The emerging role of the meningeal lymphatic and glymphatic systems in migraine pathophysiology: a systematic review

Abstract

Background/objective

Migraine is a common, debilitating neurological disorder of unclear pathophysiology. Recent evidence has implicated impaired meningeal lymphatic function and its interaction with the glymphatic system in the development of neuroinflammation and pain sensitization. This systematic review aimed to summarize existing evidence on the role of the meningeal lymphatic and glymphatic systems in migraine pathophysiology.

Methods

Following the PRISMA 2020 recommendations, we searched PubMed, Web of Science, and Scopus from inception to December 15, 2025. Inclusion criteria comprised human or animal studies examining meningeal lymphatic or glymphatic function in migraine or validated migraine animal models. SYRCLE’s risk of bias tool for animal studies and the Joanna Briggs Institute (JBI) critical appraisal tools were used for quality assessment.

Results

The search yielded a total number of 457 records, out of which 10 articles fulfilled the eligibility criteria. These comprised six imaging studies conducted on human populations and four preclinical studies performed on animal models. Studies using dynamic contrast-enhanced (DCE)-MRI have shown changes in lymphatic enhancement characteristics in both episodic and chronic migraines. Diffusion tensor imaging along perivascular spaces (DTI-ALPS) yielded heterogeneous findings, with abnormalities more consistently observed in chronic and high-frequency migraine. Preclinical models demonstrated that cortical spreading depression, nitroglycerin exposure, and familial hemiplegic migraine mutations impaired glymphatic influx and reduced cerebrospinal fluid efflux through meningeal lymphatic vessels.

Discussion

Initial research has shown that altered lymphatic and glymphatic systems may be related to migraine. Nonetheless, there is no existing literature on whether these conditions can be considered causative agents for migraines. There is a need for longitudinal imaging studies in a larger population.

Registration

PROSPERO ID: CRD420251266296.

Introduction

Headache disorders are among the most common neurological conditions worldwide [1]. The Global Burden of Disease Study estimated that 2.9 billion individuals were affected in 2023, and headache-attributed years lived with disability were more than twice as high in females as in males [2]. Within this group, migraine is a recurrent primary headache disorder that causes substantial disability and socioeconomic burden. Global burden estimates consistently rank migraine among the top causes of years lived with disability, with especially high impact in adolescents and working-age adults, including young women [3, 4]. Even with effective acute and preventive options, many people still do not receive timely diagnosis or adequate treatment in many settings [1].

Current pathophysiological frameworks emphasize trigeminovascular system activation and neuropeptide signaling, particularly calcitonin gene-related peptide (CGRP), as core components of migraine pain generation and maintenance [5, 6]. Converging evidence also supports immune and neuroinflammatory contributions at the brain–meningeal interface that may stimulate and sensitize meningeal nociceptors (trigeminovascular afferents) and amplify pain [7–9]. In addition to its nociceptive role, CGRP infusion has been reported to provoke aura in susceptible individuals, suggesting involvement in early attack phases [10]. Inflammatory and neuropeptidergic mediators generated during migraine attacks are likely to accumulate at the brain–meningeal interface, making cerebrospinal fluid (CSF)-associated transport and meningeal outflow pathways biologically relevant to trigeminovascular activation. CSF solutes were shown to enter the trigeminal ganglion and directly activate trigeminal neurons, supporting a route by which CSF-borne mediators can engage trigeminal nociceptive signaling [11].

