Neurovascular Coupling Disruption and Glymphatic Dysfunction Are Associated With Visual Impairment in Pediatric Monocular Anisometropic Amblyopia

Abstract

Purpose

To investigate abnormalities in neurovascular coupling and the glymphatic metabolic clearance in children with monocular anisometropic amblyopia, as well as their associations with visual acuity (VA) status.

Methods

Multimodal magnetic resonance imaging scans were acquired from 33 children with anisometropic amblyopia and 31 age-matched normal controls. Neurovascular coupling was quantified as the spatial correlation between fractional amplitude of low-frequency fluctuations and cerebral blood flow across 90 cortical regions. Glymphatic metabolic clearance was evaluated using diffusion tensor image analysis along the perivascular space (DTI-ALPS). VA was assessed by measuring uncorrected VA (UCVA) and best-corrected VA (BCVA).

Results

Compared with controls, children with amblyopia exhibited altered neurovascular coupling in eight visual cortical regions (false discovery rate–corrected P < 0.05) and reduced DTI-ALPS indices in the hemisphere contralateral to the amblyopic eye (1.41 ± 0.13 vs. 1.55 ± 0.11; Bonferroni-corrected P < 0.001), with lower mean DTI-ALPS indices as well (1.46 ± 0.08 vs. 1.56 ± 0.08; Bonferroni-corrected P < 0.001). Neurovascular coupling in the contralateral middle and inferior occipital gyri showed significant negative correlations with both UCVA (r = −0.698, −0.694) and BCVA (r = −0.645, −0.663) of the amblyopic eye; similarly, the contralateral DTI-ALPS and mean indices demonstrated negative correlations with UCVA (r = −0.672, −0.580) and BCVA (r = −0.668, −0.554) of the amblyopic eye (all false discovery rate–corrected P < 0.05).

Conclusions

Neurovascular coupling and glymphatic clearance are altered in monocular anisometropic amblyopia and correlate with VA impairment, supporting their roles in the pathophysiology of amblyopia.

Amblyopia affects 2% to 4% of the global population, with anisometropia being one of its most frequent causes.^1,2 Although amblyopia is recognized as a cortical disorder that results from abnormal visual experience during development, its central mechanisms remain elusive. Notably, individuals with amblyopia exhibit abnormalities in local neural activity, particularly within the visual cortices and related brain regions.³ These neural disturbances may involve the brain's metabolic regulation, because adequate neural function depends on robust metabolic support.⁴

The proper functioning of neural circuits depends on finely regulated metabolic support through neurovascular coupling.⁵ This mechanism ensures that localized cerebral blood flow (CBF) aligns with neural activity, thereby facilitating the efficient delivery of oxygen and glucose to meet metabolic demands.⁶ Under normal physiological conditions, heightened neural activity initiates vasodilatory signals that increase regional CBF, establishing a tightly coordinated supply–demand relationship.⁷ In amblyopia, impaired visual input during the critical developmental period leads to dysregulation of neurovascular coupling within the visual pathways, altering metabolic homeostasis and neuronal function.^8,9 Furthermore, this metabolic dysregulation may extend to the glymphatic system, a perivascular network essential for clearance of metabolic waste through cerebrospinal fluid (CSF) circulation.¹⁰ The anatomical continuity between the retina, optic nerve, and CSF compartments has been proposed as a potential glymphatic clearance pathway.¹¹ CBF and astrocytic aquaporin-4 (AQP4) water channels influence glymphatic metabolic clearance.^12,13 The downregulation of AQP4 reduces perivascular fluid flow and impairs waste clearance efficiency.¹⁴ Elucidating this interaction is essential for understanding amblyopic pathophysiology and guiding future therapeutic strategies. Notably, neurovascular coupling dysregulation is documented in amblyopia; however, there is a lack of evidence supporting its impact on the glymphatic system, particularly in cases of monocular anisometropic amblyopia, where neural competition driven by refractive error may induce distinct metabolic remodeling.^9,15,16

