Introduction
Visual aura is the most common transient neurological symptom experienced by patients during attacks of the so-called “migraine with aura” [
1].
In the last decades, the relationship between the aura phenomenon, most likely underlied by cortical spreading depression (CSD), and trigeminal nociceptive pathway activation, subtending headache attacks, has been widely debated [
1].
Preclinical evidence strongly supported that CSD-related up-regulation of c-fos expression in trigeminal nucleus caudalis, local release of pro-nociceptive molecules and thalamic activation may modulate the nociceptive inputs to the cortex [
2,
3]. However, despite the growing body of functional neuroimaging and electrophysiological studies investigating the neuronal correlates of pain processing in patients with migraine without aura (MwoA) [
4], to our knowledge, no similar studies have been conducted in patients with migraine with aura (MwA).
Therefore, in the present study we aimed to explore the functional brain response to trigeminal heat stimulation (THS), using whole-brain blood oxygen level-dependent (BOLD)-fMRI, in patients with MwA compared to healthy controls (HC). To examine the specificity of any observed differences between patients with MwA and HC, we further studied a group of age and sex-matched patients with MwoA.
We hypothesized that cerebral regions involved in visual aura phenomenon would be differentially responsive between patients with MwA and HC. We further hypothesized that functional changes associated with visual aura phenomenon would be present in patients with MwA but not in those with MwoA.
Discussion
In the present study, using a whole-brain BOLD-fMRI approach we found, consistently with previous observations, that THS activates cortical and subcortical areas, known to be involved in central pain processing [
9]. While all participants demonstrated this general pattern of response, patients with MwA exhibited, during moderately painful stimuli, an increased activation of higher cortical areas constituting a known distributed network associated with AVN, including lingual gyrus, inferior parietal lobule, inferior frontal gyrus and medial frontal gyrus [
10‐
13]. Moreover, a significantly greater cerebellar activation was observed in patients with MwA compared to both HC and MwoA. Interestingly, the magnitude of cerebellar activation was correlated with disease duration of patients with MwA. However, pain intensity ratings were not significantly different between the three groups at any level of THS.
In the last decades, based on the prominent role played by visual symptoms experienced by patients with MwA, visual pathways have been extensively explored to identify whether the visual pathways hyperexcitability could subtend both the hyperresponsiveness to visual stimuli and the increased propensity to experience aura or, contrarily, whether the abnormal visual processing associated with aura phenomenon leads to brain hyperresponsiveness to visual stimuli [
14].
Advances in functional neuroimaging techniques have dramatically improved our insight of mechanisms underlying aura phenomenon in migraine patients. Among these, visual pathways activation has been observed in patients with MwA, suggesting that cortical hyperresponsiveness in migraine is directly related to the presence of aura in these patients [
15]. More recently, in a multiparametric fMRI study, a stronger functional connectivity in visual network (specifically in lingual gyrus) has been demonstrated in patients with MwA compared to HC and patients with MwoA, in absence of structural or microstructural abnormalities in the same brain areas [
16]. These data are well-fitting with abnormal levels of GABA and glutamate observed by magnetic resonance spectroscopy in the primary visual cortex of MwA patients during visual stimulation [
17].
However, despite the growing body of literature exploring the neural correlates of both aura and headache phases during the migraine cycle [
18,
19], only few studies have specifically assessed brain activity in patients with MwA during experimental moderately painful stimulations. Among these, two seminal PET studies showed the peculiar reciprocal interactions between pain experience and aura phenomenon in migraine patients [
20,
21]. Specifically, these findings have demonstrated that visual pathway activation by luminous stimuli was potentiated by trigeminal nociception, strongly suggesting a “visual network-pain network” integration in these patients.
The present results, in line with above-mentioned observations show, for the first time, the hyperresponsiveness of the AVN during experimental moderately painful stimulations, in patients with MwA when compared to patients with MwoA and HC.