Over the past decade, the discovery of meningeal lymphatic vessels (MLVs) has reshaped understanding of central nervous system (CNS) fluid and immune trafficking. Glymphatic system is the perivascular route through which the exchange of cerebrospinal fluid occurs between interstitial fluid within the brain tissue. The meningeal lymphatic system (MLS) is the term used to refer to the collection of lymphatic vessels found in the dura mater layer that transport cerebrospinal fluid and immune cells towards deep cervical lymph nodes. It is important to note that these systems are separate structures and do not work together in one single functional capacity [11, 12]. Functional lymphatic vessels in the dura were found to connect to deep cervical lymph nodes and can transport CSF-associated solutes and immune cells [13]. In addition, a dural lymphatic vascular system that drains brain interstitial fluid and macromolecules further establishes the dura as an active interface for CNS–peripheral immune communication [14]. The MLS is closely linked to the glymphatic pathway [12]. Subsequent work identified skull-base lymphatic drainage routes that contribute to CSF outflow [15]. Non-invasive magnetic resonance imaging (MRI) studies have yielded evidence consistent with dural lymphatic networks and connections to cervical lymph nodes, supporting the translational relevance of these pathways [16]. More recent multimodal mapping defined ventral dural lymphatic architecture and drainage dynamics around the middle meningeal artery, broadening the anatomical framework for meningeal outflow pathways [17].

Across CNS disorders, studies increasingly point to meningeal lymphatic drainage as a workable pathway that shapes neuroimmune signaling and solute clearance [18]. Emerging migraine-focused studies suggest that meningeal lymphatic function may be directly involved in headache pathophysiology [11, 15]. Even so, the evidence remains scattered across experimental models and clinical studies. Accordingly, this systematic review aimed to synthesize and critically appraise current evidence for meningeal lymphatic or glymphatic systems function involvement in migraine pathophysiology.

Methods

This systematic review complied with the PRISMA 2020 recommendations [19]. The review has been registered in PROSPERO (ID: CRD420251266296).

Eligibility criteria

We included original studies (human or animal) that assessed the role of meningeal lymphatic or glymphatic system function in migraine pathophysiology, and we excluded reviews, case reports, conference abstracts, and non-English papers.

Information sources and search strategy

We conducted a comprehensive search via PubMed, Web of Science, and Scopus from inception to December 15, 2025. We executed the subsequent query: ((meningeal lymphatic) OR (dural lymphatic) OR (intracranial lymphatic) OR glymphatic OR ((CSF OR “cerebrospinal fluid”) AND (drain* OR clear* OR circulat* OR outflow* OR absorb*))) AND migraine*). The strategy was designed by two trained authors (A.N and R.M-H.). Besides, reference lists of included studies and relevant review articles were checked, and we created an alert for the PubMed and Scopus databases to update the search results monthly.

Selection process

Following the database searches, the acquired records were loaded into the Rayyan platform [20], and duplicate entries were removed before screening. Two reviewers (H.S and K.B) independently evaluated titles/abstracts and the full texts. All disputes were settled with a senior author (R.M-H.). EndNote 20 reference management software was utilized to arrange the citations.

Data collection process

We extracted data utilizing Microsoft Excel forms. Pairs of reviewers (N.S.D. with N.T., and H.S. with K.B.), separately gathered study characteristics, first author’s name, publication year, study setting and design, population/model, key mechanisms, and main findings on migraine pathophysiology. Any disagreements were resolved by the senior author (R.M.H).

Study risk of bias assessment

Two authors (N.S and R.M.H) independently appraised the included papers and any disagreements were reviewed by the third author (H.S.). The Systematic Review Centre for Laboratory Animal Experimentation (SYRCLE)’s risk of bias (RoB) tool, which is a modification of the Cochrane RoB tool, was utilized to evaluate the quality of animal studies. This tool consists of 10 items that address six types of bias [21]. The Joanna Briggs Institute (JBI) critical appraisal tools were also employed to evaluate the RoB in the clinical studies [22].

Synthesis methods

Given the heterogeneity in human MRI protocols, anatomical targets, and indirect imaging markers used to infer meningeal lymphatic function, as well as the diversity of preclinical model systems and outcome measures, a meta-analysis was not feasible. Therefore, the findings were synthesized narratively.

Results

Study selection

A total of 457 records were identified through database searches. After removal of 169 duplicates, 288 unique records were screened by title and abstract, and 268 were excluded. Twenty full-text articles were assessed for eligibility; ten were excluded—six did not evaluate meningeal lymphatic, glymphatic, or clearance-related outcomes [23–28], three were review articles [29–31], and one study was not based on animal models of migraine [32]—leaving six clinical [33–38] and four preclinical animal studies [39–42] for inclusion (Fig. 1).

