Background
Migraine is a common disabling headache disorder, affecting between 10% and 20% of the population worldwide, [
1] and is characterized by recurrent headaches of moderate to severe intensity, pulsating quality, and unilateral location. Headaches are aggravated by routine physical activity and are associated with nausea, photophobia, and/or phonophobia [
2].
In recent years, neuroimaging technology has provided more convenient methods for better understanding the pathogenesis of neurological disorders. Several brain magnetic resonance imaging (MRI) studies have identified structural and functional changes in patients with migraine and have suggested that the alterations of these changes may be associated with the pathophysiology of migraine [
3‐
7]. These changes have been demonstrated in patients with migraine in different phases of the disease [
8]. In the prodrome period, hypothalamus, pons, spinal trigeminal nuclues and visual cortex may be involved, and viusal cortex is associated with the aura period. In addition, various regions including thalamus as well as cingulate cortex, cerebellum, periaqueductal gray, hypothalamux, pons, spinal trigeminal nucleus, visual cortex, middle frontal cortex, somatosensory and temporo-occipital cortex can be related with ictal phase of the migraine [
8].
In the pathogenesis of migraine, the role of the thalamus has been considered as the relay center for ascending nociceptive information, via the trigemino-vascular pain pathway, from lower brain areas to various cortical regions [
9]. The presence of multisensory symptoms during migraine attacks and the central role of the thalamus indicate a potential involvement of the thalamus in the pathogenesis of migraine [
9]. Recent neuroimaging research, with advantages of brain MRI techniques, has revealed an expanding spectrum of additional structural and functional roles of the thalamus in migraine, which could provide a better understanding of the pathophysiology of migraine. In a structural study, there were no changes in the overall volume of the thalamus in patients with migraine compared with healthy controls [
6,
7]. However, in another study with 37 migraineurs, T1 relaxation time was significantly shorter in the thalamus of migraineurs compared with healthy controls [
3]. In addition, the magnetization transfer ratio was higher and the T2* relaxation time was shorter in migraineurs with aura [
3]. These data reveal broad microstructural alterations in the thalamus of migraineurs compared healthy controls, suggesting increased iron deposition and myelin content/cellularity. In a functional study, 17 patients with migraine underwent resting-state functional MRI scan during migraine attacks [
4]. The authors found increased functional connectivity between the right thalamus and several contralateral brain regions, such as superior parietal lobule, insular cortex, primary motor cortex, supplementary motor area, and orbitofrontal cortex, whereas decreased functional connectivity was noted between the right thalamus and ipsilateral brain areas, including the primary somatosensory cortex and premotor cortex [
4]. The study suggested that network connectivity between the thalamus and the pain-modulating as well as pain-encoding cortical areas were affected during migraine attacks [
4]. Another study assessed the local levels of glutamate/glutamine and gamma-aminobutyric acid in the thalamus in patients with migraine and healthy controls using proton magnetic resonance spectroscopy [
5]. That study revealed significantly increased glutamine levels in the thalamus, suggesting increased regional excitability [
5]. All of these studies commonly demonstrate the pivotal role of the thalamus in the pathogenesis of migraine.
The aim of this study was to investigate the alterations of thalamic nuclei volumes and the intrinsic thalamic network in patients with migraine compared with healthy controls. We hypothesized that there would be significant alterations of thalamic nuclei volumes or network, which could be related to the pathogenesis of migraine.
Discussion
The main finding of this study was that anteroventral and medial geniculate nuclei volumes were significantly increased with decreased volumes of parafascicular nuclei in the patients with migraine without aura compared with healthy controls, despite the fact that thalamic volumes as a whole were not changed. However, the intrinsic thalamic network was not different between them.
Only one study had investigated thalamic nucleus volumes in the patients with migraine [
13]. They found significant volume reductions in central nuclear complex, anterior nucleus, and lateral dorsal nucleus. Because the central nuclear complex belongs to the intralaminar nuclei and includes the central medical and the parafascicular nuclei, the results of the previous study was partially in agreement with our study [
13]. However, the previous study had several limitations compared to our study. Their patients were a heterogeneous group. They pooled data of patients with migraine with and without aura. In addition, the patients were enrolled in the four centers, which had different scanners with large variability in coil and gradient properties between them. In addition, they segmented the thalamus into 10 thalamic nuclei, only. We investigated the alterations of 25 individual thalamic nuclei volumes, and firstly analyzed the intrinsic thalamic network in patients with migraine.
