Functional reorganization of intranetwork and internetwork connectivity in patients with Ménière’s disease

Ménière’s disease (MD) is associated with functional reorganization not only in the auditory or sensory cortex but also in other control and cognitive areas. In this study, we examined intranetwork and internetwork connectivity differences between 55 MD patients and 70 healthy controls (HC) in 9 well-defined resting-state networks. Functional connectivity degree was lower in MD compared to HC in 19 brain areas involved in the somatomotor, auditory, ventral attention, default mode, limbic, and deep gray matter networks. In addition, we observed lower intranetwork connectivity in the auditory, ventral attention, and limbic networks, as well as lower internetwork connectivity between the somatomotor and limbic networks, and between the auditory and somatomotor, deep gray matter, and ventral attention networks, and between the deep gray matter and default mode network. Furthermore, we identified 81 pairs of brain areas with significant differences in functional connectivity between MD patients and HC at the edge level. Notably, the left amygdala’s functional connectivity degree was positively correlated with MD’s disease stage, and the ventral attention network’s intranetwork connectivity was positively correlated with the healthy side vestibular ratio. Our findings suggest that these functional network reorganization alterations may serve as potential biomarkers for predicting clinical progression, evaluating disease severity, and gaining a better understanding of MD’s pathophysiology. Large-scale network studies using neuroimaging techniques can provide additional insights into the underlying mechanisms of MD.

between the thalamus and visual cortex 13 , decreased connectivity between the left hippocampus and bilateral central opercular 14 , and reduced connectivity between the precuneus and left precentral gyrus 15 .Moreover, they exhibit decreased intranetwork functional connectivity in the right precuneus within the posterior default mode network 16 .Patients with sensorineural hearing loss have demonstrated functional reorganization not only within the auditory network 17,18 , but also in related brain areas involved in sensory processing, motor control, and cognitive functions [19][20][21] .Patients with tinnitus have exhibited abnormal functional brain network connectivity, not only within the auditory-limbic network 22 but also across other networks, including the executive network, control network and default model network [23][24][25] .While these findings are valuable, they predominantly pertain to symptoms shared by MD and similar conditions.Given that MD patients experience similar symptoms, it raises the question of whether comparable functional reorganization of brain networks occurs in individuals with MD.
In this study, our objective was to explore the patterns of functional reorganization in intranetwork and internetwork connectivity between MD patients and healthy controls.Intranetwork connectivity refers to the strength of connections within specific functional brain networks 16,26,27 , representing 'local' interactions within a network.For instance, within the auditory network, regions responsible for processing sound signals are highly interconnected, facilitating efficient communication when processing auditory information.Conversely, internetwork connectivity pertains to interactions between different functional brain networks 16,26,27 , representing 'global' communication.For example, the auditory network may interact with the visual network when processing multisensory information.
We hypothesized that MD could potentially lead to reconfigurations not only within the auditory or sensory cortex but also in other areas associated with control and cognitive functions.To test this hypothesis, we examined the functional connectivity patterns of nine well-defined resting-state networks, including the auditory network (AUN), visual network (VSN), somatomotor network (SMN), dorsal attention network (DAN), ventral attention network (VAN), limbic network (LBN), frontoparietal network (FPN), default mode network (DMN), and deep gray matter network (DGN) at three different levels: integrity level, network level, and edge level.

Clinical findings
We enrolled 55 patients with unilateral MD.The demographic characteristics of the included subjects are presented in Table 1.There was no significant difference in age (p = 0.443), sex (p = 0.988) and education level (p = 0.779) between the MD patients and the HC.
On affected side, no cochlear EH was found in 14 patients (25.5%), mild in 38 patients (69.1%), and significant in 3 patients (5.5%).The proportion of none, mild and significant EH in vestibular was the same as cochlear EH.As for non-affected side, cochlear EH was rated none in 48 patients (87.3%) and mild in 7 patients (12.7%).No vestibular EH was identified in 50 patients (90.9%) and mild in 5 patients (9.1%).