Specific neuroimaging techniques enable assessment of these pathophysiological processes. Diffusion tensor image analysis along the perivascular space (DTI-ALPS) has emerged as a noninvasive approach to assess glymphatic metabolic clearance by quantifying water diffusion anisotropy in perivascular regions.^17,18 Recent applications of DTI-ALPS to assess glymphatic activity in thyroid eye disease and asthenopia demonstrate the practical utility of this method in neuro-ophthalmology.^19,20 Arterial spin labeling (ASL) enables noninvasive quantification of CBF.²¹ The fractional amplitude of low-frequency fluctuations (fALFF) serves as an effective measure of spontaneous neural activity derived from blood oxygenation level–dependent signals.²² The spatial correlation between the fALFF and CBF has been proposed as an approach to evaluate neurovascular coupling and has been validated as a potential marker of its integrity.^23,24 Our previous work established the specificity of this index to amblyopia, which initially identified neurovascular coupling variations in anisometropic and visual deprivation amblyopia.⁸ However, critical mechanism issues remain unresolved: (1) whether neurovascular decoupling defines a monocular anisometropic-specific endophenotype, thereby establishing a therapeutically targetable and localized biomarker, and (2) whether concomitant glymphatic dysfunction contributes to the disease pathophysiology.

To address these gaps, this study used multimodal magnetic resonance imaging (MRI) to characterize neurovascular coupling and glymphatic metabolic clearance in children with monocular anisometropic amblyopia. It specifically tested the hypothesis that both neurovascular coupling and DTI-ALPS indices are altered in this population, suggesting a potential dysregulation of neurovascular and perivascular interactions. Correlations between neuroimaging indices and visual acuity (VA) were analyzed to examine their associations with amblyopic visual deficits. By integrating these measures, this study sought to provide a specific pathophysiological model for this subtype.

Methods

Participants

This prospective case control study was conducted at the First Affiliated Hospital of Zhengzhou University from May 2022 to January 2025. Figure 1 illustrates the enrollment process. According to the inclusion and exclusion criteria, 71 participants were enrolled. Three participants failed to complete the MRI examination, and four participants were excluded owing to inadequate MRI image quality. Ultimately, a total of 33 children (33 amblyopic eyes and 33 fellow eyes), including 16 males and 17 females, with an average age of 9.76 ± 2.80 years, were diagnosed with monocular anisometropic amblyopia. Additionally, 31 individuals (31 dominant eyes and 31 nondominant eyes) with normal VA, comprising 15 males and 16 females with an average age of 9.84 ± 2.10 years, were included as the normal control (NC) group. The groups were matched for age and education level. Sample size calculation was based on the primary outcome measure: the neurovascular coupling index. A priori power analysis (G*Power 3.1) for two-sample t tests (α = 0.05, two-tailed; power = 0.80; effect size Cohen's d = 0.75) indicated that 29 participants per group were required. This effect size estimate was derived from a pilot study involving 12 participants and is consistent with published functional MRI (fMRI) studies in amblyopia.^8,25 We prespecified a 10% buffer to account for potential data loss, resulting in a target enrollment of 64 participants. Ultimately, 64 participants completed all MRI scans and VA assessments without major protocol deviations (amblyopia group: n = 33; NC group: n = 31). The slight imbalance in final group sizes was due to random variations in recruitment and did not materially affect statistical power. The study adhered to the Declaration of Helsinki and received approval from the First Affiliated Hospital of Zhengzhou University Scientific Research and Clinical Trial Ethics Committee (No: 2022-KY-0394-002).

Figure 1.

Diagram illustrating participant enrollment and data analysis for this study.

Inclusion Criteria

Children were required to meet the diagnostic criteria for anisometropic amblyopia, defined as a difference in spherical power diopter between the two eyes of ≥1.50 diopters (D) or a difference in cylindrical power diopter of ≥1.00 D. The NC group consisted of individuals with normal binocular vision, meaning that the uncorrected VA (UCVA) in both eyes was ≥0.8.

Exclusion Criteria

Individuals with neurological or mental disorders, organic lesions in the eyes or brain, optic nerve hypoplasia or atrophy detected by optical coherence tomography, prior amblyopia treatment, or previous surgical intervention were excluded.

Image Quality Standard

Participants with head movements exceeding 2° of rotation or 2 mm of translation were excluded.

Ophthalmic Assessments

IOP was measured using the CT-80A noncontact tonometer (TOPCON, Tokyo, Japan). UCVA and best-corrected VA (BCVA) were assessed at a distance of 5 m using a Chinese Tumbling E chart (decimal chart) and recorded in logMAR units. Topical tropicamide (0.5%) and phenylephrine hydrochloride (0.5%) (Shenyang Xingqi Pharmaceutical Co., Shenyang, China) were administered four times at 10-minute intervals, starting 40 minutes before the examination. After this, objective refraction was measured using an autorefraction system (RM-8000, TOPCON), and refractive error was expressed in terms of the spherical equivalent.

