Initiation of spreading depression by synaptic and network hyperactivity: Insights into trigger mechanisms of migraine aura

Abstract

Background

Cortical spreading depression (SD) is thought to underlie migraine aura but mechanisms of triggering SD in the structurally normal, well-nourished cortex of migraine patients remain unknown. Synaptic and network dysfunctions appear to underlie episodic neurological disorders, including migraine. The narrative review summarizes old and recent experimental evidence for triggering SD by synaptic/network mechanisms and discusses the relevance of the data to migraine pathogenesis. Our hypothesis is that under some conditions synaptic/network hyperactivity may reliably ignite SD, and this mechanism may underlie triggering migraine aura in patients.

Findings

High-frequency tetanic stimulation of the cortex reliably triggers SD in synaptically connected regions; SD is a reliable cortical response to acute hyperexcitability (epileptic seizures), though chronic epilepsy prevents triggering SD; in the hyperexcitable cortex, SD may be triggered by sensory stimulation; compromised glutamatergic transmission plays the critical role in triggering SD.

Conclusion

SD may be triggered by dynamic network instability produced by dysfunction of calcium-dependent glutamate release. Synaptic drive from subcortical sensory processing structures (brainstem and/or thalamocortical networks) is able to evoke depolarization of hyperexcitable cortical neurons sufficient to initiate the regenerative SD process. Studying SD initiation by synaptic/network hyperexcitability may provide insights into basic mechanisms underlying SD generation in migraine brain.

Keywords

Cortical spreading depression migraine migraine triggers cerebral cortex animal models

Introduction

Cortical spreading depression (SD), a self-propagating wave of transient neuronal depolarization, is thought to underlie neurological symptoms of migraine aura (1 –4). Spreading depression is a form of intense neuronal excitation associated with reversible breakdown of ion homeostasis and transient depression of neuronal activity. Cortical SD is believed to participate in the pathogenesis of migraine with aura (MA) while a role of SD in migraine without aura (MO) is less clear. Nevertheless, it is supposed that in MO patients, SD might occur in clinically-silent areas of the brain, not producing perceived aura symptoms (2 –4). The question of whether SD/aura triggers migraine headache remains controversial. There is experimental evidence that cortical SD may cause headache by activation and sensitization of nociceptive trigeminovascular pathways (2 –5). On the other hand, some data indicate that aura and pain may be triggered by parallel mechanisms (3 –6).

Despite the widely-accepted link between migraine aura and SD, basic mechanisms underlying spontaneous initiation of SD and the onset of a migraine aura are unknown. Direct evidence for SD occurrence in the human cortex has been obtained in conditions of severe brain pathology and compromised metabolic status (7 –9) but it remains unclear how SD can be triggered in the structurally normal, well-nourished cortex of migraine patients.

Most migraine attacks are supposed to have internal or external triggers, among which stress and excessive afferent stimulation (flickering light, noise or strong smells) are the most common (10,11). Defective mechanisms regulating cortical excitability are suggested to underlie the vulnerability of the migraine brain to various risk factors (6,10,12,13) and the dysfunction is increased before the onset of a migraine attack (12 –14). It is hypothesized that a transient destabilization of the excitatory/inhibitory (E/I) balance allows internal or external triggers to induce excessive cortical activation and creates conditions for initiating the positive feedback cycle that ignites SD (4,10,15). But so far actual trigger(s) of SD and mechanisms of its initiation in the intact cortex of migraine patients remain unknown.

In experimental animals, SD is usually induced by direct exposure of the cortex to a depolarizing stimulus (high concentrations of potassium chloride or excitatory amino acids, electrical stimulation, cortical injury and so on). Among the stimuli, local (topical or intraparenchymal) application of potassium chloride and direct current stimulation are the most widely used methods for reliable induction of cortical SD, allowing estimation of tissue susceptibility to SD and screening of antimigraine drugs (5,16). Thus, experimentally, SD is induced in the healthy cortex by its damaging stimulation, whereas migraine aura is a rather benign condition occurring in the hyperexcitable cortex in response to undamaging triggers. Synaptic dysfunction and associated dynamic network instability have been identified as major determinants of multiple neurological disorders, including migraine (17 –19). Experimental evidence obtained in models of hemiplegic migraine has shown that synaptic mechanisms, primarily associated with compromised glutamatergic transmission, play a key role in triggering SD in the migraine cortex (17 –20). Therefore, evaluating a role of synaptic/network disturbances in triggering SD is of particular interest for understanding the pathogenesis of migraine aura.

The present article provides an overview of existing experimental evidence for triggering SD by synaptic/network hyperactivity, and discusses the relevance of the data to the migraine pathogenesis. Our hypothesis is that under some conditions synaptic volley may ignite SD, and the mechanism determines triggering migraine aura. The first section focuses on old and recent findings about initiation of SD by tetanic electrical stimulation in remote regions of the healthy brain (contralateral cortex and subcortical structures). The next chapters summarize evidence for spontaneous and sensory-induced occurrence of SD in the pathologically hyperexcitable (epileptic) brain. The last sections address the role of genetic factors in enhanced susceptibility to SD and the basic mechanisms underlying initiation of SD in the intact, normally metabolized brain.

Triggering SD in distant brain sites by tetanic electrical stimulation

Being locally induced in the cortex, SD non-synaptically propagates over the gray matter at the slow velocity of 2–6 mm/min (21 –23). In the human cortex, SD propagation is restricted to the occipital lobe due to significant cortical convolutions preventing long-distance non-synaptic spread of SD (24). In the lisencephalic cortex of rodents, SD spreads across all the cortex of the stimulated hemisphere, reaching distant cortical sites with a delay of several minutes and never propagating to the contralateral cortex (21,22). However early studies, which used tetanic stimulation for SD initiation, reported that SD can be ignited via synaptic mechanisms at sites other than at the stimulating location (21,25 –27). Even in the first study of SD phenomenon (21), Leao reported that a brief (1–5 s) unilateral tetanus of the rabbit cortex triggered cortical SD in the non-stimulated hemisphere by activation of corpus callosum fibers. The synaptically triggered contralateral SD appeared initially in the cortical region homotopic to the stimulated site, and spread out from there to the rest of the cortex (21). The rapid appearance of SD in cortical regions remote from a stimulation site was reported in several old studies that used 10–50 Hz electrical stimulation of 5–10 sec duration (25,27). However, occasional incidence of SD and its close association with epileptic activity in the studies have led to rare use of the tetanus for SD induction at present.

