4.0 mm, gap =0.4 mm, scan time =1 min 18 sec.
• Three-dimensional (3D) T1-weighted imaging (T1WI): sagittal, TR =1,900 ms, TE =2.5 ms, NEX =1, matrix =256×256, FOV =256 mm × 256 mm, FA =9°, slices =176, slice thickness =1.0 mm, gap =0.5 mm, scan time =4 min 26 sec.
• Diffusion tensor imaging (DTI): axial, TR =7,200 ms, TE =104 ms, NEX =1, matrix =128×128, FOV =256 mm × 256 mm, slices =47, slice thickness =3.0 mm, gap =0.3 mm, voxel size 2.0 mm × 2.0 mm × 3.0 mm, scan time =8 min 11 sec.

Image data processing

PSMD calculation

FSL software was used to preprocess the raw DTI data, including motion and eddy current correction, brain extraction (removing scalp and skull), followed by tensor fitting to generate whole-brain MD and fractional anisotropy (FA) maps. The Tract-Based Spatial Statistics (TBSS) method was employed to extract regions with FA values ≥0.2 from a standard template to generate a white matter skeleton representing the core of major white matter tracts. Using the same spatial transformation parameters obtained from FA registration, individual MD maps were projected onto this common white matter skeleton, generating skeletonized MD maps. To reduce contamination from partial volume effects [e.g., from cerebrospinal fluid (CSF)], a higher-threshold template skeleton (e.g., FA ≥0.3) was used as a mask to further refine the skeletonized MD maps. MD values from all voxels within the masked skeletonized MD map were extracted, and their distribution histogram was plotted. The PSMD value was calculated as the difference between the 95th and 5th percentiles of the MD values within the skeleton (24,25).

DTI-ALPS index calculation

FSL software was used to preprocess the raw DTI images, including data format conversion, motion and eddy current correction, gradient direction correction, and skull stripping. On the color-coded FA map at the level of the lateral ventricle body, spherical regions of interest (ROIs) with a radius of 2.5 mm were placed bilaterally in projection and association fiber regions to extract diffusion coefficients along the x, y, and z axes (29). The DTI-ALPS index was calculated as the ratio of the average diffusivity along the x-axis in the projection area (Dxproj) to the average diffusivity along the x-axis in the association area (Dxassoc), divided by the ratio of the average diffusivity along the y-axis in the projection area (Dyproj) to the average diffusivity along the z-axis in the association area (Dzassoc) (Figure 1). The DTI-ALPS index was calculated using the following formula (29):

ALPS-index=mean(Dxproj,Dxassoc)mean(Dyproj,Dzassoc)[1]

Figure 1 Schematic diagram of the DTI-ALPS processing workflow. DTI-ALPS, diffusion tensor imaging analysis along the perivascular space.

EPVS visual scoring

According to the STRIVE2 guidelines (30), visual scoring (0–4 points) of EPVS in the basal ganglia and centrum semiovale is performed on T2-weighted images. Semi-quantitative grading was conducted using the Potter scale (31): grade 0 indicates no EPVS; grade 1, 1–10 EPVS; grade 2, 11–20 EPVS; grade 3, 21–40 EPVS; and grade 4, >40 EPVS. After all slices of the relevant bilateral anatomical regions had been evaluated, counts was performed only on the slice with the highest number of EPVS on one side; If asymmetry was present, the higher score was selected (32). EPVS ratings were independently performed by two physicians with 5 years of neuroimaging experience, who were blinded to participants’ clinical information and group allocation. Inter-rater reliability was assessed using weighted Cohen’s κ based on the initial independent ratings before consensus. Any discrepancies were adjudicated by a senior physician with 15 years of neuroimaging experience, and the adjudicated scores were used for the final statistical analyses (Figure 2).

Figure 2 Schematic diagram of visual scoring for BG-EPVS and CSO-EPVS. (A) No EPVS in the basal ganglia, scored as 0. (B) 1–10 EPVS present bilaterally in the basal ganglia, scored as 1. (C) 11–20 EPVS present in the right basal ganglia, and 1–10 EPVS in the left. The side with the higher count (right) is used for scoring, resulting in a score of 2. (D) 21–40 EPVS present bilaterally in the basal ganglia, scored as 3. (E) >40 EPVS present bilaterally in the basal ganglia, scored as 4. (F) No EPVS in the centrum semiovale, scored as 0. (G) 1–10 EPVS present bilaterally in the centrum semiovale, scored as 1. (H) 11–20 EPVS present in the centrum semiovale, scored as 2. (I) 21–40 EPVS present bilaterally in the centrum semiovale, scored as 3. (J) >40 EPVS present bilaterally in the centrum semiovale, scored as 4. This study did not include patients with a score of 4. BG-EPVS, basal ganglia enlarged perivascular spaces; CSO-EPVS, centrum semiovale enlarged perivascular spaces.