Imaging Acquisition

An MRI was performed using a Siemens Prisma 3.0T scanner equipped with a 64-channel head coil. Before MRI data acquisition, all participants underwent a standardized acclimatization protocol designed to minimize motion artifacts and anxiety-related confounds. This protocol included a 10-minute mock scanner simulation that replicated the acoustic environment and confined space of the actual MRI, allowing participants to practice remaining motionless while hearing gradient noise recordings. Research staff provided detailed explanations of the scanning procedure and coached participants on breathing techniques to promote relaxation. Immediately before the scan, participants received verbal instructions to lie still, remain awake with their eyes closed, and avoid structured or task-related mental activity during resting-state sequences. Head immobilization was achieved using foam padding and a restraint system to ensure data quality. The fMRI data were collected using a gradient echo planar imaging technique with these settings: echo time of 30.0 ms, repetition time of 1000 ms, matrix size of 110 × 110, field of view of 220 × 220 mm², voxel size of 2.0 × 2.0 × 2.2 mm³, slice thickness of 2.2 mm, flip angle of 70°, 52 slices, and a total of 400 time points. The entire fMRI scan lasted 6 minutes and 52 seconds. The ASL sequence was conducted using a three-dimensional fast spin-echo method, with parameters including an echo time of 16.18 ms, repetition time of 4600 ms, matrix size of 64 × 64, field of view of 192 × 192 mm², voxel size of 1.5 × 1.5 × 3.0 mm³, slice thickness of 3.0 mm, flip angle of 180°, and 40 slices. The total duration for the ASL scan was 4 minutes and 59 seconds. DTI data were acquired using the echo-planar imaging technique for axial acquisition, incorporating a total of 64 gradient directions. The parameters were as follows: echo time of 54.0 ms, repeat time of 9200 ms, matrix size of 128 × 128, field of view of 256 × 256 mm², voxel size of 2.0 × 2.0 × 2.0 mm³, slice thickness of 2.0 mm, no slice spacing, 70 slices, and b-values of 0 and 1000. The total scanning time was 10 minutes and 27 seconds. Additionally, T1-weighted images covering the whole brain were acquired using a three-dimensional high-resolution sagittal sequence with the following parameters: echo time of 2.32 ms, repetition time of 2300 ms, matrix size of 256 × 256, field of view of 240 × 240 mm², voxel size of 0.9 × 0.9 × 0.9 mm³, slice thickness of 0.9 mm, flip angle of 8°, and 176 slices. The full acquisition time for the T1-weighted imaging was 5 minutes and 21 seconds.

Data Preprocessing

fMRI Preprocessing

The DPARSF toolbox, running on the MATLAB R2019a platform (MathWorks, Natick, MA, USA), was used for preprocessing the fMRI images.²⁶ The first 10 time points were discarded, head motion correction was performed, and images were normalized to the standard Montreal Neurological Institute (MNI) spatial template and resampled to voxels of 3 × 3 × 3 mm³. Spatial smoothing was applied using a Gaussian kernel with a full width at half maximum of 6 mm. To correct for linear drift caused by thermal noise in the scanner, a correction was implemented. Additionally, signals from white matter and CSF were averaged separately, and nuisance covariates—including the Friston-24 head motion parameters—were regressed out.

CBF Preprocessing

CBF maps were generated by processing ASL images with ExploreASL v1.10.1.²⁷ After motion correction, registration to structural T1 images, and partial volume correction, the CBF maps were normalized to MNI space and smoothed with a 6-mm full width at half maximum Gaussian kernel. Subsequently, noise suppression was performed by removing voxels with physiologically implausible CBF values of <5 or >150 mL/100 g/min.²⁸

DTI Preprocessing

The dcm2nii toolbox was used to transform DTI DICOM files into NIFTI format. Subsequently, the data were analyzed using FSL v. 6.0.6 and MRtrix3. MRtrix3 was used for denoising and artifact correction, and FSL was applied to correct for susceptibility distortion, eddy currents, and head motion.