Nevertheless, an increase in frequency of cortical stimulation remarkably shortens SD latencies in rabbits, cats and monkeys (28). SD may appear at distant cortical sites with such a short latency that only synaptic mechanisms could explain the triggering. After a brief high-frequency stimulation (100–500 Hz) of turtle cerebellum, SD sometimes appeared first at the electrode farthest away from the stimulation area, which indicates eliciting SD via activation of distant afferent pathways (29). A brief high-frequency (100 Hz) tetanus reliably induces short-latency SD in immature hippocampal rat slices (30).

Our experiments in both awake and anesthetized rats have shown that an increase in frequency of cortical stimulation from 10 Hz to 200 Hz not only remarkably shortens latencies of cortical SD, but also leads to dramatic three to five-fold reduction in SD thresholds and a twofold increase (up to 90%) in probability of SD triggering (31). As Figure 1 shows, low-frequency (10 Hz) stimulation of the cortex induces SD locally at the stimulation site, and after a delay of 60–90 sec SD appears at distant areas of the stimulated cortex, suggesting slow non-synaptic propagation of SD over the cortex. The 10 Hz tetanus never elicits cortical SD in the contralateral non-stimulated hemisphere (22,31). After stimulation at higher (50–200 Hz) frequencies, SD emerges at the remote cortical sites and in the contralateral non-stimulated cortex in 0–20 sec (Figure 1), indicating rapid synaptic ignition of SD via intracortical pathways (31). Brief unilateral 200 Hz stimulation triggers SD in the cortex of both hemispheres with 70–90% probability (31).

Figure 1.

Cortical SD induced by low- and high-frequency electrical stimulation of the parietal cortex in rats. Left: Diagram indicating location of stimulating (S) and recording (1 –4) cortical electrodes. Right: Representative examples of cortical SD induced by 10 Hz and 200 Hz stimulation of the cortex (marked by arrows) and recorded in the ipsilateral (sites 1–2) and contralateral (sites 3–4) hemispheres are shown. Calibration is 10 mV and 1 min. Note local unilateral initiation and slow non-synaptic propagation of SD induced by 10 Hz stimulation and multifocal initiation of cortical SD by 200 Hz stimulation revealing by remarkable reduction of SD latencies and triggering SD in the contralateral non-stimulated cortex (modified from (31)).

Furthermore, the high-frequency cortical stimulation also triggers short-latency (10–15 sec) SD in the thalamus and hippocampus (Figure 2) (32). An increase in stimulation frequency from 10 Hz to 100 and 500 Hz enhanced the probability of triggering SD in the thalamus from 10% to 50% and 100%, respectively (32). The cortical stimulation also induced short-latency SD in the hippocampus but the probability of the triggering did not exceed 50%, irrespective of stimulation frequency (32). The appearance of the short-latency SD in the thalamus and hippocampus after high-frequency cortical stimulation indicates initiation of SD via synaptic activation of descending cortical pathways. The difference in probability of triggering thalamic and hippocampal SDs by stimulation of the parietal cortex may reflect either a region-specific susceptibility to SD or specificity of cortical inputs to the subcortical structures. Given the critical role of glutamate and its receptors in SD generation (29,33), the easy triggering of SD in the thalamus by cortical stimulation may be explained by the glutamatergic nature of the cortico-thalamic projections.

Figure 2.

Initiation of SD in the cortex (Cx), thalamus (Th) and hippocampus (Hp) by electrical stimulation of the parietal cortex at frequencies of 50 and 500 Hz in rats. Moments and intensities of cortical stimulation (a train of 10 pulses) are marked at every fragment. Calibration is 10 mV, 1 min. Note that hippocampal SD is triggered only by intense cortical stimulation, irrespective of its frequency. Thalamic SD is easily induced by high frequency cortical stimulation (even at its minimal intensity), whereas even the strongest 50 Hz stimulation fails to elicit SD in the thalamus (modified from (32)).

In the intact brain of healthy mice and rats, cortical SD does not propagate to the hippocampus and thalamus non-synaptically (22,32), although non-synaptic penetration of cortical SD to the hippocampus via contiguous gray matter has been reported in combined neocortex-hippocampus brain slices (34). Transgenic mice R192Q and S218L with FHM-1 mutation exhibit facilitated non-synaptic propagation of cortical SD to the thalamus, hippocampus, striatum that is considered as a manifestation of genetically enhanced SD susceptibility (35,36). In S218L mice with severe FHM-1 symptoms, electrical stimulation of the occipital cortex sometimes elicited almost simultaneous appearance of SD in the cortex and subcortical structures (36). It can be speculated that in some cases SD may be triggered de novo via corticofugal axonal projections in the mutants.

Apart from the stimulation frequency, its duration is an important factor for triggering SD. If brief (1 s at 10 Hz and 0.1–0.2 s at 50–200 Hz) stimulation of the cortex reliably triggers SD, longer stimulation at the same frequency (20–30 s at 10 Hz and 1–5 s at 50–200 Hz) elicits SD of decreased amplitude (31) and can even prevent propagation of chemically or mechanically induced SD to the stimulated area (37,38). Mechanisms of potassium ion clearance play the critical role in the blockade of SD by prolonged stimulation (37,39). Effective summation of residual depolarizations during a high-frequency train of stimuli elicits a rapid rise in extracellular potassium concentrations beyond the SD threshold of 10–12 mM and triggers SD positive feedback. The same pulses applied at a lower rate and for longer periods produce gradual recruitment of neuronal populations with activation of metabolic pumps and increased potassium buffering, which raises the SD threshold and prevents SD initiation (38).

Thus, high-frequency stimulation of the cerebral cortex promotes triggering SD in synaptically connected regions, including contralateral cortex and subcortical structures, via intracortical and corticofugal projections (21,31,32). Given clinical data about increased frequency of background electrical activity in migraine patients (40) and enhanced somatosensory-induced high-frequency oscillations during and between migraine attacks (41), the reviewed experimental evidence may be relevant to pathophysiological mechanisms of triggering migraine aura. SD induced in subcortical structures by over-activation of the hyperresponsive cerebral cortex has been hypothesized to mediate several neurological deficits in migraine with aura (2).