Statistical analysis

SPSS 27.0 software was used. Continuous data conforming to a normal distribution are presented as mean ± standard deviation, while non-normally distributed data are presented as median (interquartile range). Between-group comparisons were performed using independent sample t-tests or Mann-Whitney U tests. Partial correlation analyses adjusted for age, sex and years of education were used to explore the relationships between PSMD, DTI-ALPS, and clinical or cognitive measures. A two-sided P value <0.05 was considered statistically significant before correction for multiple comparisons. Effect sizes were reported for between-group comparisons, including rank-biserial correlation for non-normally distributed variables and Cohen’s d for normally distributed variables. Ninety-five percent confidence intervals were also calculated where appropriate. Because this was a retrospective cross-sectional study, no a priori sample size calculation was performed. Instead, a post hoc power analysis was conducted using G*Power for the PSMD between-group comparison. Based on the observed effect size, a two-tailed α level of 0.05, and the current sample size, the achieved statistical power was 0.485, indicating limited statistical power. Therefore, the between-group findings, especially the PSMD result, should be interpreted cautiously. Inter-rater reliability for basal ganglia enlarged perivascular spaces (BG-EPVS) and centrum semiovale enlarged perivascular spaces (CSO-EPVS) scores was evaluated using weighted Cohen’s κ.

Results

Baseline characteristics of participants

There were no significant differences between the insomnia and healthy control groups in age, sex, years of education, or vascular risk factors (diabetes, hypertension, smoking, alcohol consumption) (P>0.05). As shown in Table 1, the insomnia group had significantly higher scores on ISI, PSQI, HAMD, and HAMA than the healthy control group (P<0.001), while MoCA and MMSE scores were significantly lower in the insomnia group (P<0.010), indicating more severe insomnia symptoms, mood disturbances, and cognitive decline in insomnia patients.

Inter-rater reliability of EPVS ratings

Inter-rater reliability for EPVS visual scoring was high. The weighted Cohen’s κvalues was 0.869 for BG-EPVS and 0.884 for CSO-EPVS, indicating excellent agreement between the two raters. After reliability assessment, discrepant ratings were resolved by consensus with a senior neuroradiologist, and the adjudicated scores were used for subsequent statistical analyses.

Between-group comparison of imaging indicators

PSMD value was higher in the insomnia group than in the healthy control group, with a statistically significant difference (P=0.047, Table 2, Figure 3). The BG-EPVS visual score differed between the insomnia and healthy control groups in the unadjusted comparison, and this difference remained significant after false discovery rate (FDR) correction (raw P=0.040; FDR-adjusted P=0.046; Table 2, Figure 4). However, no significant between-group difference was observed in the DTI-ALPS index (P=0.267; Table 2, Figure 3) or CSO-EPVS score after FDR correction (FDR-adjusted P=0.211; Table 2, Figure 4).

Figure 3 Comparison of PSMD and DTI-ALPS indices between the PI group and HC group. (A) Box plot showing PSMD values; (B) box plot showing DTI-ALPS values. CI, confidence interval; DTI-ALPS, diffusion tensor imaging analysis along the perivascular space; HC, healthy control; PI, primary insomnia; PSMD, peak width of skeletonized mean diffusivity.

Figure 4 Comparison of BG-EPVS and CSO-EPVS scores between the PI group and HC group. The color gradient represents the proportion of subjects with scores ranging from 0 to 3. BG-EPVS, basal ganglia enlarged perivascular spaces; CSO-EPVS, centrum semiovale enlarged perivascular spaces; HC, healthy control; PI, primary insomnia.

After adjustment for age, years of education, sex, diabetes, hypertension, smoking, and alcohol consumption, the association between group and BG-EPVS score did not reach statistical significance in the ordinal logistic regression model [B=−1.113, odds ratio (OR) =0.329, 95% confidence interval (CI): 0.099−1.091; P=0.069]. Therefore, although the unadjusted BG-EPVS difference survived FDR correction, this finding should be interpreted cautiously because it was attenuated after adjustment for vascular risk factors.

Correlation analysis between PSMD index and glymphatic system-related imaging indicators and clinical/cognitive measures

After controlling for age, sex, and years of education, partial correlation analyses showed the following findings:

• PSMD and DTI-ALPS: a significant negative correlation was found in the full sample (r=−0.331, P=0.009). Within the insomnia group, this negative correlation was stronger and significant (r=−0.413, P=0.026), while it was not significant within the healthy control group.
• PSMD and BG-EPVS visual score: a significant positive correlation was found in the full sample (r=0.456, P<0.001).
• PSMD and cognitive function: PSMD was significantly negatively correlated with MMSE score in the full sample (r=−0.440, P<0.001) and also within the insomnia group (r=−0.476, P=0.009). PSMD was similarly significantly negatively correlated with MoCA score in the full sample (r=−0.329, P=0.01) and within the insomnia group (r=−0.476, P=0.009) (Figure 5).
• DTI-ALPS and clinical/cognitive measures: the DTI-ALPS index was not significantly correlated with ISI or PSQI scores in the full sample or within subgroups (all P>0.05).