Analysis of Neurovascular Coupling

The preprocessed fMRI time series were transformed into the frequency domain using a fast Fourier transform to obtain the power spectrum. The fALFF was computed as the ratio of the integral of power spectral density within the low-frequency range (0.01–0.08 Hz) to the integral across the entire detectable frequency spectrum.²²

Spatial Registration and Resolution Matching

A unified two-step registration procedure was implemented to ensure precise voxel-wise correspondence between the fALFF and the CBF maps. First, individual CBF and fALFF maps were rigidly registered separately to each participant's T1-weighted structural image. Second, both modalities were warped to MNI standard space using the deformation fields derived from T1 normalization. To ensure voxel-wise correspondence, the normalized CBF maps were resampled to a 3 × 3 × 3 mm³ resolution using trilinear interpolation and masked by a binary mask representing the intersection of individually derived CBF and fALFF masks. These masks were generated in native space and transformed using identical registration parameters to ensure consistent brain coverage and minimize modality-specific edge effects.

Standardization and Correlation Calculation

Within the MNI standard space, both fALFF and CBF maps for each participant were Z-standardized individually using the mean and standard deviation of all voxels within the participant's whole brain gray matter mask, thereby excluding contributions from white matter and CSF to ensure that neurovascular coupling estimates reflect signals specific to neural tissue.^23,29 The brain was parcellated into 90 cortical and subcortical regions based on the Automated Anatomical Labeling atlas.³⁰ For each participant, neurovascular coupling was assessed by computing the correlation coefficient between the fALFF and CBF for each of the 90 regions. This approach quantifies the spatial coupling between neuronal activity and perfusion at the regional level. Correlation coefficients were Fisher Z-transformed for subsequent group-level statistical analyses.

DTI-ALPS Index Calculation

The DTI-ALPS index was calculated based on the diffusion anisotropy of the perivascular spaces.¹⁸ The analysis was conducted using the following procedure.

Diffusion Metric Calculation

The preprocessed DTI data were analyzed to generate fractional anisotropy (FA) maps and apparent diffusion coefficient maps along the x, y, and z axes. The x axis was defined as the left–right direction, the y axis as the anterior–posterior direction, and the z axis as the superior–inferior direction.

Spatial Registration

Individual FA maps were nonlinearly registered to the JHU-ICBM-FA template. The resulting transformation parameters were then applied to the directional diffusion coefficient maps to align all diffusion data within the JHU-ICBM-FA template space.

Region of Interest (ROI) Placement

Based on the JHU-ICBM-DTI-81 white matter labels atlas, four spherical ROIs, each with a diameter of 5 mm, were automatically placed in the bilateral projection and association fibers at the level of the lateral ventricle body, specifically in the superior corona radiata and superior longitudinal fasciculus. The center coordinates of the ROIs were as follows: bilateral superior corona radiata at (116, 110, 99) and (64, 110, 99), and bilateral superior longitudinal fasciculus at (128, 110, 99) and (51, 110, 99).³¹

Diffusion Coefficient Extraction and DTI-ALPS Index Calculation

For each participant, mean diffusion coefficients of the projection fibers and association fibers were obtained from four ROIs along the x, y, and z directional maps. The DTI-ALPS index was calculated for both the contralateral and ipsilateral hemispheres using the following formula (Equation 1)¹⁸:

$$\text{DTI-ALPS}=\frac{\mathrm{mean}(D_{x\_proj},D_{x\_assoc})}{\mathrm{mean}(D_{y\_proj},D_{z\_assoc})},$$ (1)

$\mathrm{where}{D}_{x\_proj}$and $D_{y\_proj}$ represent the diffusion coefficients of projection fibers along the x and y axes, respectively; and $D_{x\_assoc}$ and $D_{z\_assoc}$ represent the diffusion coefficients of association fibers along the x and z axes. The mean of the contralateral and ipsilateral DTI-ALPS indices is called the mean DTI-ALPS index.

Statistics Analysis

SPSS Statistics 23 software (IBM, Armonk, NY, USA) was used to analyze clinical indicators and demographic data. Normally distributed continuous variables were summarized as mean ± standard deviation and compared between groups using a two-sample t test. Non-normally distributed continuous variables were reported as median (interquartile range), with group comparisons conducted using Mann–Whitney U tests. Sex distributions, presented as frequency counts, were analyzed with χ² tests. All statistical tests used a significance threshold of 0.05. Neurovascular coupling values across 90 Automated Anatomical Labeling atlas regions were compared between groups using two-sample t tests, with false discovery rate (FDR) correction applied for multiple comparisons (P < 0.05). DTI-ALPS indices (contralateral, ipsilateral, and mean) were compared using two-sample t tests with Bonferroni correction (P < 0.05/3 = 0.017). To evaluate the associations between neurovascular coupling, the DTI-ALPS index, and VA measures in children with amblyopia, partial correlation analyses (FDR-corrected P < 0.05) were performed using age and sex as control variables. To compare within-group differences, paired-samples t tests (P < 0.05) were conducted. For participants with amblyopia, the tests compared the amblyopic eye with the fellow eye across all VA and imaging measures. For the NC group, the tests compared the dominant eye with the nondominant eye.

