2025 Highlights - progress towards discovery of molecular and neuroimaging biomarkers for migraine

Biomarkers are measurable indicators of biological processes that can shed light on disease mechanisms, improve diagnosis, and guide treatment monitoring. The search for reliable biomarkers in migraine remains one of the most active frontiers in headache research. Despite major advances in our understanding of migraine biology, translation into clinically relevant biomarkers, whether molecular or imaging, remains an unmet goal. In recent years, the combination of large, well-characterized patient cohorts and increasingly sensitive analytical platforms have begun to change this landscape. The year 2025 has been particularly productive, with several studies providing new and robust evidence on both the biological and neuroimaging correlates of migraine.

The search for molecular biomarkers is increasingly focused on indicators that may reflect trigeminal and autonomic activation, most notably calcitonin gene-related peptide (CGRP) and pituitary adenylate cyclase-activating polypeptide (PACAP), as well as markers that could provide a direct or indirect view of the underlying neuroinflammatory processes. CGRP remains the most promising biomarker in the migraine field and has consistently been found to be elevated in the plasma, serum, and cerebrospinal fluid of patients with migraine. In recent years, alternative biological matrices closer to the trigeminovascular system, such as saliva and tear fluid, have been explored as potentially more sensitive sources for detecting CGRP (1,2).

Among these, tear fluid has emerged as a promising matrix. In a recent study, CGRP levels measured in tears collected with a capillary tube were significantly higher in patients with migraine than in healthy controls. Moreover, in 17 individuals with episodic migraine, tear CGRP concentrations significantly increased during glyceryl trinitrate-induced experimental headache compared to baseline and significantly decreased after headache resolution, especially in those who received acute treatment (1).

Another study, using Schirmer strips to collect tear fluid, similarly demonstrated elevated tear CGRP concentrations in patients with migraine (both episodic and chronic) compared to healthy controls (2).

Notably, these levels decreased significantly after six months in five patients treated with monoclonal antibodies targeting the CGRP ligand (2). Altogether, these findings support the potential of tear fluid CGRP measurement as a novel non-invasive biomarker approach in migraine (1,2).

Translational experiments have also continued to bridge molecular findings with pathophysiological models. In a double-blind, placebo-controlled crossover study, Al-Khazali et al. (3) demonstrated that intravenous infusion of substance P provoked headache and superficial temporal artery dilation in 15 (71%) of 21 healthy participants. This work establishes substance P as another key component of the trigeminovascular system capable of triggering headache, paralleling the effects previously shown for CGRP and PACAP.

Additional important insights into both molecular and imaging biomarkers have also emerged from the Registry for Migraine (REFORM) (4). In a study including 603 individuals with migraine and 154 healthy controls, Gozalov and colleagues (4) quantified serum concentrations of glial fibrillary acidic protein (GFAP), reflecting astroglial activation, and neurofilament light chain (NfL), indicating axonal injury, using an ultrasensitive single-molecule array platform. The authors found slightly higher GFAP levels in the migraine group, whereas NfL concentrations remained unchanged. Notably, neither GFAP nor NfL levels differed according to migraine subtype (episodic vs. chronic; with vs. without aura) or headache status at the time of blood sampling (ictal vs. interictal) (4). The finding of elevated GFAP, although modest, and independent from migraine subtype, supports the idea that migraine involves low-grade glial activation and possibly subtle blood-brain barrier alterations, that, however, have not been consistently demonstrated in migraine.

Although GFAP is considered a CNS-derived marker of astrocytic activation, recent evidence indicates that satellite glial cells in the trigeminal ganglion may also express GFAP under inflammatory conditions. Therefore, circulating GFAP elevations in migraine could theoretically arise from both central and peripheral glial sources (4).

