Altered Regional Brain Functional Activity Predominantly Involving the Right Superior Temporal Gyrus in Patients With Vestibular Migraine Diagnosed According to the Diagnostic Criteria of the Bárány Society and the International Headache Society

Background: Vestibular migraine (VM) is considered one of the most common cause of episodic central vestibular disorders, the mechanism of VM is currently still unclear. It is worth investigating whether VM belongs to the migraine subtype or is a separate disorder. The development of functional nuclear magnetic resonance (fMRI) in recent years offers the possibility to explore the pathogenesis of VM in depth. The study aimed to investigate resting-state functional brain activity alterations in patients with VM diagnosed based on the diagnostic criteria of the Bárány Society and the International Headache Society. Methods: Seventeen patients with VM who received treatment in our hospital from December 2018 to December 2020 were enrolled. Clinical data of all patients were collected. Eight patients with migraine and 17 health controls (HCs) were also included. All subjects underwent fMRI examination. The amplitude of low frequency uctuation (ALFF), fractional amplitude of low frequency uctuation (fALFF) and regional homogeneity (ReHo) were calculated to observe the changes in spontaneous brain activity in patients with VM. Then brain regions with altered spontaneous brain activity were selected for seeded-based functional connectivity (FC) analysis to explore the changes in FC in patients with VM. Results: Among 17 patients with VM, there were 7 males and 10 females with an average age of 39.47±9.78 years old. All patients with VM had a history of migraine. Twelve (70.6%) patients with VM had recurrent spontaneous vertigo, 2 (11.7%) patients had visually-induced vertigo, and 3 (17.6%) patients had head motion-induced vertigo. All 17 patients with VM reported worsening of dizziness vertigo during visual stimulation. The migraine-like symptoms were photophobia or phonophobia (n=15, 88.2%), migraine-like headache (n=8, 47.1%), visual aura during VM onset (n= 7, 41.2%). 5 (29.4%) patients with VM had hyperactive response during the caloric test, and 12 (70.6%) patients had caloric test intolerance. Eleven (64.7%) patients had a history of motion sickness. VM patients showed exhibited signicantly increased ALFF and fALFF values in the right temporal lobe (STG and MTG), and signicantly increased ReHo values in the right STG, MTG and ITG in comparison with HCs. Compared with patients with migraine, patients with VM showed signicantly decreased ALFF values in the right median cingulate and paracingulate gyri, signicantly increased fALFF values in the right parietal lobe (postcentral gyrus and superior parietal gyrus), and the right frontal lobe (supplementary motor areas and dorsolateral superior frontal gyrus), as well as signicantly increased ReHo values in the right thalamus. Compared with HCs, patients with migraine showed signicantly increased ALFF values in the right limbic lobe (right parahippocampal gyrus and right fusiform gyrus), left ITG and the right frontal lobe (supplementary motor areas, right median cingulate and paracingulate gyri, and right right inferior frontal gyrus), signicantly decreased ALFF values in the pons and brainstem, signicantly decreased ReHo values in the frontal cortex (including left and right supplementary motor areas, left dorsolateral superior frontal gyrus, left median cingulate and paracingulate gyri, right paracentral lobule, right dorsolateral superior frontal gyrus, left and right middle frontal gyrus). lobe-frontal migraine both had altered but the mechanism seems to be