Discussion
The present study employed rs-fMRI to investigate the patterns of functional reorganization in intranetwork and internetwork connectivity between individuals with MD and HC.This multi-level analysis explored connectivity integrity, network-level alterations, and specific connections at the edge level to provide a comprehensive understanding of functional connectivity changes in MD.One of the most intriguing findings of this study pertains to the widespread reductions in functional connectivity degree across multiple brain areas encompassing various networks, such as the SMN, AUN, VAN, DMN, LBN, and DGN.These findings suggest a broad disruption in the integration of neural activity within these networks in MD.Furthermore, we observed lower intranetwork connectivity within the AUN, VAN, and LBN, indicating impaired communication between regions involved in auditory processing, attentional control, and emotional regulation.Additionally, we noted lower internetwork connectivity between the SMN and LBN, and between the AUN and SMN, DGN, and VAN, and between the DGN and LBN, DMN.These internetwork connectivity alterations suggest disrupted coordination and information exchange between critical functional networks involved in motor control, emotional processing, and attentional mechanisms.Taken together, these findings provide novel insight into functional organization patterns within and between functional brain networks in MD.The alterations observed in the auditory network provide important insights into the underlying mechanisms of auditory symptoms in MD.Our findings revealed functional connectivity degree was lower in MD compared to HC within the auditory network, indicating disrupted integration of neural activity among regions involved in auditory processing 28,29 .These findings align with the characteristic symptom of hearing loss experienced by individuals with MD.The observed reductions in connectivity within the auditory network may reflect impaired communication between auditory processing regions, potentially contributing to the deficits in auditory perception and discrimination observed in MD patients 1,2 .The auditory network alterations observed in our study may arise from both peripheral and central mechanisms.Peripheral vestibular dysfunction in MD, especially EH, may directly impact the cochlear and vestibular systems, leading to alterations in the auditory processing  pathways 30 .Additionally, central mechanisms involving changes in neural plasticity and compensatory processes could contribute to the observed alterations in the auditory network 20,21 .The disrupted connectivity within the auditory network may interact with and influence other symptoms associated with MD, such as tinnitus and cognitive impairments 25,31 .Our study expands on these findings by revealing additional alterations within the auditory network and shedding light on the broader functional connectivity disruptions in MD.The observed alterations in the auditory network connectivity have potential clinical implications.They may serve as objective biomarkers for diagnosing and monitoring the progression of MD.Future studies could explore the relationship between the severity of auditory symptoms and the degree of connectivity changes within the auditory network.Additionally, investigating the effects of interventions targeted at restoring normal auditory network connectivity could provide valuable insights into potential treatment strategies for MD.
The alterations observed in the SMN, VAN, LBN, and DGN provide valuable insights into the functional disruptions associated with MD.Our findings revealed that functional connectivity degree within these networks was lower in MD compared to HC, indicating compromised integration of neural activity among regions involved in motor control, attentional processing, emotional regulation, and deep gray matter structures 11,22,25 .These alterations may contribute to the motor, attentional, emotional, and autonomic symptoms experienced by individuals with MD.
The lower connectivity degree within the SMN suggests impaired integration of neural activity among motor planning and execution regions.These findings align with the characteristic motor symptoms, including balance impairments and motor coordination difficulties, commonly observed in MD 1,2 .The alterations within the VAN indicate disrupted communication between regions involved in attentional control and salience detection.These changes may underlie attentional disturbances reported by MD patients, such as difficulties in focusing and maintaining attention 1,2 .
The observed alterations within the LBN are consistent with the emotional symptoms frequently associated with MD.The LBN plays a crucial role in emotional regulation and processing, and the lower connectivity within this network may contribute to the increased prevalence of anxiety and depression 22,25 .Additionally, the connectivity changes within the deep gray matter structures, such as the basal ganglia and thalamus, may impact motor control, emotional regulation, and autonomic functions 32 .These alterations could contribute to the diverse symptoms experienced in MD, including emotional instability, autonomic dysfunction, and visceral symptoms.
Moreover, we identified 81 pairs of brain areas exhibiting significant differences in functional connectivity between MD patients and healthy controls (HC) at the edge level.These specific connections may represent critical communication pathways affected by MD and further contribute to our understanding of the disease's pathophysiology.Furthermore, our results demonstrated a positive correlation between the left amygdala's functional connectivity degree and MD disease stage, suggesting that the amygdala's involvement in emotional processing and stress regulation may play a role in disease progression.Additionally, the intranetwork connectivity within the ventral attention network was positively correlated with the healthy side vestibular response, indicating that the functional integrity of this network may be associated with the compensatory mechanisms observed in some MD patients.
It is important to acknowledge several limitations of our study.First, the cross-sectional design precludes us from drawing definitive conclusions about causality or the temporal dynamics of functional connectivity changes.Longitudinal studies would be valuable in elucidating the progressive nature of the observed alterations.Second, our sample size was relatively small, and future studies with larger cohorts are needed to validate these findings.Third, it is important to note that in this study, a 12-channel head coil was used for MRI acquisition, which may potentially lead to limitations in image spatial resolution and sensitivity.While a 12-channel head coil represents the minimum number of channels typically employed for head imaging, this choice was made based on the available equipment at our institution and its suitability for the study objectives.Finally, we acknowledge that comorbidities, medications, and other factors may have influenced our results, and further investigations controlling for these variables are warranted.In conclusion, our study provides evidence of functional reorganization in multiple brain networks in MD.These findings expand our understanding of the neural mechanisms underlying the disease and shed light on the functional implications of altered connectivity patterns.The identified connections and correlations with disease stage and vestibular response contribute to the potential development of biomarkers and therapeutic targets for MD.Future research should focus on longitudinal investigations with larger cohorts to validate these findings and explore the clinical implications of the observed alterations in functional connectivity.