Fig. 1.

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) 2020 flow diagram. Source: Page MJ, et al. BMJ 2021;372:n71. doi: 10.1136/bmj.n71. This work is licensed under CC BY 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/

Study characteristics and results of individual studies

Ten studies met the inclusion criteria: six human observational neuroimaging studies [33–38] (published in 2022–2025) and four preclinical studies [39–42] (published in 2017–2025). Clinical cohorts were recruited from headache clinics and included episodic and/or chronic migraine cases (total sample sizes n = 60–175; migraine n = 38–112), with mean ages in the mid-30s to 40 years and a female predominance (about 62.5–85% across migraine groups). Outcomes retrieved from human studies were grouped into three categories: dynamic contrast enhanced (DCE)-MRI, to assessing contrast enhancement dynamics within MLVs, diffusion tensor image analysis along the perivascular space (DTI-ALPS) as an indirect glymphatic system function marker, and other clearance-related MRI findings, such as enlarged perivascular spaces, white matter hyperintensities, and choroid plexus volume (Table 1). The preclinical studies investigated migraine-relevant models, including cortical spreading depression (CSD), nitroglycerin exposure, CGRP receptor signaling, and familial hemiplegic migraine type 2 (FHM2), to examine altered MLS function or ablation, lymphatic CGRP receptor signaling, and aquaporin-4 (AQP4)-associated glymphatic impairment (Table 2).

Table 1.

Characteristics and key findings of the clinical studies

Study	Design	Participants		MRI sequence	MLS-related assessment	Key findings	Risk of bias assessment
		Migraine and subtype (F/M) and their mean age or median IQR	HC (F/M) and mean age or median IQR
Ren, 2025 [33]	Human Observational (prospective case-control)	38 (All episodic without aura) (29/9), 37.92 ± 8.61	22 (13/9), 42.14 ± 10.6	3D T1-weighted (MP-RAGE / FSPGR	DCE-MRI	wash-in rate↓, Initial Area Under Curve↓, Ktrans in both mLV-SSS and mLV-SS ↓, Time-to-Peak↑	8/10
					DTI-ALPS	NS
					Structural MRI	Semiovale ↑, basal ganglia NS, midbrain NS
Wu, 2024 [34]	Human observational (prospective recruitment; cross-sectional multimodal MRI with clinical correlates).	112 37.0 ± 10.6 (73 episodic; 39 chronic including 22 without aura)	63, 39.7 ± 8.0	T2-weighted	DTI-ALPS	Chronic migraine ↓ Episodic migraine NS	8/8
					DCE-MRI	Time-to-Peak↑ in chronic EI↑ in chronic MTE↑ in chronic
Lee, 2022 [35]	Human observational (prospective enrollment; cross-sectional; interictal MRI)	92 (Drug-naïve episodic including 72 without aura) (72/20), 37.8 ± 11.7	80 (65/15), 38.9 ± 15.2	FLAIR	DTI-ALPS	NS	8/8
Zhang, 2023 [36]	Human observational, cross-sectional	56 (38/18), All without aura including 32 episodic (34.8 ± 15.4), and 24 chronic, 34.8 ± 16.8)	41 (25/16), 36.0 ± 10.9	SWI / SWAN	DTI-ALPS	Right in chronic migraine ↑ Right in episodic migraine NS Left in both chronic and episodic NS	7/8
Cai, 2025 [37]	Cross-sectional observational (Human; MRI-based)	74 (56/18, All episodic 46 with low frequency; 28 with high frequency), 35.0 (29.75, 38.25)	42 (30/12), 35.0 (24.75, 41.50)	TOF-MRA	DTI-ALPS	High frequency episodic migraine ↓ Low frequency episodic migraine NS	8/8
					White matter hyperintensity	Overall ↑ (periventricular ↑, deep NS)
					Perivascular space	basal ganglia NS, white matter NS
Lee, 2025 [38]	Prospective, cross-sectional (human; chronic migraine vs. episodic migraine vs. healthy controls)	59 (50/9), 39 episodic; 20 chronic), 37.7 ± 10.9	61 (52/9), 38.0 ± 8.1	DTI	Choroid plexus volume	Chronic migraine ↑ Episodic migraine NS	6/8