Introduction
Vestibular migraine (VM) is considered the most common cause of central episodic vertigo, accounting for 7 % of patients seen in neuro-otological clinics and 9% of patients seen in headache clinics [1]. In 2012, diagnostic criteria for VM was jointly formulated by the Bárány Society and the Migraine Classi cation Subcommittee of the International Headache Society, which were published in the International Classi cation of Headache Disorders, 3rd edition (beta version) (ICHD-III beta) in 2013 [2]. At present, the clinical diagnosis of VM still only relies on clinical symptoms [3]. During VM attacks, patients usually have abnormalities in perception of voluntary movement, such as rotation or swing, which are often accompanied by increased visual motion sensitivity and spatial disorientation [4]. Sometimes even during the interictal period, nystagmus may be observed in patients, suggesting that there is a certain degree of brain stem involvement [5]. Among the hypotheses about the mechanisms of VM, most hypotheses categorize VM as a subcategory of migraine and suggest that there is overlap among VM and vestibular pathway damage. However, the current vestibular testing techniques are not yet able to make the differentiation of VM from other vestibular diseases [6], there is still debate as to whether VM is central or peripheral in origin [7,8].
With the development of functional imaging techniques, such as the application of positron emission tomography (PET) and functional nuclear magnetic resonance (fMRI) provides a possibility of studying the mechanism of VM.
Previous studies have shown that patients with VM often have abnormal vestibular information processing and abnormal functional interactions between visual and vestibular networks in the cerebral cortex during tilt [9, 10].
Shin et al. [10] performed 8F-deoxyglucose PET study of two patients with VM during the interictal and ictal period, and found that compared with interictal period of VM, patients showed increased metabolism in the bilateral cerebellum, frontal cortices, temporal cortex, posterior insula, and thalami, and decreased metabolism in the occipital cortex during the ictal period of VM, increased metabolism in the temporo-parieto-insular areas and bilateral thalami indicates the activation of the vestibulo-thalamo-cortical pathway, decreased metabolism in the occipital cortex suggests reciprocal inhibition between the visual and vestibular systems. These ndings revealed that metabolic activities in the visual and vestibular cortical areas were altered during acute VM attack.
In previous studies [11,12], scholars have used task fMRI data from different tasks to investigate VM. A task-state fMRI study [11] in healthy volunteers exposed to magnetic vestibular stimulation showed that the brain regions involved in balance and spatial navigation are complex networks, including the insula, superior temporal gyrus (STG) and inferior parietal lobule, precentral gyrus and postcentral gyrus, anterior insula and inferior frontal gyrus, anterior cingulate gyrus and hippocampus. Russo et al. [12] performed a fMRI study in 12 patients with VM during ear irrigation with cold water and found that patients with VM showed signi cantly increased activation in the left dorsal thalamus during left-sided vestibular stimulation, and the authors speculated that the pain and vestibular sense in patients with VM may be related to the thalamus [13,14].
At present, fMRI studies on VM patients are mostly task-based, and most study showed that activation of multilevel vestibular integrating centers and cortical areas related to higher-level vestibular information processing is present in patients with VM. However, task-related brain regions are sometimes not associated with brain functional changes related to VM itself, which may be associated with brain functional changes related to task stimulation. Therefore, the real-time measurement of spontaneous brain functional activities by using restingstate fMRI can better re ect the pathogenesis of VM [15].
Under this background, we investigated the changes in resting-state brain spontaneous activity and functional connectivity (FC) between patients with VM, health controls (HCs) and patients with migraine by using restingstate regional brain activity measures including amplitude of low frequency uctuation (ALFF), fractional amplitude of low frequency uctuation (fALFF), regional homogeneity (ReHo), and seed-based functional connectivity analysis in this study, and aimed to explore functional brain activity alterations in patients with VM and its possible mechanisms involving the vestibular pathway.

Study subjects
A total of 17 patients with VM who received treatment in Aerospace Center Hospital, Peking University Aerospace School of Clinical Medicine between December 2018 and December 2020 were included. Clinical data of all patients were collected. Blood pressure, cardiac function, thyroid function tests, immune-related laboratory tests and lower limb electromyography were performed to exclude other medical diseases. Videonystagmography, caloric test, head impulse test, and vestibular evoked myogenic potential stimulation were performed to exclude peripheral vestibular lesions. MRI was performed to exclude severe neurological diseases.VM was diagnosed based on Bárány Society's diagnostic criteria (3) as following: (A) at least 5 episodes of vestibular symptoms of moderate or severe intensity lasting 5 minutes to 72 hours; (B) current or previous history of migraine with or without aura according to the diagnostic criteria of ICHD; (C) one or more of the following migraine features with at least 50% of the vestibular episodes: 50% of vestibular attacks are accompanied by at least one migraine symptom: (1) headache with at least two of the following characteristics: one sided location, pulsating quality, moderate or severe pain intensity, aggravation by routine activity; (2) photophobia and phonophobia; (3) visual aura; (D) not better accounted for any other vestibular or IHCD diagnosis. During the caloric test, the left and right external auditory canals were irrigated with cold (30 °C) and warm (44 °C) water with the patient in the supine position and their head raised by 30°. The right external auditory canal was rst irrigated with warm water, followed by the left external auditory canal with warm water and the right external auditory canal with cold water, and last the left external auditory canal with cold water (i.e., right warm (RW), followed by left warm (LW) and right cold (RC), and last left cold (LC)). There was a 5-min interval between irrigations. The slow-phase velocity (SPV) during irrigation was recorded, and the canal paresis (CP) value was calculated. A CP value of >25% indicates reduced unilateral horizontal semicircular canal function [17]. A sum of the SPV values of the bilateral semicircular canals of ≤12°/s suggests reduced bilateral horizontal semicircular canal function. The criteria for hyperactive responses were as follows: total peak cool response (LC + RC) of >99°/s, total peak warm response (LW + RW) of >146°/s, total peak response (LC + RC + LW + RW) of >221°/s.
Caloric test intolerance refers to the main symptoms, including obvious nausea, vomiting, numbness in hands and feet, and body stiffness. Caloric test intolerance was considered severe if its duration was >1 h.