Reliability Assessment

To objectively validate the reliability of the derived DTI-ALPS index values and assess the precision and stability of neurovascular coupling coefficients across 90 brain regions, three independent analysts—blinded to participant identities and uninvolved in the initial experiment—reprocessed the entire cohort of 31 NCs using consistent software versions and a standardized analytical pipeline. Consistency and reliability across operators were quantified using a two-way random-effects model with a single-measurement intraclass correlation coefficient (ICC[2,1]). Additionally, the standard error of measurement (SEM) and the smallest detectable change (SDC) were calculated to quantify the absolute precision of individual measurements and establish the minimum threshold for distinguishing true biological changes from measurement error, respectively.

Results

Clinical Characteristics Results

No statistically significant differences were found in terms of sex, age, BCVA of the fellow/dominant eye, spherical equivalent of the fellow/dominant eye, or IOP between the two groups. Significant differences between the amblyopia and NC groups were observed in UCVA and BCVA of the amblyopic/nondominant eye, the UCVA of the fellow/dominant eye, and spherical equivalent of the amblyopic/nondominant eye. Details are presented in Table 1.

Table 1.

Demographic Characteristics of Participants

Conditions	Amblyopia	NC	χ² / t (DF) / Z	P Value
Male/female	16/17	15/16	0.00	0.994*
Age, years	9.76 ± 2.80	9.84 ± 2.10	−0.13 (62)	0.896†
UCVA of the amblyopic or nondominant eye, logMAR	0.94 ± 0.51	−0.01 ± 0.06	10.64 (32.86)	<0.001†
UCVA of the fellow or dominant eye, logMAR	0.14 ± 0.23	0.00 ± 0.07	3.31 (38.22)	0.002†
BCVA of the amblyopic or nondominant eye, logMAR	0.68 ± 0.40	−0.02 ± 0.05	9.80 (32.89)	<0.001†
BCVA of the fellow or dominant eye, logMAR	0.02 ± 0.09	−0.02 ± 0.06	1.94 (62)	0.057†
SE of the amblyopic or nondominant eye, D	4.00 (0.38 to 6.56)	0.25 (−0.25 to 0.75)	−4.08	<0.001‡
SE of the fellow or dominant eye, D	0.75 (−0.63 to 2.00)	0.38 (−0.50 to 1.05)	−1.29	0.197‡
IOP of the amblyopic or nondominant eye, mm Hg	16.00 (14.00 to 19.00)	15.00 (13.00 to 17.00)	−1.63	0.103‡
IOP of the fellow or dominant eye, mm Hg	17.00 (14.00 to 19.15)	16.00 (15.00 to 17.00)	−1.55	0.121‡
Amblyopic or nondominant eye: right/left	17/16	14/17	0.26	0.611*

DF, degrees of freedom; SE, spherical equivalent.

Values are mean ± standard deviation, number, or median (interquartile range).

^*χ² test.

^†Two-sample t test (with Welch's correction for unequal variances).

^‡Mann–Whitney U test.

Neurovascular Coupling Results

Significant between-group neurovascular coupling differences were identified in eight brain regions (FDR-corrected P < 0.05): the contralateral inferior occipital gyrus, contralateral middle occipital gyrus, contralateral calcarine fissure, contralateral lingual gyrus, ipsilateral calcarine fissure, ipsilateral middle occipital gyrus, ipsilateral inferior occipital gyrus, and ipsilateral temporal pole (middle temporal gyrus). Contralateral and ipsilateral are defined relative to the amblyopic (or nondominant) eye. Details are presented in Figure 2 and Supplementary Table S1.

Figure 2.

Spatial distribution maps of between-group differences in neurovascular coupling. (C) Represents the contralateral hemisphere to the amblyopic eye or nondominant eye; (I) Represents the ipsilateral hemisphere to the amblyopic eye or nondominant eye.

