2025 Highlights in central and peripheral interactions in migraine

A long-standing debate in migraine pathophysiology focuses on whether migraine should be considered mainly as a central or peripheral neurological disorder, and particularly whether the initial event in migraine attacks has a central or peripheral origin. Papers published in 2025 have contributed further to the increasingly nuanced view that migraine is characterized by dynamic crosstalk between peripheral generators and central networks and cannot be defined by a single anatomical locus. Several original studies published in Cephalalgia during this year have offered additional insights into whether migraine influences brain structure and exemplified the notion of central–peripheral interaction by linking peripheral modulation (vascular ion channel opening and anti-calcitonin gene-related peptide (CGRP) monoclonal antibodies) to central effects (nociceptor activation and network remodeling). Finally, the long-standing debate about whether cortical spreading depression (CSD) is the cause of, or a parallel pathway to, headache was examined by a provocative Viewpoint that revisited the relationship between CSD, aura and pain. Taken all together, these studies provide further nuance to the central versus peripheral debate, arguing for an integrated model and offering novel points of view that could inspire future research.

Peripheral effects with central consequences

The advancements in the neuroimaging technologies over the last decade have led to new insights and an increasing understanding of the central consequences of migraine attacks. Using brain magnetic resonance imaging (MRI) with diffusion tensor imaging (DTI), Gollion et al. (1) found no differences in white matter integrity between participants with migraine and healthy controls. Interestingly, no differences emerged between participants scanned during the ictal phase and participants scanned in the headache-free period, nor among various migraine subtypes, supporting the notion that migraine, even during the pain phase, does not cause microstructural changes in cerebral white matter.

A narrative review conducted by an international panel of experts on brain morphometry imaging studies, including surface-based morphometry (SBM) and voxel-based morphometry (VBM), provided additional insights into how migraine affects the central nervous system. Despite some inconsistencies across the studies, an increased thickness was observed in several pain-processing regions, while cortical thinning emerged in other brain regions, including the orbitofrontal cortex, posterior cingulate cortex and visual cortex. Notably, preventive treatments with drugs targeting the CGRP pathway or botulinum toxin A seem to revert some of these cortical alterations, suggesting neuroplastic remodeling in treatment responders (2).

Along these lines, two original articles probed, from complementary angles, how peripherally constrained interventions or perturbations exhibit central effects. First, Szabo et al. (3) showed that treatment with a peripherally acting anti-CGRP monoclonal antibody (galcanezumab) attenuates cortical resting-state connectivity selectively in clinical responders, particularly across sensorimotor–insular–occipital circuits, implying that reducing nociceptive afferent “barrage” can drive a central recovery toward a less excitable cortical state. This longitudinal functional MRI result fits well with the accepted pharmacology that anti-CGRP mAbs have minimal blood–brain barrier (BBB) penetration under normal conditions; their primary action is believed to be peripheral, yet they can secondarily reshape central networks (3).

Second, Christensen et al. (4) provided mechanistic evidence that opening ATP-sensitive potassium (K_ATP) channels activates meningeal nociceptors in vivo, offering a plausible anatomical site and pathway by which levcromakalim infusions trigger migraine attacks in humans. They recorded single-unit activity in trigeminal ganglion neurons and observed dose-dependent activation of meningeal Aδ- and C-fiber nociceptors after intracarotid levcromakalim, directly linking a vascular/meningeal ion-channel event to ascending trigeminal pain signaling. In short, opening of a peripheral vascular channel led to a measurable nociceptor response, and, by implication, central pain signaling.