Image acquisition
All subjects were scanned using a 3.0-Tesla MR (SIEMENS MAGNETOM Skyra syngo MR D13, Germany) with a 16-channel head and neck coil. During image acquisition, subjects' head was immobilized to avoid head movement. Subjects were asked to relax with their eyes closed, stay awake throughout the scanning. Images was obtained in sagittal plane using a 3D gradient-echo T1WI sequence with the following parameters: MRI image data processing Image data were processed using DPARSFA software on the MATLAB2013 platform, ALFF analyisis ALFF analysis was done with following steps: (1) data conversion: image data were converted from DICOM format to NIFTI format that can be processed by SPM12; (2) removal of the rst 10 time points: due to the possible instability of the initial MRI signal caused by T1 relaxation and the need for participants to adapt to the scanning environment, the rst 10 time points were deleted; (3) time correction: time correction is performed to ensure that all voxels within one volume had been acquired at the same time, the 35th slice was chosen as the reference slice; (4) head movement correction: slight head motion of the subject between the time points during the scan was corrected, in order to ensure the accuracy of the position information, the subjects who had more than 1.5 mm head translation in x-, y-, or z -direction and 1.5° head rotation were removed; (5) removal of linear drift: linear trend formed due to heating of MR scanner and fatigue in subjects after long MRI scanning time was removed; (6) regression: a linear regression model was used to remove the interference signal in the BOLD signal; (7) spatial normalization: to solve the problems related to difference in brain morphology among different subjects and the inconsistencies in spatial position during scanning, fMRI images were spatially normalized to a standard space using DARTEL and resampled at a resolution of 3 mm×3 mm×3 mm; (8) spatial smoothing: smoothening was conducted with a Gaussian kernel of 8 mm × 8 mm× 8 mm to reduce registration errors and increase the normality of the data; (9) ALFF calculation: the power spectrum is obtained by converting signals in time domain into the frequency domain using the Fast Fourier Transform, the average square root of the power spectrum was calculated as the ALFF value; (9) fALFF calculation: fALFF value was calculated at frequency range of 0.01-0.1 Hz.
Statistical analysis fMRI data analysis was performed using SPM 12 software. Comparison of the difference in results of ALFF, REHO, FC between groups were tested using a two-sample T test and adjusted for covariates including age and sex. P < 0.01 was considered to be statistically different. Multiple comparisons corrections were performed using familywise-error correction (FWE). Results were visualized using XJVIEW and BrainNet.

Results
Clinical data of patients included in the study P=0.001, FWE corrected, respectively, Table 2). ALFF and fALFF refer to the intensity of spontaneous brain activity. The higher the values, the greater the spontaneous brain activity in the brain regions. Increased ALFF values in the right temporal lobe in patients with VM indicated that the intensity of spontaneous functional activity in this region was enhanced.
Compared with migraine group, VM group showed signi cantly decreased ALFF values in the right median cingulate and paracingulate gyri (X=21, Y=-30, Z=45, P=0.009, FWE corrected, Table 2), indicating that the intensity of spontaneous functional activity in this region in patients with VM was weakened. VM group also showed signi cantly increased fALFF values in the right parietal lobe (postcentral gyrus and superior parietal gyrus) (X=24, Y=-45, Z=69, P=0.011, FWE corrected,  Table 2). ReHo refers to the synchronization of spontaneous functional activities among an individual voxel and surrounding voxels. Increased ReHo values in the right temporal lobe indicated an increase in consistency of regional brain activity in this region in patients with VM.