Subjects
Fifty-five patients with MD and seventy healthy control (HC) subjects were recruited from Wuhan Union Hospital.Only adult participants aged 18 years and older were included in the study.The diagnosis of MD was established following the diagnostic guidelines of MD outlined by the Bárány Society 33 .Exclusion criteria included: (1) patients fulfilling both diagnostic criteria of vestibular migraine and MD; (2) middle or inner ear infections (otitis media, mastoiditis, labyrinthitis etc.); (3) middle or inner ear anomaly (common cavity malformation, semicircular canal dysplasia, enlarged vestibular aqueduct, etc.); (4) bilateral MD; (5) having received previous ear surgery or intratympanic injections; (6) retro-cochlear lesions (vestibular schwannoma, internal acoustic canal stenosis etc.); (7) head trauma.Inclusion criteria for HCs encompassed the following aspects: (1) the absence of any known neurological or psychiatric disorders; (2) no history of hearing impairment or vestibular disorders; (3) normal audiometric findings, including pure-tone audiometry and tympanometry within specified parameters; (4) no contraindications for MRI scans.Exclusion criteria for HCs were defined as the presence of any abnormal structural manifestations, such as stroke or brain tumor, on magnetic resonance images, as well as the presence of white matter lesions exceeding 1 mm in size, visible on structural MRI.
This study was conducted in compliance with the latest version of the Declaration of Helsinki and was approved by the Medical Ethics Committee of Tongji Medical College of Huazhong University of Science and Technology (2020IEC-J330).All participants provided written informed consent after receiving a detailed explanation of the study's purpose and procedures.

Audio-vestibular evaluations
The detailed procedures of audio-vestibular evaluations have been described in our previous studies 34 , which included the pure tone audiometry, electrocochleogram (ECochG), glycerol test, caloric test.The audio-vestibular tests were performed within an interval of 3 days.
Pure-tone audiometry was conducted in a soundproof room, at 0.125, 0.25, 0. Summating potential (SP) and action potential (AP) were recorded during ECochG.A positive ECochG result was defined as SP/ AP ratio greater than 0.4, which indicated the presence of EH.In glycerol test, two patterns of pathological results were noted, i.e., glycerol-induced hearing gain (improvement of hearing ability after glycerol intake) and rebound phenomena (deterioration of hearing ability after glycerol intake).The result of audiometric glycerol test was deemed positive when the pure-tone hearing ability was improved by: (1) ≥ 10 dB at any three or more frequencies or (2) ≥ 15 dB at one frequency at any time point after glycerol ingestion.
Bithermal caloric response was measured using infrared videonystagmography (Visual Eyes VNG, Micromedical Technologies, Chatham, IL).Patients lay supine with their upper trunk elevated at 30°.Cold (24 °C) and warm (50 °C) air stimulation was delivered into each external auditory canal alternately, and the maximum slow phase velocity (SPV max ) of the caloric nystagmus was measured after each irrigation.The value of canal paresis (CP) was calculated using the Jongkees formula.Interaural asymmetry of the caloric nystagmus ≥ 25% was taken as evidence of unilateral vestibular hypofunction.Bilateral vestibular hypofunction is considered if SPV max of each ear was less than 6°/s after caloric stimulation, or the summated SPVmax was less than 20°/s for all four stimulation conditions.