“Spontaneous” triggering SD in the hyperexcitable epileptic brain

Migraine is a disorder of brain excitability and dysfunctional regulation of the cortical E/I balance (3,4,13,15). Treatments increasing cortical excitability (e.g. electroconvulsive therapy, transcranial magnetic stimulation) can produce headache in patients (42,43) and trigger SD in experimental animals (44 –46). The SD phenomenon was discovered as a side product of experimental epilepsy research (21) and SD frequently occurs during epileptic hyperexcitation in both humans (8) and animals (23,45,47,48).

Transcranial electroconvulsive stimulation can induce SD in the cortex, hippocampus, thalamus and striatum of experimental animals (44,45). Low-intensity electroshock can trigger cortical SD without behavioral convulsions (46). A single systemic administration of PTZ reliably induces bilateral SD waves in the neocortex (in 83% of rats) and thalamus (in 95% of rats) when the first epileptic spikes appear in the EEG (47). Cortical SD can be elicited by a short-lasting and modest increase in neuronal excitability such as a brief episode of minimal brainstem seizures (48), a single epileptic spike (49) and even slight disinhibition of the cortex without its epileptic activation (50). Thus, the threshold for SD initiation in the hyperexcitable cortex may be lower than that required for seizure induction (46 –50). Accordingly, more abundant occurrence of SD compared to seizures has been reported in the acutely injured human cortex (8) and in the hyperexcitable cortex of experimental animals (47 –50). Moreover, SD may be produced by projected activation of brain sites remote from a focus of pathologic hyperexcitation. In rats with genetic audiogenic epilepsy, repeated mild brainstem epileptic activation reliably triggers a unilateral cortical SD in the absence of cortical seizures (48,51). And vice versa, acutely induced cortical seizures trigger SD in the brainstem of mice carrying mutations in potassium and sodium ion channels (52).

In contrast to the easy triggering of SD by acute epileptic hyperexcitation, chronic epilepsy increases the resistance of the cortex to SD (47,50,53,54). In pilocarpine-treated chronically epileptic rats, the efficacy of KCl injection in eliciting cortical SD is halved compared to non-epileptic animals (53). Stimulation of the cortex, which easily induces SD in nonepileptic rats, cannot elicit SD in the cortex of patients with intractable epilepsy (55). Increased resistance to SD has been shown in slices of chronically epileptic rats and humans (50,54). Our experiments have shown that chronic PTZ treatment leads to both intensification of epileptiform activity and a twofold decrease in cortical SD occurrence (from 83% to 37%) (47).

SD and seizure represent two levels of the homeostatic regulation of the ionic environment of the brain (22,23,56). Seizures are associated with steady state regulation of extracellular potassium concentration at a ceiling level of about 10 mM, while SD reflects oscillation between two extreme states produced by positive and negative feedback mechanisms (22,23). Mechanisms of triggering SD during epileptic over-excitation are supposed to involve local accumulation of potassium ions above a critical level (22,57). According to the traditional viewpoint, triggering SD requires intense epileptic activation of the cortex to exceed the “ceiling” level and to ignite self-regenerating depolarization process (22,23). However, the reviewed experimental evidence shows that abrupt hyperexcitation, even brief and low-intensity, can increase extracellular potassium concentration to the threshold level and trigger SD. On the contrary, chronic (prolonged) epileptic activation increases metabolic transport and clearance of potassium ions, counteracting potassium accumulation in the extracellular space and preventing the autoregenerative process of SD generation (22,47,53). However, in bicuculine-treated slices, SD can start at the near-to-normal extracellular concentrations of potassium ions (50), suggesting involvement of additional mechanisms in triggering SD. Resistance to SD generation in chronic epilepsy has been shown to reflect increased tolerance to high potassium concentrations rather than more effective potassium buffering (54).

Both migraine and epilepsy are disorders associated with deficient regulation of E/I balance and enhanced neuronal excitability (56). Cortical hyperexcitability is a common finding in both MA and MO patients (13,15) and disruption of E/I balance is supposed to underlie some migraine symptoms including aura (15,20). Given the co-morbidity of epilepsy and migraine (56), as well as the effectiveness of many antiepileptic drugs in migraine treatment (6,16), migraine-associated hyperexcitability might share basic mechanisms with epileptic excitation. But episodic disruption of E/I balance, which triggers a seizure discharge in the epileptic cortex, should elicit SD without epileptic activation in the cortex of migraine patients. The reviewed experimental data show that SD may be induced by a modest increase in the cortical excitability subthreshold for seizure initiation, and this mechanism may mimic triggering migraine aura by aberrant neuronal firing (15).

Triggering SD by sensory stimulation

Abnormal processing of sensory stimuli represents a general marker of neuronal dysfunction in migraine patients (58,59). The migraine brain is more reactive to various environmental stimuli compared to healthy people, and sensory stimulation can even trigger an attack of migraine with aura (11,12). Migraine patients show lowered thresholds of auditory and visual discomfort both between and during attacks, suggesting enhanced sensitivity and exaggerated responses (hyperresponsivity) to sensory activation (12). Impaired habituation to repetitive sensory stimulation contributes to dysfunctional sensory processing in migraine patients (12,58).

There are only a few experimental studies, mainly in awake animals, reporting the possibility of triggering SD by sensory stimulation. Van Harreveld and Stamm (60) were the first who demonstrated that sensory stimuli (single or repeated light flashes/acoustic stimuli) may induce SD in the hyperexcitable cortex of awake rabbits. Systemic administration of low subconvulsive doses of PTZ was used to increase cortical excitability in the animals, but no epileptiform activity was recorded before or during sensory-induced SD. The SD waves first appeared (with a latent period of a few seconds) in respective sensory regions, in the visual cortex after optic stimulation and in the auditory cortex after acoustic stimulation.

Our experiments in genetically-susceptible animals have shown that a unilateral cortical SD may be produced by overactivation of the auditory network and its ascending synaptic inputs to the cortex (48,61,62). In awake Wistar rats with innate hypersensitivity to acoustic stimulation (audiogenic epilepsy), repetitive acoustic stimulation reliably triggers SD in the hyperexcitable cerebral cortex (48,61). In these sound-susceptible rodents, short-lasting acoustic stimulation elicits brief (5–10 sec) mild epileptic activation of the brainstem, primarily of the inferior colliculus, the key site of the auditory processing network (61 –66). The minimal brainstem seizures usually have asymmetric pattern (66) and their repetition reliably triggers SD in the cortex of one hemisphere (48,61,62). Lateralization of the unilateral cortical SD coincides with that of brainstem epileptic excitation, suggesting involvement of intra-hemispheric ascending pathways in triggering the unilateral cortical SD (51,62,66). Thus, sensory (acoustic) stimulation may produce asymmetric network hyperactivity, the only cortical manifestation of which is a unilateral SD. Clinical data about increased intrinsic connectivity between the brainstem and ipsilateral cortex during spontaneous attacks of migraine with aura (67) and tight correlation between lateralizations of brainstem activation and headache (68) strengthen the relevance of the experimental data to understanding the basic mechanisms of migraine aura.

