Analysis of Changes in Brain Region and Connectomics in the Acute Stage of Sudden Sensorineural Hearing Loss in the Resting State via Functional Magnetic Resonance Imaging

,


Introduction
When the lesions are in the hair cells, auditory nerves, or hearing centers at all levels of the helical organ and the sensation of sound and conduction of nerve impulses are impaired, resulting in hearing loss, it is called sensorineural hearing loss (SHL) [1][2][3]. Hearing loss caused by hair cell disease is called SHL (cochlear or terminal hearing loss). Hearing loss caused by lesions located in the auditory nerve and its conduction pathway is called neural hearing loss (retrocochlear or postcochlear) and is characterized by a marked decline in speech discrimination. Lesions occurring in the auditory center of the cerebral cortex are called central deafness and are often accompanied by neurological symptoms [4,5]. SHL patients mainly have symptoms such as high-frequency hearing loss, reinvigorating phenomenon, tinnitus, and vertigo. Tis kind of deafness cannot change sound waves into nerve signals, or nerves, and their central way of producing obstacles cannot transmit nerve signals to the brain, which is the most difcult to treat and the most troublesome form of deafness. It is usually caused by congenital or acquired causes of cochlear auditory nerve or auditory center lesions, resulting in deafness [6][7][8]. According to current clinical and experimental studies, hearing loss often caused by SHL is irreversible and cannot be cured by drugs or surgery, which has been fully confrmed by massive basic and clinical studies and experiments over the years [9,10]. Although the data showed that a small number of people's sudden deafness can be cured automatically, for the occurrence of sudden deafness, it is still recommended for timely hospital diagnosis.
Common clinical diagnosis methods for deafness include vestibular function tests, hearing examinations, X-rays, computed tomography (CT), and MRI [11][12][13][14]. Patients with vertigo can undergo a cold temperature test and an electronystagmus examination to determine vestibular damage and exclude central dysfunction after the condition becomes stable [15]. Pure tone audiometry can show typical SHL, and the hearing curve of moderate, severe, or more severe hearing loss patients is usually fat or descending. Te pure tone supramural function test, the Bekesy test, speech audiometry, and acoustic immittance test can determine cochlear deafness. X-ray or CT scans of the inner ear canal can rule out sudden deafness caused by external neuroma or other intracranial occupying lesions [16]. However, the above methods can only make a preliminary judgment on the lesion degree of patients but cannot evaluate the changes in cerebral cortical functional activities after deafness [17]. Functional magnetic resonance imaging (fMRI) is a new neuroimaging method that uses magnetic resonance imaging to measure the hemodynamic changes caused by neuronal activity. It has been widely used to map brain activation in humans and animals. Te change of local deoxyhemoglobin concentration in the brain will lead to the change of magnetic resonance signal intensity, which indirectly refects neuronal activity and then refects the change of brain function in real time [18]. In recent years, fMRI has become the tool of choice for visualizing neural activity in the human brain due to its noninvasive nature, reproducibility, and lack of radiation exposure. In resting-state fMRI, subjects are scanned while lying fat, eyes closed, awake, and not thinking about a problem without a specifc task. Resting-state fMRI is very suitable for the study of clinical diseases, especially neurological diseases because of its simple experimental design, simple operation, and ability to observe the spontaneous brain activity of subjects without subjects completing specifc tasks [19]. Te brain is one of the major parts of the central nervous system. It can transmit, store, and process information, make people do various psychological activities, and dominate and control people's behavior. Although they are composed of various brain regions that are independent of each other in space, these brain regions are functionally interconnected and constantly fow information with each other [20]. Terefore, it is signifcant to explore sudden deafness from the perspective of brain functional network changes.
In summary, this study aimed to analyze the functional parameters and core nodes of the whole brain network based on resting-state fMRI by conducting fMRI scanning and constructing the brain functional networks of 80 patients with severe right sensorineural hearing loss and 40 healthy volunteers. It was hoped to provide a reference for the study of the changes in the functional connectivity level of the whole brain network in patients with sudden sensorineural hearing loss in the acute stage.