Regions with increased ReHo values contain regions with increased ALFF values and adjacent regions. Increased
ReHo values and ALFF values were both observed in the right temporal lobe, indicating that the regional functional activities in the right temporal lobe (STG, MTG, and ITG) in patients with VM were enhanced.
Patients with VM showed signi cantly increased ReHo values in the right thalamus compared with patients with migraine (X=12, Y=-21, Z=-3, P=0.043, FWE corrected, Table 2)), suggesting an increase in consistency of regional brain activity in this region in patients with VM. In this study, migraine without aura was most common in patients with VM. The most common migraine-like symptoms in VM patients was photophobia/phonophobia (n=15, 88.2%), followed by migraine-like headache (n=8, 47.1%), and visual aura (n=7, 41.2%). This result is consistent with previous studies made by Zhang et al.
In this study, we found that compared with HCs, spontaneous functional activity in the right temporal lobe (STG and MTG), ReHo values in the right temporal lobe (STG, MTG and ITG), and FC within the right temporal lobe were both enhanced in patients with VM, suggesting enhanced regional functional activity in the right temporal lobe in patients with VM, this may be related to the mechanism of photophobia/phonophobia. A study have shown that lesions in regions within the temporal lobe such as STG (AAL; BA22), MTG, and ITG are often associated with allocentric de cits [29]. Allocentric processing is known as allocentric spatial discrimination ability, i.e., impaired stimulus-centered advanced navigation system. Medina et al. [30] performed MRI and spatial cognition tests on 171 patients with right supratentorial ischemic stroke and found that parts of the dorsal stream of visual processing, including the right supramarginal gyrus, are involved in spatial encoding in egocentric coordinates, parts of the ventral stream, including the posterior ITG, are involved in allocentric processing. Priyanka et al. [31] performed repetitive transcranial magnetic stimulation (rTMS) on the right STG (48, −20, −8) in 11 right-handed healthy subjects, revealed impaired stimulus-centered spatial cognition in patients. Patients with VM in the present study are similar to healthy subjects who underwent rTMS on right STG in the above-mentioned study of Priyanka et al, indicating that ventral stream of visual processing and allocentric spatial cognition were impaired in patients with VM. We speculated that enhanced spontaneous functional activity in the right temporal lobe (STG, MTG and ITG) in patients with VM may be related to patients' discomfort and worsening of dizziness during visual stimulation. Xu et al. [32] found that MTG is associated with auditory information processing, enhanced regional functional activity of MTG in patients with VM is presumed to be related with symptoms of phonophobia Furthermore, our ndings showed that vertigo attacks in patients with VM may be associated with increased regional spontaneous activity in the right parietal lobe-frontal lobe-thalamus. (1) Compared with patients with migraine, spontaneous functional activity in the right parietal lobe (postcentral gyrus and superior parietal gyrus) (X=24, Y=-45, Z=69) and right frontal lobe (SMA and dorsolateral superior frontal gyrus) (X=24, Y=-45, Z=69) was enhanced in patients with VM. It is speculated that patients with VM has abnormalities in visual-spatial information integration, and higher-level vestibular cortical integration areas are in a sensitive state. Previous studies based on working memory have shown that the dorsolateral superior frontal gyrus and superior parietal lobule (superior parietal gyrus) received visual and spatial sensory stimuli and integrate visual information ow and spatial information (vestibular information related to position) [33], the right frontal lobe is related to visualspatial tasks [34]. Generally, visual-spatial information integration requires the combined effects of top-down (cognitive control and goal-directed) modulation and bottom-up inputs (capture of external sensory stimuli) in the superior parietal lobule [35,36]. Postle et al.
[37] applied rTMS stimulation over superior parietal gyrus, and found that subjects' performance was not stable during tasks requiring visual-spatial coordination, and changes in the parietofrontal network is observed, application of rTMS on superior parietal gyrus is similar to enhanced spontaneous activity observed in patients with VM in this study.
(2) Compared with patients with migraine, increase in consistency of regional brain activity in the right thalamus was observed in patients with VM. The thalamus is the relay station for all sensory input, it is also considered as an integrated area of multisensory vestibular information [38]. As mentioned above, compared with patients with migraine, VM patients showed enhanced regional functional activity in the thalamus, right parietal lobe (postcentral gyrus and superior parietal gyrus) and right frontal lobe (SMA and dorsolateral superior frontal gyrus). It is speculated that the vestibular pathway, right parietal lobe-frontal lobe-thalamus, in patients with VM is in a sensitive state, this may be related to the repeated episodes of vertigo and worsening of dizziness during visual stimulation that are different from those of patients with migraine.
In this study, we also compared the differences in the regional brain activity measures between patients with migraine and HCs. The results showed that enhanced spontaneous functional activity in the right limbic lobe (mainly including the right parahippocampal gyrus and the right fusiform gyrus) and right frontal lobe (SMA, right median cingulate and paracingulate gyri, and right inferior frontal gyrus), decreased spontaneous function activity in the pons and brainstem, as well as reduced ReHo values in the frontal cortex in patients with migraine compared with HCs. This results is consistent with the results of previous fMRI study showing that brain stem and hippocampus play a key role in the onset of migraine attack, and the limbic lobe is involved in the formation of migraine-related pain network, abnormal frontal lobe function can eventually lead to reduced pain threshold in patients with migraine [39]. Our ndings showed that brain regions with altered functional activity in patients with VM are very different from those in patients with migraine relative to HC groups, there appears to be little overlap, and the mechanisms involved may be different.