Abbreviations: EM, episodic migraine; DTI-ALPS: diffusion tensor imaging-along the perivascular space; MLV: meningeal lymphatic vessel; SSS: superior sagittal sinus; SS: sigmoid sinus; ↑ increased/higher; ↓ decreased/lower; NS, not statistically significant; TTP: time to peak; EI: enhancement integral; MTE: mean time to enhance

Table 2.

Characteristics of animal studies (n = 4)

Study	Design	Animals background and model	MLS/glymphatic assessment	Key finding(s)	Risk of bias assessment
Schain, 2017 [39]	Animal Experimental	C57BL/6J-background; Cortical spreading depression (model of migraine aura)	FITC-Dextran	Impaired glymphatic flow	High risk (sequence generation; caregiver/investigator blinding; outcome assessor blinding; other bias); other domains unclear
Huang, 2023 [40]	Animal Experimental	C57BL/6 male mice; nitroglycerin	Glymphatic influx of cerebrospinal fluid tracer	Reduced tracer influx, AQP4 expression, and AQP4 polarization; AQP4 inhibition further aggravated migraine-related pathological changes	Mostly unclear; low risk for assessor blinding (behavioral outcomes)
Nelson-Maney, 2024 [41]	Animal Experimental (chronic NTG + lymphatic CGRP-receptor manipulation) + Lab Experimental (human cells, in vitro)	C57BL/6J and CalcrliLEC; nitroglycerin	CSF drainage/uptake to meningeal lymphatic vessels	Lymphatic CGRP-receptor signaling reduced CSF uptake and altered immune endpoints	Mostly unclear; low risk for assessor blinding
Duan, 2025 [42]	Lab-Based (mouse in vivo imaging/tracing + in vitro astrocyte assays)	C57BL/6; FHM2 vs. wild-type	FITC-Dextran	FHM2 mice showed reduced tracer influx/clearance, greater PVS compression, and reduced AQP4 expression	Unclear across domains

Abbreviations: FHM2: familial hemiplegic migraine type 2; AQP4, aquaporin-4; CGRP, calcitonin gene-related peptide; CSF, cerebrospinal fluid

Results of syntheses, human neuroimaging findings

Migraine and DCE-MRI meningeal lymphatic enhancement/kinetics

Two studies used DCE-MRI to assess contrast enhancement dynamics within MLVs. Ren (2025) reported attenuated meningeal lymphatic enhancement kinetics in interictal episodic migraine without aura [33], whereas Wu (2024) observed delayed enhancement kinetics in chronic migraine compared with episodic migraine [34].

Migraine and DTI-ALPS glymphatic

Five studies assessed DTI-ALPS, an indirect MRI-derived marker of water diffusivity along perivascular spaces, considered measure for glymphatic-related transport. The findings were heterogeneous: two episodic migraine cohorts showed no difference from controls, two studies found lower indices in higher-burden migraine, and one chronic migraine study reported higher right-hemisphere ALPS values. Directionality therefore, differed across migraine subtypes and across studies. Lee (2022) [35] and Ren (2025) [33] reported no difference in episodic migraine, Cai (2025) reported lower indices in high-frequency episodic migraine [37], Wu (2024) reported lower indices in chronic migraine with medication-overuse headache [34], and Zhang (2023) reported higher right-hemisphere indices in chronic migraine [36].