Magnetic resonance imaging acquisition
All participants were scanned on a 3.0 Tesla MRI scanner (MAGNETOM trio; Siemens Healthcare, Erlangen, Germany) using a 12-channel head coil.All subjects were instructed to lie supine and keep still, with headphones covering their ears to reduce scanner noise.Head motion was minimized using a foam cushion.Anatomical images were acquired with a 3D high-resolution T1-weighted magnetization-prepared rapid acquisition gradient echo (MP-RAGE) sequence, with the following parameters: repetition time (TR) = 2250 ms, echo time (TE) = 2.26 ms, inversion time (TI) = 900 ms, flip angle = 9°, voxel size = 1.0 × 1.0 × 1.0 mm 3 , field of view (FOV) = 256 mm × 256 mm, slice thickness = 1.00 mm, and 176 sagittal slices covering the entire brain.Functional images were acquired using a gradient echo type echo planar imaging (EPI) sequence, with TR = 2000 ms, TE = 30 ms, flip angle = 90°, voxel size = 3 × 3 × 3 mm 3 , and FOV = 200 mm × 200 mm, resulting in 240 functional images.A T2-weighted sequence was also acquired to evaluate the peripheral auditory system, with the following parameters: TR = 1000 ms, TE = 132 ms, slice thickness = 0.5 mm, slice number = 64, flip angle = 120°, FOV = 200 mm × 200 mm, and averages = 2. Two radiologists independently reviewed the MR images (1 with 10 years of experience and one with more than 20 years of experience).Subjects with MR abnormalities, such as otitis media, acoustic neuroma, and brain tumors, were excluded from the analysis.Intratympanic Gd injection and MRI examination was performed as previously described 34 .The Gd-DTPA-dimeglumine solution (Mul-tiHance; Braccosine, Shanghai, China) was used as the contrast agent.Delayed 3D-FLAIR MRI was used to evaluated the degree of EH.According to the published criteria 35 , the degree of EH in the vestibule and cochlea was classified into 3 grades: none, mild, and significant (Table 2).The vestibular hydrops ratio (VR) represents the proportion of the endolymphatic space area in relation to the entire lymphatic space in the vestibule.It was calculated using the following equation: VR = (number of negative pixels for the endolymph in the region of interest (ROI)/total number of pixels in the ROI).

Magnetic resonance imaging data processing
Functional MRI (fMRI) images were preprocessed using the Data Processing and Analysis for Brain Imaging (DPABI) software (http:// rfmri.org/ dpabi) toolbox 36 which is based on the Statistical Parametric Mapping (SPM12) toolkits (http:// www.fil.ion.ucl.ac.uk/ spm/ softw are/ spm12/) on the Matlab (R2017a) platform.The first 10 volumes of the resting-state blood oxygen level-dependent (BOLD) data were discarded to account for T1 equilibration effects, leaving 230 time points for analysis.Slice timing and motion correction were performed on the remaining volumes using a six-parameter rigid-body transformation.Nuisance signals, including linear trends, white matter signal, cerebrospinal fluid signal, and Friston 24 head motion parameters, were regressed out from the functional signal.The functional images were then normalized to the Montreal Neurological Institute (MNI) template using the Diffeomorphic Anatomical Registration Through Exponentiated Lie algebra (DARTEL) tool, after being coregistered with the corresponding structural images, which were also normalized to the MNI template.Finally, the images were smoothed with a 4 mm full width at half-maximum Gaussian kernel and band-pass filtered with a frequency range of 0.01-0.1 Hz.Participants with a maximum head motion greater than 1.5 mm in displacement or 1.5° in rotation, as well as those with a mean frame-wise displacement (FD) calculated by the Jenkinson method greater than 0.2, were excluded from the study.

Functional connectivity analyses
To examine the functional connectivity patterns of the various brain networks, we used the 90 Automated Anatomical Labeling (AAL) regions 37 , which were clustered into seven networks, including visual (VSN), somatomotor (SMN), dorsal attention (DAN), ventral attention (VAN), limbic (LBN), frontoparietal (FPN), and default mode (DMN) networks, based on previous studies 26,38,39 .Furthermore, the bilateral caudate, putamen, pallidum, and thalamus were identified and clustered as the deep gray matter network (DGN) 26,39,40 .As the auditory network (AUN) plays a unique role in MD disease, we separated it from the SMN network for subsequent analysis.Thus, in our subsequent analysis, we investigated the patterns of functional connectivity among these nine brain subnetworks.
To quantify the functional connectivity between brain regions, we calculated Pearson's correlation coefficients (r values) between the mean times of each pair of regions, resulting in a 90 × 90 correlation matrix for each subject.We then converted the correlation coefficients to nodal connectivity degree η using an exponential function, as described in previous studies 26,27 .We assessed intranetwork and internetwork connections by averaging the transformed correlation coefficients of all ROI pairs within or between each particular brain network 26,27,39,40 .
We performed functional connectivity analyses at the integrity, network, and edge levels to study the shared general and distinct specific connectivity patterns between the MD and HC groups 26,27,39 .At the integrity level, we investigated the information flow received from specific nodes, which is equivalent to the "degree centrality" in graph theory 27 .At the network level, we performed intranetwork and internetwork analysis.Intranetwork analysis measured connectivity by averaging the transformed correlation coefficients within a network, while internetwork analysis measured the connectivity between two networks 41 .At the edge level, we explored the functional connectivity between all possible node pairs.This approach allowed us to accurately study connectivity between nodes and reveal focal phenomena obscured by the integration of connected brain regions as networks 41 .