DTI-ALPS Index Results

In the amblyopia group, the contralateral, ipsilateral, and mean DTI-ALPS indices were 1.41 ± 0.13, 1.50 ± 0.10, and 1.46 ± 0.08, respectively. In the NC group, the corresponding values were 1.55 ± 0.11, 1.56 ± 0.11, and 1.56 ± 0.08, respectively. After Bonferroni correction with a significance threshold set at a P value of <0.017, children with amblyopia had significantly lower contralateral (P < 0.001) and mean (P < 0.001) DTI-ALPS indices. The ipsilateral reduction was not statistically significant (P = 0.025). Contralateral and ipsilateral are defined relative to the amblyopic (or nondominant) eye. Details are presented in Figure 3 and Supplementary Table S2.

Figure 3.

Differences in DTI-ALPS index between amblyopia and NC groups. Two-sample t test (with Welch's correction for unequal variances) with Bonferroni correction, at a threshold of P < 0.05/3 = 0.017. The contralateral/ipsilateral DTI-ALPS index is the DTI-ALPS index of the hemisphere contralateral or ipsilateral to the amblyopic or nondominant eye.

Correlation Analysis Results

After adjusting for age and sex, neurovascular coupling in contralateral occipital regions (middle/inferior occipital gyri) and DTI-ALPS indices (contralateral/mean) were negatively correlated with UCVA and BCVA of the amblyopic eye (all FDR-corrected P < 0.05) (Table 2). Additionally, the contralateral DTI-ALPS index was positively correlated with neurovascular coupling in the middle and inferior occipital gyri (FDR-corrected P < 0.05) (Fig. 4). Contralateral is defined relative to the amblyopic eye.

Table 2.

Partial Correlation Between the NVC, DTI-ALPS Index, and VA Measures in Amblyopia (n = 33)

		UCVA of the Amblyopic Eye, LogMAR	UCVA of the Fellow Eye, logMAR	BCVA of the Amblyopic Eye, Logmar	BCVA of the Fellow Eye, Logmar
NVC in the contralateral Occipital_Mid	R	−0.698*	−0.163	−0.645*	−0.132
	P	<0.001	0.581	<0.001	0.617
NVC in the ipsilateral Occipital_Mid	R	−0.371	−0.276	−0.256	−0.259
	P	0.106	0.254	0.287	0.287
NVC in the contralateral Occipital_Inf	R	−0.694*	−0.108	−0.663*	−0.144
	P	<0.001	0.669	<0.001	0.586
NVC in the ipsilateral Occipital_Inf	R	−0.393	0.075	−0.405	−0.116
	P	0.087	0.735	0.086	0.651
NVC in the contralateral calcarine fissure	R	−0.345	−0.083	−0.155	−0.102
	P	0.140	0.720	0.581	0.671
NVC in the ipsilateral calcarine fissure	R	−0.300	−0.336	−0.298	−0.176
	P	0.206	0.140	0.206	0.581
NVC in the ipsilateral Temporal_Pole_Mid	R	−0.338	−0.159	−0.160	−0.091
	P	0.140	0.581	0.581	0.702
NVC in the contralateral lingual gyrus	R	−0.369	−0.393	−0.154	−0.148
	P	0.106	0.087	0.581	0.585
Contralateral DTI-ALPS index	R	−0.672*	−0.126	−0.668*	−0.055
	P	<0.001	0.623	<0.001	0.800
Mean DTI-ALPS index	R	−0.580*	−0.017	−0.554*	−0.035
	P	0.002	0.926	0.004	0.869

NVC, neurovascular coupling; Occipital_Inf, Inferior occipital gyrus; Occipital_Mid, Middle occipital gyrus; Temporal_Pole_Mid, Temporal pole (middle temporal gyrus).

^*Represents a significant correlation after FDR correction, with P < 0.05. Contralateral and ipsilateral are defined relative to the amblyopic eye.

Figure 4.

Correlation between the DTI-ALPS index and NVC in the hemisphere contralateral to the amblyopic eye (n = 33). AE, amblyopic eye; NVC, neurovascular coupling; Occipital_Inf, inferior occipital gyrus; Occipital_Mid, middle occipital gyrus.