Limitations
The limitation of the study is the small sample size, further study with larger sample size is needed to further con rm our results.

Conclusions
Our ndings revealed enhanced spontaneous functional activity in the right temporal lobe (STG, MTG and ITG) in patients with VM, indicating that ventral stream of visual processing and allocentric spatial cognition may be impaired. Vertigo attacks in patients with VM may be associated with increased spontaneous activity in the right parietal lobe-frontal lobe-thalamus. Patients with VM and migraine both had altered brain function, but the mechanisms involved seem to be different. Our ndings provide evidence for the regional functional properties of brain regions in patients with VM and for further functional connectivity studies based on ROI of multilevel vestibular pathways, as well as brain network connectome studies based on small-world brain graph theory.

Declarations
The criteria for hyperactive responses were as follows: total peak cool response LC + RC of >99°/s, total peak warm response LW + RW of >146°/s, total peak response LC + RC + LW + RW of >221°/s. Caloric test intolerance refers to the main symptoms, including obvious nausea, vomiting, numbness in hands and feet, and body stiffness. Caloric test intolerance was considered severe if its duration was >1 h.
Table 2 Altered regional brain functional activity in patients with VM measured by fMRI Figure 1 Differences in ALFF values between three groups. Vestibular migraine (VM) group exhibited signi cantly increased ALFF values in the right temporal lobe compared with health control (HC) group (X=51, Y=-54, Z=6), indicating enhanced spontaneous functional activity in the right temporal lobe in patients with VM. No signi cant difference was found between migraine and HC groups, as well as VM and migraine groups.

Figure 2
Differences in fALFF values between three groups. Vestibular migraine (VM) group had signi cantly increased fALFF values in the right temporal lobe compared with health control (HC) group (X=69, Y=-45, Z=3), indicating that spontaneous functional activity in this area was enhanced in patients with VM, which corroborated the ALFF results. No signi cant difference was found between migraine and HC groups, as well as VM and migraine groups.  Differences in functional connectivity (FC) between the three groups. Vestibular migraine (VM) group showed signi cantly increased FC in the right temporal lobe (X=51, Y=-45, Z=15) compared with health control (HC) group, indicating that functional connectivity between this region and the seed regions was enhanced (X=69, Y=-45, Z=3), which corroborated the ReHo results. No signi cant difference was found between migraine and HC groups, as well as VM and migraine groups.