At the early kindling stages, when excitability of the cortex is only slightly increased and episodic brainstem excitation triggers cortical SD without seizure activation of the cortex, SD is always unilateral and appears in the fronto-parietal cortex after a significant time delay (1–2 min), indicating local triggering SD in a restricted cortical area of one hemisphere and its subsequent non-synaptic propagation (48,61). At the late kindling stages, acoustically-induced brainstem excitation triggers strong generalized seizures and multifocal bilateral short-latency SD in the highly hyperexcitable cortex (48,66,69).

High susceptibility to audiogenic brainstem seizures, which are similar to those triggering unilateral cortical SD in Wistar rats, has been recently described in mice bearing mutation of the proline-rich transmembrane domain protein 2 (PRRT2) gene (70) found in patients with different types of migraine (71). It is hypothesized that PRRT2 protein, which is involved in synaptic function and highly expressed in sensory processing regions, participates in the pathogenesis of sensory-triggered paroxysmal events, including migraine attacks (18,19).

Somatosensory stimulation (light tactile stimulus) is able to elicit cortical SD-like waves (peri-infarct depolarizations) in metastable hot zones around penumbra during supply-demand mismatch transients in the mouse model of stroke (72). In healthy mice, prolonged (for several minutes) visual and somatosensory stimulation facilitates potassium-induced SD (73).

Triggering cortical SD by sensory stimulation certainly involves activation of specific sensory-processing networks via synaptic mechanisms. The above experimental evidence indicates that the synaptic drive from subcortical sensory processing structures (brainstem and/or thalamocortical networks) is able to evoke depolarization of hyperexcitable (or injured) cortical neurons sufficient for initiation of the regenerative SD process. Increased excitability of the cortex appears to be the prerequisite for triggering SD by sensory stimuli. The experimental data are in line with the clinical evidence of strong subcortical involvement in triggering migraine attacks (3,4,6). Interictal hypersensitivity to sensory stimuli and abnormal processing of sensory information are key manifestations of migraine (12,58). The aberrant sensory sensitivity is maximal 12–24 h before the migraine attack, i.e. during the premonitory phase (58). Despite normalization just before or during the migraine attack, the abnormality may predispose the migraine cortex to triggering SD/aura by sensory-induced hyperactivity. Further studies are needed to clarify the pathways and mechanisms of triggering SD by sensory stimulation.

Genetic factors predisposing the migraine brain to initiation of SD

Migraine is under a strong genetic influence (74). There is growing evidence that genetically-determined dysfunctions of ion channels, synapses and networks underlie episodic neurological disorders, including migraine (18,19).

Genetic models of familial hemiplegic migraine (FHM), a rare autosomal dominant subtype of migraine with aura, provide important information about mechanisms involved in triggering migraine attack and cortical SD. Mice carrying human FHM mutations show enhanced susceptibility to SD along with defective regulation of E/I balance, glutamatergic transmission and synaptic plasticity (75,76).

Mutations of voltage-gated neuronal Cav2.1 (P/Q-type) channels associated with FHM type 1 determine a gain-of-function increase in calcium conductance and enhanced excitatory, but not inhibitory, transmission at cortical synapses (77). Apart from abnormal synaptic calcium homeostasis, structural changes in synaptic morphology (17) and enhanced basal cortical excitability (36) have been described in the FHM-1 mutants. Knock-in mice carrying the human FHM-1 mutations exhibit enhanced susceptibility to SD manifested as facilitation of triggering cortical SD by electrical and chemical stimulation (75,78,79) and its propagation to subcortical structures (35,36). The mutant mice with loss-of-function P/Q Ca channels display a tenfold higher SD threshold (80). The evidence suggests that synaptic dysfunction, namely increased calcium influx into presynaptic terminals via mutated P/Q type calcium channels and enhanced synaptic release of glutamate, promote the initiation of SD.

FHM type 2 is caused by loss-of-function mutations in the astrocytic isoform of the Na-K-ATPase, and mice carrying the mutation show reduced density of GLT-1a glutamatergic transporters in cortical perisynaptic astrocyte processes and reduced rates of glutamatergic and potassium ion clearance by cortical astrocytes during neuronal activity (76,81). Mice with the human mutation exhibit facilitated induction of SD by electrical stimulation of the cortex (81). A causative relationship between the decreased SD threshold and the reduced rate of glutamate clearance at cortical synapses with minimal contribution of the reduced rate of potassium clearance by astrocytes has been shown in FHM2 knock-in mice (76).

Thus, the data obtained in FHM mutant models suggest that excessive glutamatergic synaptic transmission and defective glutamate clearance, increasing the glutamate concentration in the synaptic cleft, may underlie triggering SD in the migraine cortex (76,79).

Mutations in the PRRT2 gene are implicated in the pathogenesis of various paroxysmal disorders, including migraine (19,71). A protein encoded by PRRT2 localizes at synaptic contacts, participating in presynaptic release of neurotransmitters, multisensory integration and sensorimotor coordination (19,82). All mice bearing the PRRT2 mutation express mild brainstem seizures in response to acoustic stimulation (70). Since these sound-induced seizures reliably trigger cortical SD (48,62), it can be speculated that the PRRT2 mutant mice may exhibit sensory-evoked cortical SD.

Basic mechanisms of triggering SD in the normally metabolized brain

The critical condition for triggering SD is strong sustained depolarization of neurons. In the normally metabolized cortex, such depolarization may result either from a failure of homeostatic mechanisms regulating extracellular potassium/glutamate levels or from abnormal sensitivity to the extracellular concentration changes (83). The initial stimulus-induced depolarization elicits explosive opening of cation conductances and activation of net persistent inward current in apical dendrites of pyramidal neurons (23,83,84). The self-sustaining inward current initiates a positive-feedback cycle that triggers accelerating, regenerative, all-or-none type depolarization typical for SD.