Interocular and Interhemispheric Differences

Within the amblyopia group, paired-samples t tests revealed significant differences in UCVA and BCVA between the amblyopic and fellow eyes, as well as hemispheric differences in DTI-ALPS values and neurovascular coupling in the inferior and middle occipital gyri (Supplementary Table S3). No such differences were found in controls (Supplementary Table S4).

Reliability Analysis Results

Reliability analysis demonstrated high consistency across the three independent analysts for both the DTI-ALPS index and the neurovascular coupling coefficients of 90 brain regions (Supplementary Table S5). For the DTI-ALPS index, the ICC(2,1) was 0.952 (95% confidence interval, 0.920–0.970), with a SEM of 0.016 and a SDC of 0.045. Similarly, the regional neurovascular coupling coefficients exhibited excellent interanalyst reliability, with ICC(2,1) values ranging from 0.905 to 0.976, SEM values from 0.004 to 0.051, and SDC values from 0.010 to 0.141. These results confirm the high precision and stability of the derived DTI-ALPS index and neurovascular coupling coefficients.

Discussion

This study demonstrates alterations in both neurovascular coupling and glymphatic metabolic clearance in children with monocular anisometropic amblyopia using multimodal neuroimaging. The study found that these children exhibit altered neurovascular coupling in multiple visual cortical regions and a reduced DTI-ALPS index. These abnormalities correlated with VA deficits in the amblyopic eye and showed predominant lateralization to the hemisphere contralateral to the amblyopic eye (Fig. 5). These findings extend the current understanding of amblyopia pathophysiology to include dysregulation of metabolic supply and waste clearance mechanisms.

Figure 5.

Diagram illustrating the relationship among neurovascular coupling, DTI-ALPS, and VA measures. The contralateral represents the contralateral hemisphere to the amblyopic eye, and the ipsilateral represents its ipsilateral hemisphere. Red represents the amblyopic eye and its contralateral hemisphere. Blue represents the fellow eye and the hemisphere ipsilateral to the amblyopic eye. Occipital_Inf, Inferior occipital gyrus; Occipital_Mid, Middle occipital gyrus.

Neurovascular coupling links neural activity to CBF, ensuring dynamic matching of oxygen and nutrient delivery to metabolic demand.^32,33 At the cellular level, neurovascular coupling is regulated by neuronal signaling and astrocytic mediation.^34,35 Neuronal firing induces ionic shifts and metabolic byproducts that trigger astrocyte release of vasoactive substances, modulating arteriolar tone, and local vasodilation.^36–38 Our previous work reported generalized neurovascular coupling deficits across both anisometropic and deprivation amblyopia.⁸ In contrast, the current study demonstrates that monocular anisometropic amblyopia exhibits a distinct neurovascular phenotype characterized by region-specific occipital decoupling—specifically in the middle and inferior occipital gyri—rather than diffuse cortical involvement. Such localized decoupling can alter the metabolic support requisite for synaptic plasticity and neural signal integration, processes essential for normal visual function, especially during development.^5,33 These findings challenge the assumption of uniform neurovascular pathology across amblyopia subtypes and suggest that neural competition driven by refractive errors induce distinct metabolic remodeling.

The glymphatic system facilitates the clearance of metabolic waste through perivascular fluid exchange.^13,39 This clearance mechanism is vital for brain and ocular health.^40,41 A reduced DTI-ALPS index in children with amblyopia suggests altered perivascular water diffusivity.^20,42 Alterations in neurovascular coupling reduce CBF dynamics, potentially slowing CSF circulation and metabolic clearance.^9,43 The positive correlation between DTI-ALPS and neurovascular coupling is consistent with a potential link between fluid transport and neurovascular decoupling. During visual development, balanced afferent signals are essential for synaptic pruning and neural circuit refinement.⁴⁴ Regionally heterogeneous astrocyte populations represent a common cellular substrate linking neurovascular coupling and the glymphatic system.^45,46 Astrocytes mediate both processes by transducing neural activity into vasoactive signals that regulate CBF and, concurrently, form the anatomical substrate of the glymphatic pathway through AQP4 channels clustered at perivascular endfeet.^13,47 In anisometropic amblyopia, reduced neural activity along the amblyopic pathway can induce morphological and functional remodeling of astrocytes.⁹ Astrocytic remodeling alters brain function through two interconnected pathways. First, it dysregulates vasoactive signaling and calcium dynamics, leading to altered neurovascular coupling.^37,48 Second, it causes a loss of AQP4 polarity, characterized by the redistribution of AQP4 from perivascular endfeet to the soma, which induces changes in glymphatic clearance.^49,50 These alterations form a bidirectional cycle: glymphatic dysfunction reduces the clearance of metabolic waste, exacerbating astrocytic and neurovascular variations, and altered neurovascular regulation further modifies the perivascular architecture essential for the glymphatic system.^51,52