These studies reinforce a unifying message: peripheral interventions that effectively trigger and treat migraine attacks in humans affect central physiology. Thus, observation of central changes does not, by itself, indicate a central origin of attacks. However, if the system can be ignited by peripheral mechanisms, it should also be noted that central mechanisms may initiate the CGRP cascade as well. Indeed, by using an ex vivo model with preserved anatomical continuity between the trigeminal nucleus caudalis (TCN) and trigeminal ganglion (TG), Christiansen et al. (5) showed that activating the TCN alone elicited a significant CGRP release in the TG, while the opposite (stimulation of the TG) did not cause the same effect in the TCN, supporting the bidirectional link between central-peripheral mechanisms and leaving the debate about origins of migraine attacks still unsolved.

Rethinking migraine with aura: Cause, consequence, or parallel pathway?

Whether CSD can directly be responsible for igniting the headache phase is still debated. A Viewpoint article by Moskowitz (6) argues that CSD – not aura per se – causes the headache of migraine with aura via a “buildup” of noxious mediators that eventually trigger pain from pial/dural afferents. The piece highlights a recently published stereo-EEG capture of human CSD recorded during a typical aura (7), confirming that CSD is the electrophysiological substrate of migraine aura, and contends that “silent” spread beyond eloquent cortex can explain the variable timing between aura and headache. Thus, headache beginning before or during aura may be explained by CSD propagating from non-eloquent to eloquent cortex and aura without headache could be due to an insufficient spread of CSD to trigger meningeal pain. While this hypothesis is stimulating, some clinical evidence published over the last decade challenges the view of a purely CSD-triggered model of headache in migraine with aura. Indeed, anti-CGRP mAbs, unlikely to cross the BBB, prevent migraine attacks with and without aura. If CSD (an entirely central event) were the obligatory first step, these drugs should specifically prevent the headache phase while patients would still experience aura; yet anti-CGRP antibodies appear to reduce both (8). Supporting this view, intravenously administered CGRP itself, which likewise does not readily enter the CNS, reliably provokes migraine without aura in patients with a history of migraine without aura, and occasionally provokes migraine with aura and accompanying migraine headache in patients with a history of migraine with aura. Taken together, a common peripheral mechanism may theoretically lead to initiation of migraine aura (CSD) as well as migraine pain via a separate pathway. ⁷ Thus, CSD may be sufficient for pain in some attacks, but it could be questioned whether it is universally necessary. As proposed in the Viewpoint, challenging studies requiring correlation of real-time mapping of CSD in patients with headache are needed to ultimately clarify the relationship between CSD and migraine headache.

Vestibular migraine: A challenging entity

Another crucial question that remains to be answered is whether the co-occurrence of migraine and episodic vertigo establishes a specific pathophysiology unique enough to justify a distinct disease entity. Indeed, from a clinical nosology perspective, the diagnosis of vestibular migraine (VM) remains contentious and recent reviews underscore persistent uncertainty regarding mechanisms and the limited, low-certainty evidence base for VM-specific treatments (9). On balance, VM should be considered a syndrome definition under active validation – useful for structuring research but not yet a pathophysiologically anchored entity. The study by Zhai et al. (10) is an important step in this process, offering a new hypothesis for a microglia-centric mechanism into the VM. Using a mouse model combining nitroglycerin with vestibular injury, they found that microglial TRPV2-NLRP3 signalling is implicated in central sensitization within the spinal trigeminal nucleus caudalis, showing that TRPV2 inhibition shifts microglial polarization away from a pro-inflammatory state and reduces cytokine signaling (9). These data are hypothesis-generating and could offer new potential targets; translation will require human biomarker and treatment-response studies that outperform non-specific migraine management.

Conclusions

The 2025 papers collectively advance our understanding of migraine as a complex disorder arising from interplay between central and peripheral processes. Further progress will require studies that seek to determine the temporal sequence and causal relation among pathophysiological events. Peripherally acting migraine provocation paradigms paired with high-temporal neurophysiology and longitudinal neuroimaging should establish whether cortical events (including CSD) precede, follow or are absent during peripherally triggered pain, as well as quantify the extent to which peripherically acting therapies normalize central networks. Translation will ideally also require validated human aura models for investigating the potential link between CSD and pain in patients.