Figure 5 Partial correlation analysis between PSMD and glymphatic system-related imaging metrics as well as clinical cognitive measures in the full sample after controlling for age, sex, and years of education. (A) Correlation between PSMD and DTI-ALPS index (r=−0.331, P=0.009); (B) correlation between PSMD and BG-EPVS score (r=0.456, P<0.001); (C) correlation between PSMD and MMSE score (r=−0.44, P<0.001); (D) correlation between PSMD and MoCA score (r=−0.329, P=0.01). BG-EPVS, basal ganglia enlarged perivascular spaces; DTI-ALPS, diffusion tensor imaging analysis along the perivascular space; MoCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination; PSMD, peak width of skeletonized mean diffusivity.

To further examine whether mood symptoms confounded the imaging findings, additional adjusted analyses were performed by including HAMD and HAMA scores as covariates.

PSMD remained significantly negatively correlated with the DTI-ALPS index (r=−0.314, P=0.015), MMSE score (r=−0.411, P=0.001), and MoCA score (r=−0.298, P=0.022). PSMD also remained significantly positively correlated with BG-EPVS score (r=0.433, P<0.015). These findings suggest that the associations between PSMD and glymphatic-related imaging markers as well as cognitive performance were not fully explained by depressive or anxiety symptoms. In the full sample, after FDR correction for multiple comparisons, PSMD was significantly positively correlated with BG-EPVS score (r=0.456, FDR-adjusted P=0.0011) and significantly negatively correlated with the DTI-ALPS index (r=−0.331, FDR-adjusted P=0.0194). PSMD was also significantly negatively correlated with MoCA score (r=−0.329, FDR-adjusted P=0.0194) and MMSE score (r=−0.440, FDR-adjusted P=0.0013). The correlations of BG-EPVS with MoCA, MMSE, or DTI-ALPS did not remain statistically significant after FDR correction.

Within the insomnia group, after FDR correction, only the negative correlation between PSMD and MMSE score remained statistically significant (r=−0.476, FDR-adjusted P=0.045). The correlations between PSMD and the DTI-ALPS index (r=−0.413, FDR-adjusted P=0.065) and between PSMD and BG-EPVS score (r=0.413, FDR-adjusted P=0.065) showed trends but did not survive FDR correction.

After additional adjustment for HAMD and HAMA scores, PSMD remained significantly positively correlated with BG-EPVS score in the full sample (r=0.433, FDR-adjusted P=0.0031) and significantly negatively correlated with the DTI-ALPS index (r=−0.314, FDR-adjusted P=0.0382), MoCA score (r=−0.298, FDR-adjusted P=0.0439), and MMSE score (r=−0.411, FDR-adjusted P=0.0040). These findings suggest that the associations between PSMD, glymphatic-related imaging markers, and cognitive performance were not fully explained by depressive or anxiety symptoms (Figure 6).

Figure 6 Partial correlation analysis between PSMD and glymphatic system-related imaging metrics as well as clinical cognitive measures in the full sample after controlling for age, sex, years of education, score of HAMD and HAMA. (A) Correlation between PSMD and DTI-ALPS index (r=−0.314, P=0.015); (B) Correlation between PSMD and BG-EPVS score (r=0.433, P<0.001); (C) Correlation between PSMD and MMSE score (r=−0.411, P=0.01); (D) Correlation between PSMD and MoCA score (r=−0.298, P=0.022). BG-EPVS, basal ganglia enlarged perivascular spaces; DTI-ALPS, diffusion tensor imaging analysis along the perivascular space; HAMA, Hamilton Anxiety Rating Scale; HAMD, Hamilton Depression Rating Scale; MoCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination; PSMD, peak width of skeletonized mean diffusivity.

Discussion

This study jointly applied PSMD, DTI-ALPS, and BG-EPVS scores to investigate white matter microstructural alterations and glymphatic-related imaging features in patients with primary insomnia. Patients with insomnia showed higher PSMD values, suggesting greater global white matter microstructural heterogeneity. As a diffusion MRI-derived marker, PSMD may sensitively capture diffuse microstructural alterations within the white matter skeleton. This finding supports the hypothesis that chronic insomnia may induce white matter pathological changes such as axonal injury or demyelination (33,34), while also aligning with evidence from animal experiments: chronic sleep deprivation can directly lead to reduced myelination and microglia-mediated neuroinflammation, thereby impairing white matter integrity (35,36). Importantly, PSMD reflects global white matter microstructural heterogeneity and is not a disease-specific biomarker. Elevated PSMD has been reported in various conditions including cerebral small vessel disease and Alzheimer’s disease (25,37). Therefore, the observed PSMD increase in patients with primary insomnia should be interpreted as evidence of diffuse white matter vulnerability rather than a pathognomonic signature of insomnia pathophysiology. Nonetheless, PSMD remains a sensitive quantitative metric for detecting insomnia-associated white matter damage. These findings may be further interpreted within an orexin-mediated hyperarousal framework (12). Excessive or mistimed orexinergic activity may sustain wakefulness, fragment sleep, and reduce restorative NREM sleep, thereby indirectly impairing sleep-dependent glymphatic clearance. In this context, increased BG-EPVS may reflect disturbed perivascular fluid dynamics, while the negative correlation between PSMD and DTI-ALPS suggests a link between white matter microstructural vulnerability and less efficient glymphatic transport. This study extends the application of PSMD, a sensitive but non-specific biomarker already validated in cerebrovascular diseases (38), to the field of insomnia research as a tool for quantifying white matter injury burden.