Other clearance-related MRI markers

Three studies assessed ancillary MRI markers potentially associated with brain fluid homeostasis, perivascular dynamics, or neurovascular dysfunction. These markers included enlarged perivascular spaces, white matter hyperintensities, and choroid plexus volume. Although these measures do not directly assess glymphatic transport or meningeal lymphatic drainage, they may reflect structural or physiological alterations relevant to CNS fluid regulation. Collectively, the findings suggested abnormalities in these related pathways; however, the direction and clinical significance of the observed changes varied across studies. Ren (2025) [33], Cai (2025) [37], and Lee (2025) [38] reported findings related to enlarged perivascular spaces, white matter hyperintensities, and choroid plexus volume.

Results of syntheses, preclinical animal studies

Four preclinical studies examined migraine-relevant mechanisms in animal models. One study using CSD found acute impairment of glymphatic flow [39]. A study using the nitroglycerin model showed reduced glymphatic tracer influx and altered AQP4 expression and polarization [40]. A CGRP-receptor signaling study showed reduced CSF uptake into MLVs and altered immune endpoints [41]. An FHM2 model showed reduced tracer influx and clearance with greater perivascular space compression and lower AQP4 expression [42] (Table 2).

Risk of bias

Overall, the clinical cross-sectional studies demonstrated moderate-to-high methodological quality, with JBI scores ranging from 6/8 to 8/8. All five studies clearly defined inclusion criteria, described participants and settings, used valid and reliable exposure and outcome measures, applied objective diagnostic criteria, and employed appropriate statistical analyses. The main recurring limitation was confounding control: although confounders were generally identified, strategies to address confounding were not reported in Zhang (2023) [36], and both identification and management of confounding were absent in Lee (2025) [35], lowering confidence in interpretation (Table 3). The case–control study (Ren, 2025) [33] scored 8/10, with most domains satisfied (standardized identification of cases/controls; valid, consistent exposure measurement; confounding identified and addressed; adequate exposure window; appropriate analysis). However, it did not demonstrate baseline comparability between groups and did not report appropriate matching, introducing potential selection and comparability bias. Across animal experiments, SYRCLE judgments were predominantly unclear risk of bias across several domains, reflecting limited reporting of sequence generation, allocation concealment, random housing, and blinding. One study showed high risk of bias across multiple domains (Schain, 2017) [39], while the remainder were largely unclear with isolated low-risk judgements for specific outcomes (e.g., assessor blinding for behavioral outcomes in Huang, 2023 [40]; Nelson-Maney, 2024 [41]) (Table 4).

Table 3.

The results of risk of bias assessments based on the JBI critical appraisal tools

Authors (year)	Were the criteria for inclusion in the sample clearly defined?	Were the study subjects and the setting described in detail?	Was the exposure measured in a valid and reliable way?	Were objective, standard criteria used for measurement of the condition?	Were confounding factors identified?	Were strategies to deal with confounding factors stated?	Were the outcomes measured in a valid and reliable way?	Was appropriate statistical analysis used?
Zhang, 2023	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes
Lee, 2022	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Cai, 2025	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Wu, 2024	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Lee, 2025	Yes	Yes	Yes	Yes	No	No	Yes	Yes

Table 4.

SYRCLE risk of bias assessment for included animal studies

Study	Sequence Generation	Baseline Characteristics	Allocation Concealment	Random Housing	Blinding (Caregivers/ Investigators)	Random Outcome Assessment	Blinding (Outcome Assessor)	Incomplete Outcome Data	Selective Reporting	Other Bias	Overall
Schain, 2017	High	Unclear	Unclear	Unclear	High	Unclear	High	Unclear	Unclear	High	High
Duan, 2025	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Huang, 2023	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Low (behavioral) / Unclear (others)	Unclear	Unclear	Unclear	Unclear
Nelson-Maney, 2024	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Low	Unclear	Unclear	Unclear	Unclear

Discussion

In interictal scans of episodic migraine cohorts, DCE-MRI demonstrated diminished wash-in rates in MLVs along the superior sagittal sinus and extended time-to-peak, with the initial area under the curve also significantly reduced. The modifications exhibited an inverse correlation with headache intensity, independent of age and sex in multivariate analysis. Alterations in CNS fluid-regulation pathways have been reported in chronic migraine; however, the specific compartment involved may depend on the imaging modality used. DTI-ALPS findings indices have been associated with greater migraine severity and symptom burden [33]. On the other hand, Wu et al. demonstrated that, DTI-ALPS values were diminished in chronic migraine. No significant differences were observed between the presence and absence of headache during the MRI for all migraine types, episodic migraine, chronic migraine with medication-overuse headache, and chronic migraine without medication-overuse headache in this study [34].