Research Objects.
Eighty patients with severe right SHL who visited the hospital from October 2019 to October 2021 were selected as study subjects (patient group), including 34 males and 46 females. Te average age was 35.36 ± 10.34 years (range, 25-67 years). After pure tone audiometry, the hearing threshold of the afected ear of the patient group was more than 70 dB, and the hearing threshold of the healthy ear was in the normal range. In addition, 40 healthy volunteers who underwent physical examination in the hospital during the same period were selected as a control group, including 15 males and 25 females. Te age ranged from 24 to 65 years, with an average age of 37.11 ± 8.95 years. After pure-tone audiometry, the hearing threshold of both ears was less than 25 dB. Tis study was approved by the ethics committee of the hospital. Te subjects and their families were informed of the study and signed informed consent forms.
Inclusion criteria are as follows: (i) all subjects were right-handed; (ii) subjects had complete basic information; (iii) subjects had no previous history of mental illness; and (iv) subjects were older than 18 years.
Exclusion criteria are as follows: (i) subjects with contraindications to MRI examination; (ii) subjects with history of cardiovascular disease; and (iii) subjects complicated with serious damage to the heart, liver, and kidney organs.

Resting fMRI Examination.
A 3.0 T superconducting magnetic resonance imaging system with 12 channel coils was used. First, the subjects were placed in a supine position and put on an eye mask and soundproof earplugs to avoid interference from the outside environment. Te subjects were guided to remain in a resting state. After the scan, if the patient was found not to remain in a resting state during the scan, the sample would be removed. A total of 300 time points were collected by resting whole brain volume scanning for each subject. Te scanning parameters were as follows: repeat time (TR) � 1.800 ms, echo time (TE) � 20 ms, layer thickness � 3.5 ms, layer spacing � 0, layer number � 65, feld of vision � 256 × 256, matrix � 256 × 256, and fip angle � 90 degrees.

Image Preprocessing.
Te data processing assistant for resting-state fMRI (DPARSF) based on MATLAB R2014b-2020b was used to preprocess the obtained images. Te volume data of the frst 10 points of each subject were removed. Time correction and head correction were carried out for the remaining 290 time points. Sample data of head movement exceeding 2 mm and rotation exceeding 2 degrees were removed. Linear regression was adopted to remove white matter and cerebrospinal fuid signals from the image data, and then it was standardized to the spatial coordinates of the Montreal Neurological Institute for resampling. Spatial smoothing was performed on standardized data, and time-domain bandpass fltering with a bandwidth of approximately 0.05 Hz was used to reduce the infuence of low-frequency drift and high-frequency physiological noise.

Construction of the Brain Functional Network.
GRETNA, developed by Professor Yong He's team from Beijing Normal University, was used to construct the brain functional networks of subjects. Te subjects' brains were registered on an automatic anatomical marker map, which divided the cerebral cortex into 90 regions. Each region represented a node, and the time correlation of the blood oxygen level-dependent signal in each brain region was called an edge. Te average time series of all voxels in each brain region was extracted, and the Pearson correlation coefcient between them was calculated to obtain the correlation matrix R. Ten, the Fisher Z transformation was performed on the matrix R to obtain the correlation matrix Z, which was converted to the correlation matrix C for parameter calculation.

Brain Functional Network Parameters.
Given that the network node set is U and the network edge set is I, the complex network can be represented by the image F(U and I). Ten, the node clustering coefcient C x is expressed as follows: L x represents the number of nodes constituting subgraph F x , L y represents the number of nodes constituting subgraph F y , and E represents local efciency. Te average value of C x of all nodes is the network clustering coefcient C p , which can be used to evaluate the degree of network collectivization. Te average value of E x of all nodes is the network local efciency E local , which can be used to evaluate the network defense capability. Te average maximum shortcircuit length L p is the average time l xy of all nodes to ensure efective integration and fast transmission of information in distant brain regions. Global efciency E global evaluates the global transport capacity of a network.
Te standardized coefcient c refects the transformation from the real brain network to the regular network, and the standardized average shortest path length λ refects the transformation from the real brain network to the random network, which can be expressed as follows: (2) C rand represents the clustering coefcient of the random network, L rand represents the shortest path length of the random network, and σ > 1 indicates that the network has a high clustering coefcient and a short characteristic path length.
Te intermediate centricity method was used to fnd the core nodes of the whole brain network. In this experiment, the threshold range was 0.1-0.2 with an interval of 0.02.