There is a general consensus that an increase in extracellular potassium concentration above a critical value (12–15 mM) is a key initiating event for activation of the positive feedback mechanism of SD (22,23,25,85). The time required to attain the SD threshold is determined by a balance between stimulus-induced accumulation and dispersion of potassium ions by different buffering mechanisms (22,85,86). High KCl and cathodal (dc) stimulation acts by raising the local extracellular potassium concentration below the stimulation site and the elevation may be rather moderate and occur in a relatively small volume (85,86).

On the other hand, there is strong experimental evidence favoring the critical role of glutamate in the SD initiation process. Application of glutamate or NMDA agonists evokes SD, whereas NMDA antagonists inhibit SD initiation (29,33). Significant release of glutamate from neurons and glia occurs with SD, and activation of glutamate receptors precedes massive potassium efflux during SD (87,88). The glutamate overflow into interstitial space could be mediated either by its excessive release in a Ca-dependent vesicular and Ca-independent non-vesicular manner via glutamate transporters, or by activation of presynaptic NMDA receptors (87 –89). It is supposed that glutamate derived from synaptic Ca-dependent vesicular release predominates during SD initiation in the well-nourished brain, whereas Ca-independent non-vesicular glutamate release contributes to triggering SD by anoxia (4). Data from FHM mutants indicate that excessive glutamate release and activation of NMDA receptors are the key steps in triggering SD in migraine (17,76,78,90). In FHM-1 mutants, the decreased threshold for SD induction results from enhanced synaptic release of glutamate due to P/Q channel gain-of-function mutation selectively affecting glutamatergic neurons, but not GABA-ergic inhibitory interneurons (77).

Van Harreveld suggested that an increase in extracellular glutamate, irrespective of extracellular potassium increase, is able to trigger SD and two types of SD, one mediated by potassium ions and the other dependent on glutamate, may occur (91). Under conditions favoring both glutamate- and potassium-based mechanisms, hybrid SDs mediated by both may occur (23,83,91). It has been shown that various methods of SD induction may differ in underlying mechanisms (4,92). For example, sodium and P/Q calcium channels implicated in the pathogenesis of FHM (17,74,90) are the key for triggering SD by mechanical and electrical stimulation, whereas potassium-induced SD critically depends on calcium and potassium channels, but not sodium ones (92,93). Which of the mechanisms determines triggering cortical SD (aura) in the migraine brain is still unclear.

Most migraine triggers, that is, factors inducing migraine headache in susceptible individuals (such as food or sleep deprivation), evoke a migraine attack after a delay of a few hours (94), and the triggers may represent premonitory symptoms of an already ongoing attack (6). Migraine triggers usually induce rather slow and mild homeostatic changes, while the above experimental evidence shows that initiation of SD requires rapid and profound changes in neuronal function. Therefore, it is likely that most migraine triggers reduce the threshold for generation of SD/aura by synaptic/network mechanisms rather than directly trigger SD/aura. For example, calcitonin-gene related peptide (CGRP), a well-known player in the migraine pathogenesis and one of possible migraine triggers (6,95), can induce delayed migraine-like headache and even aura in MA patients (96) but fails to induce cortical SD in animals (97). It is likely that the hypersensitivity of migraineurs to CGRP is not sufficient to trigger SD/aura directly and requires additional (neuronal) mechanisms.

Conclusion

A growing body of data supports the idea that migraine is a complex neurological disorder associated with significant homeostatic changes, disregulation of neuronal E/I balance and network instability (3,6,13,18,59,90). A migraine attack is a multistep event that starts with premonitory symptoms long before development of aura and pain. Disruption of body and brain homeostasis at the onset of a migraine attack generates prodromal symptoms and may progressively increase sensory processing dysfunction, cortical excitability, and vulnerability to various risk factors (3,6,13,14). It is possible that at a certain level of the cortical dysfunction, internal or external triggers start to induce hyperactivity of cortical circuits sufficient to trigger SD/aura.

To date, there is no understanding of how SD is ignited in the uninjured brain of migraineurs, and there is no experimental model relevant to the spontaneous triggering SD. Mechanisms underlying initiation of SD in the structurally intact and normally metabolizing brain of migraine patients should certainly involve synaptic/network mechanisms. The experimental evidence reviewed in the present article show that in the uninjured hyperexcitable cortex, SD may occur spontaneously, that is, as a result of intrinsic aberrant network activity, or in response to sensory activation. Direct experimental evidence for the ability of pure neuronal mechanisms to trigger SD in the intact brain has been recently obtained in a study showing initiation of SD by optogenetic activation of cortical neurons (98). Genetic factors determining excessive glutamatergic transmission predispose the migraine brain to generation of SD during synaptic hyperactivity and/or network instability.

It is increasingly recognized that mechanisms of SD ignition differ from those of SD propagation (23,92). Although unfolding and propagation of SD are clearly mediated by non-synaptic mechanisms and do not require synaptic signaling (22,23), synaptic processes may participate in the initial activation of the positive feedback mechanisms at the earliest stage of SD generation. Once the positive feedback is activated, synaptic mechanisms may no longer be necessary and non-synaptic interactions drive subsequent progression and propagation of SD (87).

Recent experiments with potassium-induced SD have demonstrated non-uniform susceptibility of the rat cortex to SD, with the lowest SD threshold in the somatosensory cortex (73) and the most reliable SD incidence in the visual cortex (99). The data on electrically- and sensory-induced SD discussed in the present review (21,48) indicate that the SD may be triggered by synaptic volley in a restricted cortical area, which is especially vulnerable to SD, and from this site SD may non-synaptically propagate to surrounding neuronal tissue, shaping the well-known slowly propagating phenotype of migraine aura. At a higher level of hyperexcitability, simultaneous synaptic triggering SD in several regions of the occipital cortex may occur, and results in multifocal patterns of migraine aura described in patients (94,100).

The reviewed data show that SD may be a reliable response to aberrant activation of long-range synaptic pathways, including upstream brainstem projections and thalamocortical networks, which are known to be involved in the generation of migraine attacks. Elevated excitability of the cortex is prerequisite for synaptic triggering cortical SD by the subcortical drive. Initiation of SD in subcortical structures has been speculated to produce premonitory symptoms and trigger headache in migraine without aura (2).

Studying of SD ignition by synaptic/network hyperexcitability may provide an important insight into basic mechanisms underlying generation of SD/aura in the migraine brain.