Although the PSMD difference between the insomnia and healthy control groups reached nominal statistical significance, this finding should be interpreted with caution. The effect size was modest, with a rank-biserial correlation coefficient of 0.25 and a 95% confidence interval of 0.00 to 0.47. In addition, the post hoc power analysis showed an achieved statistical power of only 0.485, indicating that the present sample size may not provide sufficient power to detect small-to-moderate between-group differences reliably. Therefore, the PSMD finding should be considered preliminary and requires confirmation in larger independent cohorts.

The negative correlations between PSMD and MoCA/MMSE scores suggest that greater white matter microstructural heterogeneity is associated with poorer global cognitive performance in patients with primary insomnia. This finding is consistent with previous evidence linking PSMD to cognitive function in other neurological conditions (23,39). Nevertheless, these associations should be interpreted cautiously because mood symptoms and other unmeasured factors may also contribute to cognitive performance.

Importantly, these associations remained significant after additional adjustment for HAMD and HAMA scores, suggesting that the observed relationships were not solely driven by depressive or anxiety symptoms. Nevertheless, given the higher HAMD and HAMA scores in the insomnia group, mood symptoms should still be considered important potential confounders.

After FDR correction, PSMD remained significantly associated with BG-EPVS, the DTI-ALPS index, and cognitive scores in the full sample. These associations also remained significant after additional adjustment for HAMD and HAMA scores, suggesting that white matter microstructural heterogeneity may be related to glymphatic-related imaging alterations and poorer cognitive performance. However, within the insomnia group, the associations of PSMD with DTI-ALPS and BG-EPVS did not survive FDR correction, and only the negative association between PSMD and MMSE score remained significant. Therefore, the relationship between white matter microstructural alterations and glymphatic-related imaging markers should be interpreted mainly based on the full-sample findings, whereas subgroup findings should be regarded as exploratory and interpreted cautiously.

Perivascular spaces are important pathways for cerebrospinal fluid and interstitial fluid exchange and are involved in the clearance of metabolic waste from the brain (40). EPVS have been considered indirect imaging markers related to glymphatic function and cerebral small vessel disease (41-43). In this study, the insomnia group showed higher BG-EPVS scores than the healthy control group in the unadjusted comparison. However, this association did not remain statistically significant after adjustment for age, years of education, sex, diabetes, hypertension, smoking, and alcohol consumption. Therefore, the BG-EPVS finding should be interpreted cautiously and should not be considered definitive evidence of insomnia-related glymphatic dysfunction. No similar increase was observed in CSO-EPVS scores, suggesting possible regional heterogeneity. Previous studies have suggested that BG-EPVS is more closely related to vascular risk factors and cerebral small vessel arteriolosclerosis (44,45), whereas CSO-EPVS may be more associated with β-amyloid clearance impairment, cerebral amyloid angiopathy, and Alzheimer’s disease (46). Future studies with larger samples, longitudinal designs, and more direct glymphatic or vascular imaging markers are needed to determine whether insomnia is independently associated with regional EPVS burden.

Although PSMD was negatively associated with DTI-ALPS, the absence of a significant between-group difference in DTI-ALPS indicates that convergent evidence for overt glymphatic dysfunction is limited. DTI-ALPS reflects water diffusion along perivascular spaces in selected white matter regions and may be influenced by local fiber architecture and methodological factors (47,48). Therefore, the correlation between PSMD and DTI-ALPS should be interpreted as a potential link between white matter integrity and glymphatic-related diffusion, rather than as proof of impaired glymphatic clearance.

Although the insomnia group showed increased BG-EPVS burden, no significant between-group difference was observed in the DTI-ALPS index. Therefore, our findings should not be interpreted as direct or definitive evidence of glymphatic dysfunction. EPVS enlargement is an indirect and non-specific imaging marker that may reflect altered perivascular fluid dynamics, vascular pathology, or other age and risk factor-related processes. In contrast, DTI-ALPS captures diffusion characteristics along perivascular spaces in specific white matter regions and may be less sensitive to subtle or early-stage alterations. Thus, the present results indicate partial and indirect glymphatic-related imaging alterations rather than convergent evidence of global glymphatic impairment.