Statistical
Methods. SPSS 19.0 statistical software was used for data processing in this study. Te mean ± standard deviation (x ± s) was used for measurement data, and the percentage (%) was used for counting data. Pairwise comparison was performed by one-way ANOVA. P < 0.05 suggests that the diference was statistically signifcant. Table 1, the ratio of male to female, age, course of disease, height, weight, years of education, THI score, and left ear PTA score in the patient group were not signifcantly diferent from those in the normal group (P > 0.05). Te PTA score of the patient group (85.71 ± 9.01) was signifcantly higher than that of the normal group (20.77 ± 3.81), and the diference was statistically signifcant (P < 0.05). Figure 1 shows the comparison of the fMRI parameters E global , E local , C p , and L p between the two groups. Te fMRI parameters E global , E local , and C p of the two groups showed an increasing trend with an increasing threshold, while L p showed a decreasing trend with an increasing threshold. At all thresholds, the fMRI parameters, C p and L p, in patients were dramatically superior to those in the control group, and the diferences were considerable (P < 0.05). At most of the thresholds, the fMRI parameter E global of the patient group was substantially lower than that of the control group, and the diferences were considerable (P < 0.05). At all thresholds, there was no considerable diference in the fMRI parameter E local between the patient group and the control group (P > 0.05). Figure 2 shows the comparison of fMRI parameters c, λ, and σ between the two groups. Te fMRI parameters c, λ and σ of the two groups showed a decreasing trend with an increasing threshold value. At most of the thresholds, the fMRI parameter λ of the patient group was dramatically superior to that of the control group, and the diferences were considerable (P < 0.05). At all thresholds, fMRI parameters c and σ in the patient group were not substantially diferent from those in the control group (P > 0.05). Figure 3 shows MRI images of one patient in the patient group. Physical examination suggested a clear mind, vague language, decreased vision, no nystagmus, deafness in both ears, muscle strength level V, bilateral calcaneus shin test (+), and acceptable fnger and nose rotation. Obviously, the surface of the midbrain, pons, vermis of the cerebellum and sulci, bilateral lateral fssure, underside of the frontal lobe, and spinal cord were surrounded by short T2 signals with bands. Figure 4 shows the MRI data of a healthy volunteer in the control group. Te T1 weighted image presented a high signal, and the T2 weighted image presented an equal signal. Te position of the brain stem and cerebellum was normal without displacement or deformation, and the surface of the midbrain, pons, vermis, and sulci was normal.

Analysis Results of Functional Core Nodes of the Whole
Brain Network. In Figure 5, this analysis showed that there were 10 specifc network core nodes in the patient group, including the right parahippocampal gyrus, right suproccipital gyrus, left suproccipital gyrus, right fusiform gyrus, right parietal lobule, right subparietal lobule, right superior temporal gyrus, left superior marginal gyrus, and right superior temporal gyrus temporal pole. Tere were eight specifc network core nodes in the control group, including the left inferior frontal gyrus orbital, right insula, left insula, left parietal lobule,, right pea shell, left superior temporal gyrus temporal pole, and right middle temporal gyrus temporal pole. Subjects in both groups had a total of fve network core nodes, including the right inferior frontal gyrus orbital region, left posterior central gyrus, right superior marginal gyrus, left middle cingulate gyrus, and right superior frontal gyrus medial orbital region.