Key findings

Basic mechanisms that drive spontaneous initiation of cortical spreading depression (SD), the phenomenon underlying migraine aura, are unknown. The reviewed experimental evidence indicate that SD may be triggered by synaptic/network hyperactivity under conditions that are expected in the migraine brain.

Experimental data demonstrate that brief high-frequency stimulation of the cerebral cortex promotes triggering SD in synaptically connected regions, including subcortical structures.

SD is a reliable cortical response to a modest acute increase in neuronal excitability.

In the hyperexcitable cortex, SD may be triggered by sensory stimulation, suggesting that synaptic drive from subcortical sensory processing structures (brainstem and/or thalamocortical networks) can depolarize cortical neurons to levels sufficient for SD ignition.

Footnotes

Declaration of conflicting interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

References

Leao

AAP

Morrison

. Propagation of spreading cortical depression. J Neurophysiol 1945; 8: 33–45.

Ayata

. Cortical spreading depression triggers migraine attack: Pro. Headache 2010; 50: 725–730.

Charles

Baca

. Cortical spreading depression and migraine. Nat Rev Neurol 2013; 9: 637–644.

Pietrobon

Moskowitz

. Chaos and commotion in the wake of cortical spreading depression and spreading depolarizations. Nat Rev Neurosci 2014; 15: 379–393.

Zhang

Levy

Noseda

et al.

Activation of meningeal nociceptors by cortical spreading depression: Implications for migraine with aura. J Neurosci 2010; 30: 8807–8814.

Goadsby

Holland

Martins-Oliveira

et al.

Pathophysiology of migraine: A disorder of sensory processing. Physiol Rev 2017; 97: 553–622.

Dreier

Woitzik

Fabricius

et al.

Delayed ischaemic neurological deficits after subarachnoid haemorrhage are associated with clusters of spreading depression. Brain 2006; 129: 3224–3237.

Fabricius

Fuhr

Willumsen

et al.

Association of seizures with cortical spreading depression and peri-infarct depolarization in the acutely injured human brain. Clin Neurophysiol 2008; 119: 1973–1984.

Lauritzen

Strong

. “Spreading depression of Leão” and its emerging relevance to acute brain injury in humans. J Cereb Blood Flow Metab 2017; 37: 1553–1570.

10.

Lambert

Zagami

. The mode of action of migraine triggers: A hypothesis. Headache 2009; 49: 253–275.

11.

Hauge

Kirchmann

Olesen

. Characterization of consistent triggers of migraine with aura. Cephalalgia 2011; 31: 416–438.

12.

Coppola

Pierelli

Schoenen

. Is the cerebral cortex hyperexcitable or hyperresponsive in migraine? Cephalalgia 2007; 27: 1429–1439.

13.

Aurora

Wilkinson

. The brain is hyperexcitable in migraine. Cephalalgia 2007; 27: 1442–1453.

14.

Charles

. The evolution of a migraine attack – a review of recent evidence. Headache 2013; 53: 413–419.

15.

Welch

KMA

D’Andrea

Tepley

et al.

The concept of migraine as a state of central neuronal hyper excitability. Headache 1990; 8: 817–828.

16.

Ayata

Jin

Kudo

et al.

Suppression of cortical spreading depression in migraine prophylaxis. Ann Neurol 2006; 59: 652–661.

17.

Eikermann-Haerter

Arbel-Ornath

Yalcin

et al.

Abnormal synaptic Ca²⁺ homeostasis and morphology in cortical neurons of familial hemiplegic migraine type 1 mutant mice. Ann Neurol 2015; 78: 193–210.

18.

Russell

Ptáček

. Episodic neurologic disorders: Syndromes, genes, and mechanisms. Ann Rev Neurosci 2013, 8; 36: 25–50.

19.

Valtorta

Benfenati

Zara

et al.

PRRT2: From paroxysmal disorders to regulation of synaptic function. Trends Neurosci 2016; 39: 668–679.

20.

Gasparini

Smith

Griffiths

. Genetic insights into migraine and glutamate: A protagonist driving the headache. J Neurol Sci 2016; 367: 258–268.

21.

Leao

AAP

. Spreading depression of activity in cerebral cortex. J Neurophysiol 1944; 7: 359–390.

22.

Bures

Buresova

Krivanec

. The mechanisms and applications of Leao’s spreading depression of electroencephalographic activity, Prague, Czech Republic: Academia, 1974.

23.

Somjen

. Mechanisms of spreading depression and hypoxic spreading depression-like depolarization. Physiol Rev 2001; 81: 1065–1096.

24.

Hadjikhani

del Rio

et al.

Mechanisms of migraine aura revealed by functional MRI in human visual cortex. PNAS 2001; 98: 4687–4692.

25.

Grafstein

. Mechanism of spreading cortical depression. J Neurophysiol 1956; 19: 154–171.

26.

Rosenblueth

García Ramos

. Some phenomena usually associated with spreading depression. Acta Physiol Lat Am 1966; 16: 141–179.

27.

Och

Hunt

Booker

. Spreading depression using chronically implanted electrodes. Am J Physiol 1961; 200: 1211–1214.

28.

van Harreveld

Stamm

Christensen

. Spreading depression in rabbit, cat and monkey. Am J Physiol 1956; 184: 312–320.

29.

Lauritzen

Rice

Okada

et al.

Quisqualate, kainate and NMDA can initiate spreading depression in the turtle cerebellum. Brain Res 1988; 475: 317–327.

30.

Javedan

Ellsworth

et al.

Gamma oscillation underlies hyperthermia-induced epileptiform-like spikes in immature rat hippocampal slices. BMC Neurosci 2001; 2: 18–18.

31.

Koroleva

. Synaptically evoked spreading cortical depression in short-term high-frequency electrical stimulation of the rat brain. Neirofiziologiia 1989; 21: 789–796.

32.

Koroleva

Vinogradova

. Spreading depression in the thalamus, hippocampus and caudate nucleus of the rat during electrical stimulation of the parietal area of the cortex. Neirofiziologiia 1990; 22: 36–44.

33.

Marrannes

Willems

De Prins

et al.

Evidence for a role of the N-methyl-D-aspartate (NMDA) receptor in cortical spreading depression in the rat. Brain Res 1988; 457: 226–240.

34.

Martens-Mantai T, Speckmann EJ and Gorji A. Propagation of cortical spreading depression into the hippocampus: The role of the enthorhinal cortex. Synapse 2014; 68: 574–584.