This study did not observe a significant difference in the DTI-ALPS index between the primary insomnia and healthy control groups, which was not entirely consistent with our initial hypothesis. This negative finding may be partly attributable to the limited sample size and insufficient power to detect subtle changes. In addition, insomnia-related glymphatic alterations, if present, may be mild or compensatory and may not be readily captured by the DTI-ALPS index. The DTI-ALPS index reflects water diffusion along perivascular spaces in a specific region near the lateral ventricle, where projection and association fibers intersect (21). Therefore, local fiber architecture and structural asymmetry may influence its sensitivity. Glymphatic dysfunction is often more evident in conditions with pronounced neuropathological or vascular abnormalities, such as β-amyloid deposition, severe sleep apnea, or advanced cerebral small vessel disease (7,49,50). The absence of these major comorbid conditions in our participants may partly explain the non-significant DTI-ALPS result. The sensitivity of the DTI-ALPS index may also depend on whether white matter or vascular alterations reach a detectable threshold (51). Moreover, orexin-mediated hyperarousal may affect glymphatic-related processes through dynamic changes in sleep architecture, which may not be fully captured by a single cross-sectional DTI-ALPS measurement. Future studies with larger samples, longitudinal designs, objective sleep monitoring, and multimodal glymphatic imaging are needed.

Limitations

This study has several limitations. First, the cross-sectional design precludes causal inference; therefore, the relationships among insomnia, PSMD, EPVS, DTI-ALPS, and cognitive performance should be interpreted as associations rather than causal effects. Second, the sample size was relatively small and all participants were recruited from a single center, which may limit statistical power and generalizability. Although PSMD showed a nominally significant between-group difference, the post hoc power analysis indicated limited statistical power, and this result should therefore be considered preliminary. Third, insomnia diagnosis and severity assessment were based on subjective measures, including ISI and PSQI, without objective sleep assessments such as polysomnography or actigraphy. This limited our ability to evaluate sleep architecture, particularly deep sleep, and its relationship with glymphatic-related imaging markers. Fourth, EPVS visual scoring may involve some subjectivity, although inter-rater reliability was assessed. Future studies may benefit from automated or quantitative EPVS measurements. Fifth, PSMD is a global and non-specific marker of white matter microstructural heterogeneity, and elevated PSMD cannot be interpreted as specific to insomnia-related pathology. Finally, mood symptoms and vascular risk factors may still have residual confounding effects despite additional adjusted analyses. Future studies with larger samples, longitudinal designs, objective sleep monitoring, detailed cognitive assessments, and multimodal imaging are needed to validate and extend the present findings.

Conclusions

This cross-sectional study showed that patients with primary insomnia had higher PSMD values than healthy controls, whereas the DTI-ALPS index did not differ significantly between groups. Although BG-EPVS was higher in the unadjusted comparison, this association was attenuated and no longer statistically significant after adjustment for demographic and vascular risk factors. Higher PSMD was associated with poorer cognitive performance and lower DTI-ALPS values. These findings suggest that primary insomnia is associated with white matter microstructural vulnerability, whereas glymphatic-related imaging alterations should be interpreted as preliminary and indirect. However, causal relationships cannot be established from this study, and longitudinal studies are needed to determine whether these imaging changes are causes, consequences, or correlates of insomnia.

Acknowledgments

The authors would like to thank Edanz for providing expert editorial review and language support during the preparation of this manuscript.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0804/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0804/dss