Decreased VA in the amblyopic eye is associated with a reduction in the DTI-ALPS index and altered neurovascular coupling, specifically in the cerebral hemisphere contralateral to the affected eye. This finding reflects a close relationship between neural dysfunction and alterations in the brain's microenvironmental homeostasis. Although each cerebral hemisphere processes the contralateral visual field through binocular integration, this integration is functionally disrupted in amblyopia owing to chronic interocular suppression.^53,54 Critically, the visual input from each eye projects to both hemispheres via the partially crossed optic pathways. However, monocular inputs from the two eyes remain anatomically segregated into ocular dominance columns within V1.⁵⁵ Consequently, the partial decussation of retinal fibers at the optic chiasm means that the hemisphere contralateral to a given eye receives the dominant thalamocortical input from that eye via crossed nasal retinal projections, while also receiving weaker input from the ipsilateral eye through uncrossed temporal fibers. In anisometropic amblyopia, interocular suppression of signals from the amblyopic eye effectively renders cortical activation functionally monocular: the fellow eye dominates perception, while the amblyopic eye's contribution is attenuated or actively suppressed during binocular viewing.^53,56,57 This functional inhibition manifests as follows: the degraded input from the amblyopic eye selectively suppresses neural activity within its driven ocular dominance columns—specifically those representing the contralateral visual field in the contralateral hemisphere—thereby reducing local metabolic demand and causing corresponding reductions in CBF.^9,56,58,59 Dysregulation of neurovascular coupling is posited to impair the perfusion dynamics that facilitate glymphatic clearance, consistent with evidence linking cerebral perfusion to interstitial fluid dynamics.^43,60 Long-term neurometabolic disturbances impair CSF–interstitial fluid exchange through mechanisms such as altered astrocyte function and disrupted AQP4 polarization, potentially creating a cycle linking neurovascular decoupling and glymphatic dysfunction.⁶¹ Critically, the consistent lateralization of both DTI-ALPS reduction and hemodynamic alterations to the hemisphere contralateral to the amblyopic eye does not imply a direct eye-to-hemisphere correspondence. Rather, it reflects the spatial concentration of eye preference–specific circuits within the contralateral visual field map. This pattern, highlighted by the amblyopic eye, demonstrates that functional and microenvironmental changes co-occur within the hemisphere bearing the primary burden of disrupted afferent input.^9,25 Within-group analyses have indicated that this asymmetric pattern was significant in the amblyopia group, but was not observed as a similar interocular difference in the control group. This finding highlights the specificity of this pattern to monocular anisometropic amblyopia.

Limitations

Although this study innovatively explored the alterations in the glymphatic system and neurovascular coupling in children with monocular anisometropic amblyopia, several methodological and interpretive limitations must be acknowledged. First, the sample size was relatively modest, which may affect the generalizability and statistical power of the results. Second, this study was cross-sectional and cannot examine the temporal dynamics of changes in the glymphatic system and neurovascular coupling, nor can it clarify the causal relationships involved. In future studies, we will expand the sample size, conduct 12- and 24-month follow-ups of amblyopic children undergoing standard treatment, and investigate the directional relationships between these physiological indices and the recovery of visual function. Third, this study focused on the DTI-ALPS index and neurovascular coupling in specific brain regions, while both systems may involve multiple indicators and mechanisms. Fourth, the DTI-ALPS index was an indirect measure of glymphatic metabolic clearance and can be influenced by factors such as head motion, white matter integrity, and regional anatomical variations. Future studies should use a multimodal imaging framework that integrates more direct measures of glymphatic metabolic clearance (e.g., intrathecal contrast-enhanced MRI) with a broader range of neurovascular coupling indicators, enabling a comprehensive, multiparametric assessment.

Conclusions

Children with monocular anisometropic amblyopia exhibit significant alterations in both neurovascular coupling and glymphatic metabolic clearance, which are associated with VA loss. These findings indicate that anisometropic amblyopia involves dysregulation of metabolic supply and clearance mechanisms. The observed abnormalities offer potential pathophysiological insights and imply candidate biomarkers.