35.

Eikermann-Haerter

Yuzawa

Qin

et al.

Enhanced subcortical spreading depression in familial hemiplegic migraine type 1 mutant mice. J Neurosci 2011; 31: 5755–5763.

36.

Cain

Bohnet

LeDue

et al.

In vivo imaging reveals that pregabalin inhibits cortical spreading depression and propagation to subcortical brain structures. Proc Natl Acad Sci USA 2017; 114: 2401–2406.

37.

Reddy

Bures

. Cortical (K+)e and the stimulation induced blockade of spreading depression in the rat cerebral cortex. Neurosci Lett 1980; 17: 243–247.

38.

Koroleva

Oitzl

Bures

. Threshold of synaptically elicited cortical spreading depression: drug-induced changes in rats. Electroenceph Clin Neurophysiol 1985; 60: 55–64.

39.

Heinemann

Lux

. Ceiling of stimulus induced rises in extracellular potassium concentration in the cerebral cortex of cat. Brain Res 1977; 120: 231–249.

40.

Bramanti

Grugno

Vitetta

et al.

Migraine with and without aura: Electrophysiological and functional neuroimaging evidence. Funct Neurol 2005; 20: 29–32.

41.

Liao

Lioa

Fuh

et al.

Subcortical hyperexcitability in migraineurs: A high-frequency oscillation study. Can J Neurol Sci 2011; 38: 309–316.

42.

Aurora

Ahmad

Welch

et al.

Transcranial magnetic stimulation confirms hyperexcitability of occipital cortex in migraine. Neurology 1998; 50: 1111–1114.

43.

Antal

Arlt

Nitsche

et al.

Higher variability of phosphene thresholds in migraineurs than in controls: A consecutive transcranial magnetic stimulation study. Cephalalgia 2006; 26: 865–870.

44.

Bures

Buresová

. Cerebral [K+]e increase as an index of the differential susceptibility of brain structures to terminal anoxia and electroconvulsive shock. J Neurobiol 1981; 12: 211–220.

45.

de Azeredo

Perret

. Cortical slow potential changes during convulsions induced by maximal electroshock or penicillin focus. Metab Brain Dis 1992; 7: 101–113.

46.

Rebert

Pryor

Schaeffer

. Slow cortical potential consequences of elctroconvulsive shock in rats. Physiol Behav 1974; 12: 131–134.

47.

Koroleva

Vinogradova

Bures

. Reduced incidence of cortical spreading depression in the course of pentylenetetrazol kindling in rats. Brain Res 1993; 608: 107–114.

48.

Vinogradova

. Comparative potency of sensory-induced brainstem activation to trigger spreading depression and seizures in the cortex of awake rats: Implications for the pathophysiology of migraine aura. Cephalalgia 2015; 35: 979–986.

49.

Koroleva

Bures

. Cortical penicillin focus as a generation of spike-triggered waves of spreading depression in rats. Exp Brain Res 1983; 51: 291–297.

50.

Kohling

Koch

Hagemann

et al.

Differential sensitivity to induction of spreading depression by partial disinhibition in chronically epileptic human and rat as compared to native rat neocortical tissue. Brain Res 2003; 975: 129–134.

51.

Vinogradova

Shatskova

. Lateral asymmetry of early seizure manifestations in experimental generalized epilepsy. Neuroscience 2012; 213: 133–143.

52.

Aiba I and Noebels JL. Spreading depolarization in the brainstem mediates sudden cardiorespiratory arrest in mouse SUDEP models. Sci Transl Med 2015; 7: 282ra46.

53.

Guedes

Cavalheiro

. Blockade of spreading depression in chronic epileptic rats: Reversion by diazepam. Epilepsy Res 1997; 27: 33–40.

54.

Maslarova

Alam

Reiffurth

et al.

Chronically epileptic human and rat neocortex display a similar resistance against spreading depolarization in vitro. Stroke 2011; 42: 2917–2922.

55.

McLachlan

Girvin

. Spreading depression of Leao in rodent and human cortex. Brain Res 1994; 666: 133–136.

56.

Rogawski

Migraine and epilepsy – shared mechanisms within the family of episodic disorders. In: Noebels

Avoli

Rogawski

et al. (eds). Jasper’s basic mechanisms of the epilepsies, 4th edition. Bethesda, MD: National Center for Biotechnology Information (US), 2012.

57.

Haglund

Schwartzkroin

. Role of Na-K pump potassium regulation and IPSPs in seizures and spreading depression in immature rabbit hippocampal slices. J Neurophysiol 1990; 63: 225–239.

58.

Shoenen

Wang

Albert

et al.

Potentiation instead of habituation characterizes visual evoked potentials in migraine patients between attacks. Eur J Neurol 1995; 2: 115–122.

59.

Charles

. Migraine: A brain state. Curr Opin Neurol 2013; 26: 235–239.

60.

van Harreveld

Stamm

. Cortical responses to metrazole and sensory stimulation in the rabbit. EEG Clin Neurophysiol 1955; 7: 363–370.

61.

Vinogradova

Kuznetsova

Coenen

AML

. Unilateral cortical spreading depression induced by sound in rats. Brain Res 2009; 1286: 201–207.

62.

Vinogradova

. Audiogenic kindling and secondary subcortico-cortical epileptogenesis: Behavioral correlates and electrographic features. Epilepsy Behav 2017; 71 (Pt B): 142–153.

63.

Marescaux

Vergnes

Kiesmann

et al.

Kindling of audiogenic seizures in Wistar rats: An EEG study. Exp Neurol 1987; 97: 160–168.

64.

Faingold

. Neuronal networks in the genetically epilepsy-prone rats. Adv Neurol 1999; 79: 311–321.

65.

Jobe

Browning

Mammalian models of genetic epilepsy characterized by sensory-evoked seizures and generalized seizure susceptibility. In: Pitkanen

Schwartzkroin

Moshe

(eds). Models of seizures and epilepsy, San Diego: Elsevier Academic Press, 2006, pp. 261–271.

66.

Vinogradova

Grinenko

. Ictal electrographic pattern of focal subcortical seizures induced by sound in rats. Brain Res 2016; 1635: 161–168.

67.

Hougaard

Amin

Larsson

et al.

Increased intrinsic brain connectivity between pons and somatosensory cortex during attacks of migraine with aura. Hum Brain Mapp 2017; 38: 2635–2642.