Supporting the pivotal contribution of the trigeminovascular system to migraine pathophysiology, Tohyama et al. (5) employed a multimodal imaging framework integrating ultra–high-field 7 T diffusion tensor imaging, functional magnetic resonance imaging (fMRI), and 11C-PBR28 positron emission tomography (PET). Through this approach, the authors identified microstructural alterations of the trigeminal root entry zone accompanied by elevated neuroinflammatory signal and a reduced functional response of the spinal trigeminal nucleus during innocuous trigeminal stimulation. This convergent structural–inflammatory–functional pattern reinforces the concept that trigeminal system remodeling is a central feature of migraine pathophysiology and represents a promising avenue for biomarker development (5).

Beyond the activation of the trigeminovascular system and neuropeptide release, migraine has increasingly emerged as a complex brain disorder involving widespread neural networks (6). Accumulating neuroimaging evidence reveals dynamic functional and structural alterations within cortical and subcortical regions implicated in pain and visual processing, occurring across different phases of the migraine cycle and among distinct patient subtypes. Together, these findings suggest a continuum between transient, state-dependent plasticity and enduring network reorganization that may shape individual vulnerability to migraine attacks (7).

Another recent application of MRI has examined whether glymphatic dysfunction contributes to migraine. Although previous findings remain mixed, the study by Ornello et al. (8) using diffusion tensor imaging along the perivascular space (DTI-ALPS) provides a novel perspective that may help clarify prior conflicting results. Their findings showed that lower DTI-ALPS values, reflecting reduced glymphatic function, were linked to a higher number of monthly headache days. Notably, this association was evident only in individuals with poor sleep quality, suggesting that sleep disruption may heighten the impact of impaired perivascular fluid dynamics. Overall, these findings support glymphatic MRI-derived metrics as a promising, though still evolving, biomarker avenue in migraine (8).

Beyond advancing our understanding of migraine pathophysiology, neuroimaging is shedding light on how treatments act on the migraine brain. Using a fMRI approach mapping neurotransmitter-related activity, Fedeli et al. (9) showed that a six-week mindfulness program, added to standard care for chronic migraine and medication overuse headache, including withdrawal of overused drugs, tailored prophylaxis, and patient education, increased the functional coupling within serotonin- and dopamine-related brain networks, particularly in regions involved in pain modulation and emotional regulation, such as the caudate, accumbens, and insula. Along the same lines, Szabo et al. (10) demonstrated that a three-month treatment with galcanezumab reduced the functional connectivity between temporal and occipital regions involved in visual processing, as well as between these areas, the insula, and sensorimotor cortices, in responders with and without aura. Together, these studies suggest that both pharmacological and behavioral interventions can modulate cortical networks involved in the processing of pain and visual stimuli, providing novel insights into the neurobiological bases of therapeutic effect (10). However, as neither study reported significant baseline differences in functional connectivity between the patient groups examined, and no healthy control group was included for comparison, it remains difficult to determine whether these longitudinal changes reflect a true normalization of brain network function (10).

In recent years, the establishment of large, well-characterized patient cohorts and the refinement of analytical methods have gradually improved the quality of molecular biomarker research in migraine, although variability in assay performance and standardization remains a major challenge.

Looking ahead, molecular profiling in migraine will likely benefit from ultrasensitive assays capable of detecting low-abundance circulating proteins, as well as from the exploration of alternative biological matrices, such as tear fluid and saliva, that may more closely mirror underlying pathophysiological processes of migraine and may offer non-invasive methods of collection.

At the same time, neuroimaging is redefining migraine as a dynamic disorder of network plasticity and offering a framework to identify imaging biomarkers that capture its pathophysiology and the effects of migraine treatments. Emerging evidence of visual-nociceptive network interplay across migraine with and without aura points to shared cortical mechanisms and underscores the need for further studies to disentangle their mechanistic diversity. Despite growing efforts to perform imaging studies with larger and methodologically rigorous cohorts, additional large-scale investigations will be essential. Integrating multimodal imaging with molecular and behavioral data will be pivotal in driving the field toward mechanism-based, personalized migraine care.