Funding: This work was supported by the Foreign Cooperation Project of Fujian Science and Technology Program (No. 2025I0042), Science and Technology Innovation Platform Project of Fujian Science and Technology Plan (No. 2022Y2017) and the Science and Technology Innovation Joint Fund Program (No. 2025Y9720).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0804/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of 900th Hospital of PLA Joint Logistic Support Force (approval No. 2026-007). Individual consent was waived because of the retrospective nature of the study and the use of anonymized data.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Benjafield AV, Sert Kuniyoshi FH, Malhotra A, Martin JL, Morin CM, Maurer LF, Cistulli PA, Pépin JL, Wickwire EM. Estimation of the global prevalence and burden of insomnia: a systematic literature review-based analysis. Sleep Med Rev 2025;82:102121. [Crossref] [PubMed]
2. Jahrami HA, Alhaj OA, Humood AM, Alenezi AF, Fekih-Romdhane F, AlRasheed MM, Saif ZQ, Bragazzi NL, Pandi-Perumal SR. BaHammam AS, Vitiello MV. Sleep disturbances during the COVID-19 pandemic: A systematic review, meta-analysis, and meta-regression. Sleep Med Rev 2022;62:101591. [Crossref] [PubMed]
3. Zhang W, Yu M, Xu Y, Li X, Zuo H, Huang Z, Gao X. Self-reported sleep status and influencing factors: a web-based national cross-sectional survey in China. Ann Med 2023;55:2287706. [Crossref] [PubMed]
4. Sateia MJ. International classification of sleep disorders-third edition: highlights and modifications. Chest 2014;146:1387-94. [Crossref] [PubMed]
5. Zamore Z, Veasey SC. Neural consequences of chronic sleep disruption. Trends Neurosci 2022;45:678-91. [Crossref] [PubMed]
6. Liu XX, Wang Z, Chen SJ, Vitiello MV, Wing YK, Morin CM, Shi J, Lu L. Sleep health in China: status, challenges, and promotion strategies. Lancet Public Health 2025;10:e1055-65. [Crossref] [PubMed]
7. De la Herrán-Arita AK. When sleep fails, brain clearance suffers: the role of glymphatic impairment in clinical neurology. Acta Neurol Belg 2025; Epub ahead of print. [Crossref]
8. Hauglund NL, Nedergaard M. Is glymphatic clearance the secret to restorative sleep? Brain 2026;149:1819-31. [Crossref] [PubMed]
9. Chong PLH, Garic D, Shen MD, Lundgaard I, Schwichtenberg AJ. Sleep, cerebrospinal fluid, and the glymphatic system: A systematic review. Sleep Med Rev 2022;61:101572. [Crossref] [PubMed]
10. Deantoni M, Reyt M, Berthomier C, Muto V, Hammad G, De Haan S, Dourte M, Taillard J, Lambot E, Cajochen C, Reichert CF, Maire M, Baillet M, Schmidt C. Association between circadian sleep regulation and cortical gyrification in young and older adults. Sleep 2023;46:zsad094. [Crossref] [PubMed]
11. Hauglund NL, Andersen M, Tokarska K, Radovanovic T, Kjaerby C, Sørensen FL, Bojarowska Z, Untiet V, Ballestero SB, Kolmos MG, Weikop P, Hirase H, Nedergaard M. Norepinephrine-mediated slow vasomotion drives glymphatic clearance during sleep. Cell 2025;188:606-622.e17. [Crossref] [PubMed]
12. Mogavero MP, Silvani A, Lanza G, DelRosso LM, Ferini-Strambi L, Ferri R. Targeting Orexin Receptors for the Treatment of Insomnia: From Physiological Mechanisms to Current Clinical Evidence and Recommendations. Nat Sci Sleep 2023;15:17-38. [Crossref] [PubMed]
13. Guo AC, Jewells VL, Provenzale JM. Analysis of normal-appearing white matter in multiple sclerosis: comparison of diffusion tensor MR imaging and magnetization transfer imaging. AJNR Am J Neuroradiol 2001;22:1893-900.
14. Seehafer S, Larsen N, Aludin S, Jansen O, Schmill LA. Perivascular spaces and where to find them - MR imaging and evaluation methods. Rofo 2024;196:1029-36. [Crossref] [PubMed]
15. Kilinc D, da Costa ASFB, Barkhof F, Boomsma J, van Harten A, Exalto L, Moonen J, Wardlaw J, van der Flier W, Biessels GJ. The relationship between enlarged perivascular spaces and clinical outcomes in patients with vascular cognitive impairment. Cereb Circ Cogn Behav 2024;6:100277.
16. Kamagata K, Andica C, Takabayashi K, Saito Y, Taoka T, Nozaki H, Kikuta J, Fujita S, Hagiwara A, Kamiya K, Wada A, Akashi T, Sano K, Nishizawa M, Hori M, Naganawa S, Aoki SAlzheimer’s Disease Neuroimaging Initiative. Association of MRI Indices of Glymphatic System With Amyloid Deposition and Cognition in Mild Cognitive Impairment and Alzheimer Disease. Neurology 2022;99:e2648-60. [Crossref] [PubMed]