Discussion
Sudden sensorineural deafness refers to the sudden occurrence of SHL of unknown cause, which can develop from mild hearing loss to severe hearing loss within minutes, hours, or 1-2 days [21]. Tere is no specifc diagnosis and treatment method for approximately 90% of cases of sudden sensorineural deafness because the cause is unclear. In addition, the lack of high-quality clinical research and evidence-based medical proof makes the diagnosis and treatment of sudden inductive neural deafness controversial. Terefore, seeking an efcient diagnostic method is a reliable way to achieve clinical treatment optimization [22]. In this study, 80 patients with severe right SHL who visited the hospital were selected as the study subjects (patient group), and 40 healthy volunteers who underwent physical examination in the hospital during the same period were selected as the control group. FMRI imaging was performed on all patients. GRETNA was used to construct the subjects' brain functional networks, and the functional parameters and core nodes of the whole brain network were analyzed. First, the basic data of the two groups of subjects were compared. Te results showed that the ratio of males to females, age, course of disease, height, weight, years of education, THI score, and left ear PTA score in the patient group were not signifcantly diferent from those in the control group (P > 0.05), which provided a baseline for subsequent studies. In addition, the PTA score of the patient group (85.71 ± 9.01) was dramatically superior to that of the control group (20.77 ± 3.81), and the diference was signifcant (P < 0.05), which was consistent with the control study between the patient group and control group.
Te functional network of the human brain is a complex and real network between a random network and a regular network, and its small-world attributes can ensure that the brain can efciently integrate and separate information under relatively low energy consumption [23]. Tis study found that at all thresholds, the fMRI parameters C p and L p of the patient group were dramatically superior to those of the control group. At most of the thresholds, the fMRI parameter E global of the patient group was substantially lower than that of the control group, and the diferences were considerable (P < 0.05). Tis is generally consistent with the research results of Minosse et al. [24]. Although the whole brain of patients with sudden sensorineural hearing loss still had the small-world property, some characteristics of the brain network had changed. Te increase of parameters meant that the information processing ability of some brain regions of patients was enhanced, indicating that the contralateral superiority of hearing conduction was weakened, and intrachannel reorganization occurred in the auditory cortex. Te increase in parameters L p and E local suggested that the information processing ability of the patient's remote brain region was weakened, which was consistent with the general deletion hypothesis [25]. At most of the thresholds, the fMRI parameter λ of the patient group was dramatically superior to that of the control group, and the diferences were considerable (P < 0.05). At all thresholds, fMRI parameters c and σ in the patient group were not substantially diferent from those in the control group (P > 0.05). Wang et al. [26] also pointed out that this indicates that the brain functional network of patients with sudden sensorineural hearing loss has a trend of changing to a regular network, with enhanced information integration ability and weakened separation transmission ability. Te results indicate that the brain functional network of patients  Figure 1: Comparison of the fMRI parameters E global , E local , C p , and L p between the two groups. (a-d) represented E global , E local , C p , and L p , respectively. * represents that the fMRI parameters of the patient group, including E global , C p , and L p, were signifcantly diferent from those in the normal group (P < 0.05).  with sudden SHL tended to change to a regular network, and the ability of information integration was enhanced, while the ability of separation and transmission was weakened. All the above results confrm to some extent that the connection between auditory brain areas and language and vision-related functional areas was weakened in patients with burst SHL, and their language perception and understanding ability were decreased [27].
Tere were ten specifc network core nodes in patients, including the right parahippocampal gyrus, right supraoccipital gyrus, left suboccipital gyrus, right fusiform gyrus, right parietal lobule, right subparietal lobule, right superior temporal gyrus, and left superior marginal gyrus. Among them, the superior occipital gyrus, fusiform gyrus, superior marginal gyrus, temporal pole, inferior parietal lobule, and parietal lobule are related to visual and auditory information processing. Te increase in the importance of these core nodes indicates that other sensory channels are compensatory and that the advantage of coping is weakened after hearing loss. Tere were eight specifc network core nodes in the control group, including the left inferior frontal gyrus orbital, right insula, left insula, left parietal lobule, left parietal lobule, right pea shell, left superior temporal gyrus temporal pole, and right middle temporal gyrus temporal pole. Most of these nodes belong to the limbic or parlimbic system, but they were absent in the patient group, possibly because of the recombination of these nodes after deafness [28].

Conclusion
In this study, the functional parameters and core nodes of the whole brain network were analyzed by fMRI scanning and GRETNA software to construct the brain functional network of subjects. Te results showed that the brains of patients with sudden sensorineural deafness still had a smallworld nature. However, some characteristics of the brain network had changed, and there was a trend to transform to a regular network. Te connection between the auditory brain region and language and visual related functional regions was weakened, and the distribution of core nodes was changed, which enhanced the information integration ability of local brain regions and weakened the separation transmission ability. However, the limitation of this study is that the included sample size was small and all of them were from the same hospital, and the evaluation of patient samples was insufcient. Te visual and language function evaluation data of the patients were not obtained from the side, and the correlation between right deafness and diferent functions could not be verifed. Terefore, the sample size will be expanded in the later stage, and the patient case data will be increased to explore the diference in brain network function between left and right deafness patients. In conclusion, this study provides data reference for the study of brain network function in patients with sudden right sensorineural hearing loss.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te author declares that there are no conficts of interest.