68.

Afridi

Matharu

Lee

et al.

A PET study exploring the laterality of brainstem activation in migraine using glyceryl trinitrate. Brain 2005; 128: 932–939.

69.

Vinogradova

. Audiogenic kindling in Wistar and WAG/Rij rats: Kindling-prone and kindling-resistant subpopulations. Epilepsia 2008; 49: 1665–1674.

70.

Michetti

Castroflorio

Marchionni

et al.

The PRRT2 knockout mouse recapitulates the neurological diseases associated with PRRT2 mutations. Neurobiol Dis 2017; 99: 66–83.

71.

Riant

Rose

Barbance

et al.

PRRT2 mutations cause hemiplegic migraine. Neurology 2012; 79: 2122–2124.

72.

Von Bornstädt

Houben

Seidel

et al.

Supply-demand mismatch transients in susceptible peri-infarct hot zones explain the origin of spreading injury depolarizations. Neuron 2015; 85: 1117–1131.

73.

Bogdanov

Middleton

Theriot

et al.

Susceptibility of primary sensory cortex to spreading depolarizations. J Neurosci 2016; 36: 4733–4743.

74.

Sutherland

Griffiths

. Genetics of migraine: Insights into the molecular basis of migraine disorders. Headache 2017; 57: 537–569.

75.

Ophoff

Terwindt

Vergouwe

et al.

Familial hemiplegic migraine and episodic ataxia type-2 are caused by mutations in the Ca2+ channel gene CACNL1A4. Cell 1996; 87: 543–552.

76.

Capuani

Melone

Tottene

et al.

Defective glutamate and K+ clearance by cortical astrocytes in familial hemiplegic migraine type 2. EMBO Mol Med 2016; 8: 967–986.

77.

Vecchia

Tottene

van den Maagdenberg

AMJM

et al.

Mechanism underlying unaltered cortical inhibitory synaptic transmission in contrast with enhanced excitatory transmission in Ca_V2.1 knockin migraine mice. Neurobiol Dis 2014; 69: 225–234.

78.

Van den Maagdenberg

AMJM

Pietrobon

Pizzorusso

et al.

A Cacna1a knock-in migraine mouse model with increased susceptibility to cortical spreading depression. Neuron 2004; 41: 701–710.

79.

Tottene

Conti

Fabbro

et al.

Enhanced excitatory transmission at cortical synapses as the basis for facilitated spreading depression in Ca(v)2.1 knockin migraine mice. Neuron 2009; 61: 762–773.

80.

Ayata

Shimizu-Sasamata

et al.

Impaired neurotransmitter release and elevated threshold for cortical spreading depression in mice with mutations in the alpha1a subunit of P/Q type calcium channels. Neuroscience 2000; 95: 639–645.

81.

Leo

Gherardini

Barone

et al.

Increased susceptibility to cortical spreading depression in the mouse model of familial hemiplegic migraine type 2. PLoS Genetics 2011; 7: e1002129–e1002129.

82.

Niu

Zhu

et al.

PRRT2 mutant leads to dysfunction of glutamate signaling. Int J Mol Sci 2015; 16: 9134–9151.

83.

Kager

Wadman

Somjen

. Conditions for the triggering of spreading depression studied with computer simulations. J Neurophysiol 2002; 88: 2700–2712.

84.

Cannals

Makarova

Lopez-Aguado

et al.

Longitudial depolarization gradients along the somatodendritic axis of CA1 pyramidal cells: A novel feature of spreading depression. J.Neurophysiol 2005; 94: 943–951.

85.

Tang

Mendez

Theriot

et al.

Minimum conditions for the induction of cortical spreading depression in brain slices. J Neurophysiol 2014 15; 112: 2572–2579.

86.

Reid

Marrannes

De Prins

et al.

Potassium translocation and spreading depression induced by electrical stimulation of the brain. Exp Neurol 1987; 97: 345–364.

87.

Herreras

Largo

Ibarz

et al.

Role of neuronal synchronizing mechanisms in the propagation of spreading depression in the in vivo hippocampus. J Neurosci 1994; 14: 7087–7098.

88.

Larrosa

Pastor

López-Aguado

et al.

A role for glutamate and glia in the fast network oscillations preceding spreading depression. Neuroscience 2006; 141: 1057–1068.

89.

Zhou

Rungta

Malik

et al.

Regenerative glutamate release by presynaptic NMDA receptors contributes to spreading depression. J Cereb Blood Flow Metab 2013; 33: 1582–1594.

90.

Pietrobon

. Function and dysfunction of synaptic calcium channels: Insights from mouse models. Curr Opin Neurobiol 2005; 15: 257–265.

91.

Van Harreveld

. Two mechanisms for spreading depression in the chicken retina. J Neurobiol 1978; 9: 419–431.

92.

Akerman

Holland

Goadsby

. Mechanically-induced cortical spreading depression associated regional cerebral blood flow changes are blocked by Na+ ion channel blockade. Brain Res 2008; 1229: 27–36.

93.

Sugaya

Takato

Noda

. Neuronal and glial activity during spreading depression in cerebral cortex of cat. J Neurophysiol 1975; 38: 822–841.

94.

Charles

Hansen

. Migraine aura: New ideas about cause, classification, and clinical significance. Curr Opin Neurol 2015; 28: 255–260.

95.

Hansen

Ashina

. Calcitonin gene-related peptide and migraine with aura: A systematic review. Cephalagia 2014; 34: 695–707.

96.

Hansen

Hauge

Olesen

et al.

Calcitonin gene-related peptide triggers migraine-like attacks in patients with migraine with aura. Cephalagia 2010; 30: 1179–1186.

97.

Tozzi

de Iure

Di Filippo

et al.

Critical role of calcitonin gene-related peptide receptors in cortical spreading depression. PNAS 2012; 109: 18985–18990.

98.

Houben

Loonen

Baca

et al.

Optogenetic induction of cortical spreading depression in anesthetized and freely behaving mice. J Cereb Blood Flow Metab 2017; 37: 1641–1655.

99.

Kaufmann

Theriot

Zyuzin

et al.

Heterogeneous incidence and propagation of spreading depolarization. J Cereb Blood Flow Metab 2016; 37: 1748–1762.

100.

Charles

. Sounding out migraine-related interactions between the brainstem and cortex. Cephalagia 2015; 35: 941–943.