17. Schirge PM, Perneczky R, Taoka T, Ruiz-Rizzo AL, Ersoezlue E, Forbrig R, et al. Perivascular space and white matter hyperintensities in Alzheimer’s disease: associations with disease progression and cognitive function. Alzheimers Res Ther 2025;17:62. [Crossref] [PubMed]
18. Kang SH, Kim S, Kim YJ, Yoo H, Lee EH, Jang H, et al. Glymphatic dysfunction links vascular pathology to Alzheimer’s biomarkers and cognitive decline. Alzheimers Res Ther 2026;18:69. [Crossref] [PubMed]
19. Khalafi M, Shirbandi K, Zhou L, Butler TA, Chen K, Dartora WJ, Keil S, Li Y, Chiang GC. Diffusion Tensor Imaging along the Perivascular Space for Characterizing Cerebral Interstitial Fluid Dynamics in Alzheimer Disease: A Systematic Review and Meta-Analysis. AJNR Am J Neuroradiol 2026;47:401-8. [Crossref] [PubMed]
20. Xiong R, Feng J, Zhu H, Li C, Hu P, Zou Y, Zhou M, Wang Y, Tang X. Evaluation of glymphatic system dysfunction in patients with insomnia via diffusion tensor image analysis along the perivascular space. Quant Imaging Med Surg 2025;15:1114-24. [Crossref] [PubMed]
21. Taoka T, Ito R, Nakamichi R, Nakane T, Kawai H, Naganawa S. Diffusion Tensor Image Analysis ALong the Perivascular Space (DTI-ALPS): Revisiting the Meaning and Significance of the Method. Magn Reson Med Sci 2024;23:268-90. [Crossref] [PubMed]
22. Ringstad G. Glymphatic imaging: a critical look at the DTI-ALPS index. Neuroradiology 2024;66:157-60. [Crossref] [PubMed]
23. Baykara E, Gesierich B, Adam R, Tuladhar AM, Biesbroek JM, Koek HL, Ropele S, Jouvent E, Chabriat H, Ertl-Wagner B, Ewers M, Schmidt R, de Leeuw FE, Biessels GJ, Dichgans M, Duering M. A Novel Imaging Marker for Small Vessel Disease Based on Skeletonization of White Matter Tracts and Diffusion Histograms. Ann Neurol 2016;80:581-92. [Crossref] [PubMed]
24. Raposo N, Zanon Zotin MC, Schoemaker D, Xiong L, Fotiadis P, Charidimou A, Pasi M, Boulouis G, Schwab K, Schirmer MD, Etherton MR, Gurol ME, Greenberg SM, Duering M, Viswanathan A. Peak Width of Skeletonized Mean Diffusivity as Neuroimaging Biomarker in Cerebral Amyloid Angiopathy. AJNR Am J Neuroradiol 2021;42:875-81. [Crossref] [PubMed]
25. Luckey AM, Ghosh S, Wang CP, Beiser A, Bernal R, Li Z, et al. Biological validation of peak-width of skeletonized mean diffusivity as a VCID biomarker: The MarkVCID Consortium. Alzheimers Dement 2024;20:8814-24. [Crossref] [PubMed]
26. Cui RP, Hu HY, Han SL, Liang SY, Liu M, Gao PY, Chen YM, Wang YD, Guo F, Xue RJ, Tan MS, Hu H, Tan L. Association of Peak Width of Skeletonized Mean Diffusivity With Neurofilament Light Chain in Non-Dementia Adults. J Neurochem 2025;169:e70076. [Crossref] [PubMed]
27. Oberlin LE, Respino M, Victoria L, Abreu L, Hoptman MJ, Alexopoulos GS, Gunning FM. Late-life depression accentuates cognitive weaknesses in older adults with small vessel disease. Neuropsychopharmacology 2022;47:580-7. [Crossref] [PubMed]
28. Frey BM, Petersen M, Schlemm E, Mayer C, Hanning U, Engelke K, Fiehler J, Borof K, Jagodzinski A, Gerloff C, Thomalla G, Cheng B. White matter integrity and structural brain network topology in cerebral small vessel disease: The Hamburg city health study. Hum Brain Mapp 2021;42:1406-15. [Crossref] [PubMed]
29. Zhang W, Zhou Y, Wang J, Gong X, Chen Z, Zhang X, Cai J, Chen S, Fang L, Sun J, Lou M. Glymphatic clearance function in patients with cerebral small vessel disease. Neuroimage 2021;238:118257. [Crossref] [PubMed]
30. Duering M, Biessels GJ, Brodtmann A, Chen C, Cordonnier C, de Leeuw FE, et al. Neuroimaging standards for research into small vessel disease-advances since 2013. Lancet Neurol 2023;22:602-18. [Crossref] [PubMed]
31. Potter GM, Chappell FM, Morris Z, Wardlaw JM. Cerebral perivascular spaces visible on magnetic resonance imaging: development of a qualitative rating scale and its observer reliability. Cerebrovasc Dis 2015;39:224-31. [Crossref] [PubMed]
32. Wang XX, Cao QC, Teng JF, Wang RF, Yang ZT, Wang MG, Cao ZH. MRI-visible enlarged perivascular spaces: imaging marker to predict cognitive impairment in older chronic insomnia patients. Eur Radiol 2022;32:5446-57. [Crossref] [PubMed]
33. Fu X, Wan XJ, Liu JY, Sun Q, Shen Y, Li J, Mao CJ, Ma QH, Wang F, Liu CF. Effects of sleep fragmentation on white matter pathology in a rat model of cerebral small vessel disease. Sleep 2024;47:zsad225. [Crossref] [PubMed]
34. Li S, Tian J, Bauer A, Huang R, Wen H, Li M, Wang T, Xia L, Jiang G. Reduced Integrity of Right Lateralized White Matter in Patients with Primary Insomnia: A Diffusion-Tensor Imaging Study. Radiology 2016;280:520-8. [Crossref] [PubMed]