The AVN, encompassing areas of the dorsal processing stream and fronto-lateral regions, is relevant for processing the spatial attributes of visual information, memory-guided saccades, spatial working memory and the executive control of spatial attention, as well as in different types of saccadic eye movements as demonstrated by several studies in human and non-human primates [
11‐
13,
22,
23]. The hypothesis that neural mechanisms subtending aura phenomenon may be strictly correlated with specific hyperresponsiveness of brain areas involved in AVN has been recently supported [
10]. Indeed, in the same cortical areas an increased BOLD response to visual stimulation has been observed in patients with MwA, with side-fixed visual aura attacks, in hemispheres involved by aura phenomenon when compared with the hemisphere not involved by aura phenomenon in the same patients (even during the interictal phase) and HC [
10]. Finally, the involvement of the medial frontal gyrus has been observed in structural and functional studies in patients with migraine, specifically in those experiencing aura phenomenon [
24,
25].
Beside the activation of AVN, we observed a significantly increased BOLD-response to THS in the midline and lower structures of the cerebellum (e.g.: uvula, tonsils and biventer lobule) in patients with MwA when compared with both patients with MwoA and HC. Although its role has been previously underestimated in human pain perception, it is now well-known that cerebellum is involved in pain processing [
26,
27] as demonstrated also by altered experimental pain perception after cerebellar infarction [
28]. More specifically, during trigeminal nociception, the activation of specific cerebellar areas as well as their functional connections with both the descending anti-nociceptive network and cortical hubs of pain processing have been observed [
29].
Based on these observations, several structural and functional studies have been conducted in the attempt to clarify the putative role played by the cerebellum in migraine [
30]. Among these, an increased activity of the cerebellar crus with a concomitant decreased functional connectivity with thalamus have been demonstrated during trigeminal stimulations. The findings suggest a recruitment of cerebellar additional resources to overcome the diminished functional cerebellar-thalamic connectivity in patients with MwoA [
30]. However, the cerebellar compensatory mechanism seems to be ineffective, inducing a dysfunctional thalamic gating of external stimuli addressed to cortical areas and leading to susceptibility for migraine attacks [
31]. Interestingly, a reduced cerebellar inhibitory control on cerebral cortex has been supported also in patients with MwA, using a transcranial magnetic stimulation protocol [
32].
The increased cerebellar activity founded in the present study suggest the same compensatory and ineffective, in other word “maladaptive”, mechanism in patients with MwA during trigeminal nociceptive experience.
Considering our cortical and cerebellar findings, based on the well-known connections between cerebellar structures (e.g. pyramid, uvula, tonsils and biventer lobule) and AVN [
33,
34] we speculate that the AVN hyperresponsiveness to THS might be due to the lack of cerebellar inhibitory control on thalamic sensory gating and, thus, on the brain cortex.
Our BOLD-fMRI cortical (e.g.: lingual gyrus, inferior parietal lobule, inferior frontal gyrus and medial frontal gyrus) findings showed no correlations with MwA clinical features.
Finally, although previous data supported that the activity of the cerebellum is modulated by the perceived intensity of pain, the increased activity of cerebellum and AVN does not affect pain intensity perception in patients with MwA, being the pain ratings not significantly different between the three study groups, consistently with previous experimental data [
6,
7,
35].
Taken together, our findings suggest that MwA may be characterized by a peculiar hyperresponsiveness of the distributed AVN not only to visual, as previously demonstrated [
10] but also to moderately painful trigeminal stimulations, suggesting a “visual-pain networks integration”, likely by means of a “maladaptive” activity of cerebellum.
Finally, our secondary analyses did not show statistically significant association between BOLD-fMRI response and clinical parameters of disease severity except for disease duration, showing statistically significant association with BOLD-fMRI response not surviving when a more conservative
p-value threshold (Bonferroni correction) has been used. Thereby, we believe that visual aura phenotype per se could justify the observed increased activity in higher cortical areas of the network associated with advanced visual processing and in the cerebellum in patients with migraine with aura, as previously demonstrated also for resting state visual network functional connectivity [
16].
Our study is not exempt from some limitations. First of all, we cannot completely exclude the putative effect of habituation or sensitisation phenomenon due to repetitive THS on our findings, despite pain intensity ratings were not changed in the course of experimental stimulations, the random modality of the THS and the well-known absence of habituation in patients with migraine. Moreover, we cannot compare the AVN response to THS with response to visual stimulations (not included in the experimental paradigm). In addition, patients with MwoA and MwA experienced different frequency of attacks. On the other hand, because this is general to the migraine population, we believe that it may give us the opportunity to evaluate the “natural history and phenotype” of the two migraine subtypes.