The anteroventral nucleus receives dense limbic-related projections from the mammillary nuclei via the mammillo-thalamic tract and the medial temporal lobe via the fornix [
14]. The output of this nucleus is primarily directed to the cingulate gyrus through the anterior limb of the internal capsule [
14]. Thus, the anteroventral nucleus is an important synaptic station in the Papez circuit, which is related to emotion and memory acquisition [
14]. We found that the volume of the anteroventral nucleus was increased in patients with migraine. Migraine is related to highly specific conditioning or sensitization to pain-related stimuli [
15]. Sensitization refers to the process in which neurons become increasingly responsive to nociceptive or non-nociceptive stimulation with decreased response thresholds [
15]. In the patients with migraine, not only nociceptive pain but also emotional words or emotional negative affect may act as migraine triggers [
16]. Emotional stress is one of the most common triggers of acute migraine attack, attributed to about 80% of attacks [
17]. In addition, patients with migraine have higher levels of perceived stress than healthy controls [
18]. From these results, we can speculate that the increased volume of the anteroventral nucleus may produce a sensitization to emotional stress in patients with migraine.
The medial geniculate nucleus is considered to be part of the lateral thalamic nuclear group. The medial geniculate nucleus receives ascending auditory input via the brachium of the inferior colliculus and projects to the primary auditory cortex in the temporal lobe [
14]. Thus, the medial geniculate nucleus is thought to be primarily responsible for auditory perception. Patients with migraine often report an aversion to various sensory stimuli during an acute attack, such as sound (phonophobia) [
19]. Phonophobia can be defined as aversion to normally nonaversive sounds. This symptoms are reported in 70% to 80% of patients with migraine during an acute attack [
20]. It is a plausible explanation that increased volumes of the medial geniculate nucleus might be related to phonophobia in patients with migraine. An interesting report using positron emission tomography examined the changes in regional cerebral blood flow as an index of neuronal activity in the human brain during migraine attacks [
21]. During the attacks, increased blood flow was found in the auditory association cortex [
21].
The parafascicular nucleus is one of the intralaminar nuclei, which are characterized by their projections to the neostriatum and to other thalamic nuclei, along with diffuse projections to the cerebral cortex [
14]. Thalamic regions have been traditionally proposed to support pain processing and arousal [
22,
23]. We found that patients with migraine had significantly decreased parafasicular nucleus volumes. In an animal study on familial hemiplegic migraine (FHM) that examined the effects of an FHM-1 mutation in the central trigeminal nociceptive pathway, transgenic mice expressing the FHM-1 mutation had more activation in the intralaminar nuclei following nociceptive trigemino-vascular stimulation when compared with wild-type animals [
24]. FHM is a subtype of migraine characterized by hemiplegic aura, and two-thirds of patients with FHM also experience typical migraine attacks [
25]. We can assume that alterations of the parafascicular nucleus could be related to the pathogenesis of migraine attacks. Furthermore, the parafascicular nucleus projects to the rostral and lateral areas of the frontal lobe but is more closely related with frontal lobe [
14]. A previous meta-analysis on changes in gray matter in patients with migraine showed that the migraineurs had decreases in gray matter volume mainly in the frontal lobe, such as the inferior frontal gyrus, precentral gyrus, and middle frontal gyrus [
26]. This result is in agreement with the findings of our present study.
We also investigated the intrinsic thalamic network based on individual thalamic nuclei volumes and found no alterations of the thalamic network in the patients with migraine compared with healthy controls. Structural connectivity refers to anatomical connections linking a set of neural elements [
27]. There is a lot of evidence on the abnormal thalamo-cortical network in patients with migraine using functional MRI data [
28] and diffusion tensor imaging study [
29]. However, we focused on the intrinsic thalamic network and demonstrated its well-preserved status in patients with migraine despite alterations in the volume of individual thalamic nuclei.
The strength our study was that we enrolled only newly diagnosed patients with migraine and included migraine without aura to increase the homogeneity of subjects group. In addition, this is the first migraine research study to investigate the various thalamic nuclei volumes and focus on alterations of the thalamic nuclei volume and network compared with healthy controls.
However, this study has several limitations. First, this study used a cross-sectional design. This design makes it difficult to discover the causal relationship between structural changes and clinical features and the role of the thalamus in the pathophysiology of the disorder. Longitudinal studies with larger sample sizes are needed to confirm our results. Second, we used only the FreeSurfer program based on the T1-weighed images for the segmentation of the thalamic nuclei. It was suggested that a multimodal imaging strategy (T1- and T2-weighted images as well as diffusion tensor images) could improve the accuracy of thalamic segmentation [
30]. However, among the most sophisticated programs of MRI analysis currently available, the FreeSurfer program represents a set of automated tools most widely used to reconstruct the brain’s structures. FreeSurfer usually offers a higher and more robust reproducibility compared with other neuroimaging analysis techniques [
31]. In addition, a previous study using MRI-based thalamic nuclei volume analysis like our study demonstrated a good agreement with previous histological studies and showed an excellent test-retest reliability [
32]. Third, a recent systemic review of previous functional connectivity sutides in migraine showed a poor level of reproducibility and no migraine specific pattern in functional network [
33]. Migraine is a heterogeneous disorder, which might cause variation in results between studies. In addition, no sample size or power calculation guidelines are available for functional connectivity studies. It would be needed to consider multicenter studies to allow for better and more reproducible studies [
33]. However, we investigated the structural connectivity based on thalamic nuclei volumes, which could be a more stable method than functional connectivity study.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.