35. Simayi R, Ficiarà E, Faniyan O, Cerdán Cerdá A, Aboufares El Alaoui A, Fiorini R, Cutignano A, Piscitelli F, Maroto AS, Santos A, Del Gallo F, de Vivo L, De Santis S, Bellesi M. Sleep loss induces cholesterol-associated myelin dysfunction. Proc Natl Acad Sci U S A 2026;123:e2523438123. [Crossref] [PubMed]
36. Mao X, Han D, Guo W, Zhang W, Wang H, Zhang G, Zhang N, Jin L, Nie B, Li H, Song Y, Wu Y, Chang L. Lateralized brunt of sleep deprivation on white matter injury in a rat model of Alzheimer’s disease. Geroscience 2024;46:2295-315. [Crossref] [PubMed]
37. Hu Q, Zhou X, Xiao Z, Zhao Q, Ding D, Zhang J. White matter injury, plasma Alzheimer’s disease, and neurodegenerative biomarkers on cognitive decline in community-dwelling older adults: A 10-year longitudinal study. Alzheimers Dement 2025;21:e14594. [Crossref] [PubMed]
38. Maillard P, Lu H, Arfanakis K, Gold BT, Bauer CE, Zachariou V, et al. Instrumental validation of free water, peak-width of skeletonized mean diffusivity, and white matter hyperintensities: MarkVCID neuroimaging kits. Alzheimers Dement (Amst) 2022;14:e12261. [Crossref] [PubMed]
39. Hu HY, Li HQ, Gong WK, Huang SY, Fu Y, Hu H, Dong Q, Cheng W, Tan L, Cui M, Yu JT. Microstructural white matter injury contributes to cognitive decline: Besides amyloid and tau. J Prev Alzheimers Dis 2025;12:100037. [Crossref] [PubMed]
40. Nedergaard M. Neuroscience. Garbage truck of the brain. Science 2013;340:1529-30.
41. Brown R, Benveniste H, Black SE, Charpak S, Dichgans M, Joutel A, Nedergaard M, Smith KJ, Zlokovic BV, Wardlaw JM. Understanding the role of the perivascular space in cerebral small vessel disease. Cardiovasc Res 2018;114:1462-73. [Crossref] [PubMed]
42. Hayden MR, Tyagi N. The integrity of perivascular spaces is absolutely essential for proper function of the glymphatic system waste and excess water removal from the brain. Histol Histopathol 2025;40:1897-926. [Crossref] [PubMed]
43. Chen F, Heng T, Feng Q, Hua R, Wu J, Shi F, Liao Z, Qiao K, Zhang Z, Miao J. Quantitative assessment of brain glymphatic imaging features using deep learning-based EPVS segmentation and DTI-ALPS analysis in Alzheimer’s disease. Front Aging Neurosci 2025;17:1621106. [Crossref] [PubMed]
44. Riba-Llena I, Jiménez-Balado J, Castañé X, Girona A, López-Rueda A, Mundet X, Jarca CI, Álvarez-Sabin J, Montaner J, Delgado P. Arterial Stiffness Is Associated With Basal Ganglia Enlarged Perivascular Spaces and Cerebral Small Vessel Disease Load. Stroke 2018;49:1279-81. [Crossref] [PubMed]
45. Liu S, Hou B, You H, Zhang Y, Zhu Y, Ma C, Zuo Z, Feng F. The Association Between Perivascular Spaces and Cerebral Blood Flow, Brain Volume, and Cardiovascular Risk. Front Aging Neurosci 2021;13:599724. [Crossref] [PubMed]
46. Perosa V, Oltmer J, Munting LP, Freeze WM, Auger CA, Scherlek AA, van der Kouwe AJ, Iglesias JE, Atzeni A, Bacskai BJ, Viswanathan A, Frosch MP, Greenberg SM, van Veluw SJ. Perivascular space dilation is associated with vascular amyloid-β accumulation in the overlying cortex. Acta Neuropathol 2022;143:331-48. [Crossref] [PubMed]
47. Zhao D, Wang J, Zhang F, Wang Q, Zang M, Niu H, Tong Z. Cerebral glymphatic system: Structure, regulation, ageing, and mechanisms of encephalopathy. Ageing Res Rev 2026;114:102986. [Crossref] [PubMed]
48. Venkat P, Chopp M, Zacharek A, Cui C, Zhang L, Li Q, Lu M, Zhang T, Liu A, Chen J. White matter damage and glymphatic dysfunction in a model of vascular dementia in rats with no prior vascular pathologies. Neurobiol Aging 2017;50:96-106. [Crossref] [PubMed]
49. Iliff JJ, Wang M, Liao Y, Plogg BA, Peng W, Gundersen GA, Benveniste H, Vates GE, Deane R, Goldman SA, Nagelhus EA, Nedergaard M. A paravascular pathway facilitates CSF flow through the brain parenchyma and the clearance of interstitial solutes, including amyloid β. Sci Transl Med 2012;4:147ra111. [Crossref] [PubMed]
50. Bubu OM, Andrade AG, Umasabor-Bubu OQ, Hogan MM, Turner AD, de Leon MJ, Ogedegbe G, Ayappa I, Jean-Louis G G, Jackson ML, Varga AW, Osorio RS. Obstructive sleep apnea, cognition and Alzheimer’s disease: A systematic review integrating three decades of multidisciplinary research. Sleep Med Rev 2020;50:101250. [Crossref] [PubMed]
51. Zeng B, Zeng P, Xiang Y, Peng Y, Luo D, Li B, Chai Y, Wang D, Nie L, Li Y. Expansion of white matter hyperintensities is associated with the glymphatic system dysfunction in cerebral small vessel disease. Brain Res Bull 2025;227:111391. [Crossref] [PubMed]