Statistical analysis
For the demographic data, data processing and statistical analysis were conducted using SPSS 17.0 software (SPSS Inc., Chicago, IL, USA).Continuous variables were reported as medians and interquartile intervals, while categorical variables were presented as frequencies and proportions.
To test for group differences in functional connectivity, we conducted independent samples t-tests on the Z scores for each level (integrity, network, and edge) between the MD and HC groups, with age, gender, and education level included as covariates in a general linear model.Statistical significance was set at a corrected p-value of < 0.05, using a permutation-based method.In addition, we performed multiple regression analyses to examine the relationship between functional connectivity and the clinical indices in MD patients, including parameters such as MD duration, MD stages, among others, while controlling for covariates such as age, gender, and education level.

Figure 1 .
Figure 1.Altered node degree in patients with MD at the integrity level.(A) MD presented significantly (p < 0.05, FDR corrected) lower degree compared to HC in 19 brain regions (displayed on the brain surface).(B) The histogram shows the comparison between MD and HC of network degree.VSN visual network, SMN somatomotor network, AUN auditory network, VAN ventral attention network, DAN dorsal attention network, FPN frontoparietal network, DMN default mode network, DGN, deep gray matter network, LBN, limbic network, MD, Ménière's disease, HC, healthy control, *indicates p < 0.05 after FDR correction.

Figure 2 .
Figure 2. Altered functional connectivity among nine brain networks at intranetwork and internetwork level between MD and HC.(A) A colorful circle indicates the composite of nine brain networks.The red triangles indicate that the AUN, VAN and LBN showed altered intranetwork connections.The line linking two brain networks indicates altered internetwork connections.(B) The histogram shows the comparison between MD and HC of intranetwork connections of the nine brain networks.(C) The histogram shows the comparison between MD and HC of internetwork connections.VSN visual network, SMN somatomotor network, AUN auditory network, VAN ventral attention network, DAN dorsal attention network, FPN frontoparietal network, DMN default mode network, DGN, deep gray matter network, LBN, limbic network, MD, Ménière's disease, HC, healthy control.

Figure 3 .
Figure 3. Altered functional connectivity of nodes at the edge level between MD and HC.(A) Brain surface rendering displaying altered functional connectivity (p < 0.05, FDR corrected) between MD and HC.The same color of spheres represented areas were from the same brain network.The thickness of the lines corresponds to the T values of the connectivity strength.(B) Circular representation illustrating altered functional connectivity patterns between MD and HC.VSN visual network, SMN somatomotor network, AUN auditory network, VAN ventral attention network, DAN dorsal attention network, FPN frontoparietal network, DMN default mode network, DGN deep gray matter network, LBN, limbic network.

Figure 4 .
Figure 4. Correlations between altered brain network index and clinical information.(A) Correlations between the stage of MD and the degree of the left amygdala.(B) Correlations between the VR scores on the healthy side and the intranetwork connections of the VAN.MD Ménière's disease, VAN ventral attention network, HVR vestibular hydrops ratio of healthy side.
5, 1.0, 2.0, 4.0 and 8.0 kHz.The clinical stage of MD was determined based on three-frequency pure tone average (PTA), calculated as simple arithmetic means of 0.5, 1.0, and 2.0 kHz pure tone threshold.PTA of < 26 dB hearing loss (HL) was classified as Stage I; PTA of 26-40 dB HL as Stage II; PTA of 41-70 dB HL as Stage III; and PTA of > 70 dB HL as Stage IV.

Table 1 .
Demographic and clinical variables.A Pearson chi-squared test was used for gender comparison.independent samples t-tests were used for age, education comparisons.MD Ménière's disease, HC healthy control, PTA pure tone average, SP Summating potential, AP, action potential, VR: vestibular hydrops ratio.

Table 2 .
Grading of EH using magnetic resonance imaging.EH endolymphatic hydrops.