Although glymphatic disruption is known to occur in neurodegenerative conditions, there is no correlation between clinical symptoms of migraines and glymphatic functioning measures, especially during interictal and daytime phases [35, 43]. Nevertheless, knowledge regarding mechanisms of action demonstrates that activation of the trigeminovascular pathway and CGRP plays an essential role in the physiology of migraine attacks, whereby CGRP-mediated vasodilation, nitric oxide signaling, and neuroinflammatory processes may influence glymphatic-associated transport; however, the direction and clinical significance of these effects in migraine remain uncertain [36, 44]. Furthermore, structural parameters like the presence of enlarged perivascular spaces, along with associations with headache intensity, attacks’ frequency, and neuropsychiatric disturbances, point towards possible development of glymphatic disturbance in chronic migraine phenotypes. It has been hypothesized that altered fluid homeostasis may contribute to metabolite accumulation, neuroinflammation, and central sensitization in some migraine phenotypes [37]. Recurrent CGRP-mediated neurovascular and inflammatory processes, along with impaired clearance of inflammatory mediators. Such mechanisms remain hypothetical and require validation in longitudinal and mechanistic studies.

Similarly, in C57/BL6 mice, Huang et al. demonstrated that the glymphatic influx of CSF tracer was diminished in mice with the migraine, along with lower expression and compromise of AQP4, suggesting impaired glymphatic-associated transport in the migraine mouse model. Subsequent inhibition of glymphatic activity by TGN-020 (an AQP4 antagonist) exacerbated the pathological alterations associated with migraines in mice [40]. Duan et al. also reported marked abnormalities in glymphatic-associated transport in FHM2 animals, with inappropriate neurovascular coupling largely contributing to reduced AQP4 expression in astrocytes [42]. In an animal model study, Mice without the CGRP receptor in their lymphatic vasculature (CalcrliLEC mice) that were subjected to nitroglycerin-mediated chronic migraine demonstrate less pain and light aversion relative to nitroglycerin-treated littermate controls. The gene expression profiles of lymphatic endothelial cells extracted from the meninges of Rpl22HA/+;Lyve1Cre RiboTag animals administered with nitroglycerin demonstrated heightened MLV-immune interactions relative to cells from untreated mice. The relative number of mucosal vascular addressing cell adhesion molecule 1–interacting (MAdCAM1-interacting) CD4 + T cells was elevated in the deep cervical lymph nodes of nitroglycerin -treated control mice, but not in nitroglycerin -treated CalcrliLEC animals. In vitro treatment of cultured human limbal epithelial cells with CGRP peptide prompted rearrangement of vascular endothelial–cadherin and decreased functional permeability [41].

Furthermore, peripherally activated T-cells can traverse the blood-brain barrier into the CNS and engage with neurons or glial cells. Interferon-gamma generated by Th1 cells enhances neural sensitivity to nociceptive signals, whereas diminished quantities or compromised functionality of regulatory T cells may attenuate inflammation suppression, facilitating the onset of chronic migraines [45]. In another animal study, authors showed that CSD in FHM1 mutant mice, mouse grimace scale scores normalized at 48 h, but in WT mice, they had already reverted to baseline at 24 h. At the 24-hour mark, contralateral neuroinflammation was elevated in FHM1 mutants compared to wild-type mice. At 48 h, neuroinflammatory responses in FHM1 mutant mice had reverted to aberrant, increased baseline levels relative to WT mice [8]. Findings complement established CSD models in migraine aura, where optogenetic induction in familial hemiplegic migraine mice triggers prolonged inflammation and behavior changes. CSD literature needs to be considered with caution because there are differences in endpoints being investigated. In Schain et al., it was shown that CSD results in the quick closure of the perivascular space and reduces glymphatic transport in mice [39]. The discrepancy may reflect differences in timing, endpoint selection, and experimental readout rather than true biological contradiction [39]. Besides, Schain et al. showed that CSD affects glymphatic clearance based on the decreased speed at which dye injected into the parenchyma was removed from the cortex into the PVS in connection with an abnormal cortical activity related to a neurological condition [39]. In contrast, in Kaag Rasmussen et al., the direct activation of trigeminal ganglion cells by CSF molecules as well as changes in CSF protein content. It is followed by directing flow into the venous perivascular and perineuronal spaces, resulting in the removal of the solutes into meningeal and cervical lymphatics from the neuropil. In mice, glymphatic flow occurs mainly during sleep when there is a doubling of clearance of toxic solutes like amyloid beta (Aβ) compared to the wakeful state [46].

Limitations and future directions

There are some limitations in the existing body of evidence. Clinical evidence is characterized by small sample sizes, and an abundance of single-center cross-sectional studies. Substantial heterogeneity in migraine phenotypes, MRI acquisition protocols, image-processing methods, and outcome measures further limits direct comparison across studies. An additional limitation relates to the interpretation of DTI-ALPS findings. Although the ALPS index is widely used as an indirect MRI-derived biomarker of glymphatic function, it does not directly measure glymphatic flow, CSF clearance, or meningeal lymphatic drainage. The precise physiological basis of the ALPS index remains incompletely understood, and its validity as a measure of glymphatic function, particularly in migraine populations, continues to be investigated. Consequently, findings derived from DTI-ALPS should be interpreted with caution, and inferences regarding specific clearance pathways should be considered preliminary until validated by complementary imaging modalities and longitudinal studies. In the future, longitudinal human imaging research with DCE-MRI, as well as validation of ALPS with a more objective method of measuring drainage, is necessary. In addition, preclinical studies require blinding and a better definition of bias domains.

Conclusion

This systematic review suggests that alterations in meningeal lymphatic drainage, glymphatic-associated transport, and perivascular fluid dynamics may be associated with migraine pathophysiology. However, the available evidence remains preliminary, largely cross-sectional, and insufficient to establish causality or definitive mechanistic relationships.

Author contributions

Kimia Bagheri: Methodology; Investigation; Writing - Original Draft. Hadi Sahrai: Methodology; Investigation; Writing - Original Draft; Writing - Review & Editing. Negin Safari Dehnavi: Investigation; Writing - Original Draft. Niloofar Taheri: Investigation; Writing - Original Draft; Visualization. Ali Norouzi: Investigation; Writing - Original Draft; Writing - Review & Editing. Mahnaz Talebi: Methodology; Validation; Investigation; Writing - Review & Editing; Supervision. Reza Mosaddeghi-Heris: Conceptualization; Methodology; Validation; Investigation; Writing - Review & Editing; Visualization; Supervision; Project administration; Funding acquisition. Amirreza Naseri: Methodology; Validation; Investigation; Writing - Review & Editing; Supervision; Project administration; Funding acquisition.

Funding

The research was supported by the Student Research Committee, Tabriz University of Medical Sciences (grant number: 78799).

Data availability

The extracted data supporting the conclusions of this article are included within the article.

Declarations

Ethical approval

This systematic review synthesizes published data and did not involve primary human or animal research. All included studies were assessed for reported ethical compliance, with human studies noting institutional review board approvals and informed consent, and animal studies adhering to ARRIVE guidelines where specified. The protocol for this study was approved by the ethics committee of Tabriz University of Medical Sciences (Ethics code: IR.TBZMED.VCR.REC.1404.385).

AI declaration

The authors used ChatGPT [Large language model] by OpenAI for language editing of the initial draft. The final content was reviewed and approved by the authors, who take full responsibility for its validity.

Competing interests

The authors declare no competing interests.