Standard protocol approval, registration and patient consent
This research involved human participants. All authors have declared that this research was approved by the Institutional Review Board (IRB). This study was approved by the ethics committees of our research institution (Beijing Friendship Hospital, Capital Medical University, 2016-P2-012). Written informed consent was obtained from all study subjects.
Subjects
All patients and healthy volunteers were recruited in our institution. In this study, thirty-three patients with idiopathic tinnitus were enrolled. The tinnitus sound was described as a persistent, high-pitched sound in both ears. The inclusion criteria were as follows: (1) 18 to 65 years old; (2) right handedness; (3) tinnitus duration ranging from 6 to 48 months; (4) no significant hearing loss (hearing thresholds ≥ 25 dB HL at frequencies of 0.250, 0.500, 1, 2, 3, 4, 6, and 8 kHz determined by pure tone audiometry (PTA) examination; and (5) willingness to receive sound therapy for 24 weeks and then return to be re-examined after treatment. The exclusion criteria included (1) other kinds of tinnitus (such as pulsatile tinnitus), Meniere’s disease, sudden deafness, or otosclerosis; (2) neurological signs and/or a history of neurological disease; (3) current chronic medical illness; (4) a history of head trauma; (5) cardiovascular, pulmonary or systemic disease; (6) claustrophobia experienced during the MRI simulator session; and (7) subjects who could not pitch-match their tinnitus. Twenty-six age-, sex-, education-, and handedness-matched HC subjects were enrolled as controls. None of the HCs suffered from tinnitus in the past year. Other exclusion criteria were the same as previously described. The characteristics of the subjects are presented in Table 1.
Sound therapy and clinical evaluation
All of the enrolled tinnitus patients were examined for tinnitus loudness matching (L = loudness of tinnitus), pitch matching (Tf = tinnitus frequency), minimum masking level, and residual inhibition to characterize the tinnitus and prepare for treatment. Narrow band sound therapy was administered to participants in the tinnitus group for 24 weeks, three times a day for 20 minutes each time. For each tinnitus patient, the loudness of sound for treatment was set as L + 5 dB. The frequency was set as a 1 kHz narrow band when setting Tf as the middle point of the delivered sound (Tf ± 0.5 kHz; for example, Tf = 4 kHz, low sound cut = 3.5 kHz, high sound cut = 4.5 kHz).
We also asked the patients to fill out the Tinnitus Handicap Inventory (THI) to assess the severity of tinnitus before and after treatment. The primary outcome of this prospective study was the change in THI score after treatment. A reduction of at least 16 points in the THI score was considered effective treatment (41).27 The HC group was not given any kind of sound during the study.
Data acquisition and data preprocessing
For each patient, to evaluate the change in brain activity under treatment, structural MRI data were collected at baseline and at the end of therapy (24th week). The HC group was also scanned at baseline as well as at the 24th week. Images were obtained using a 3.0T MRI system (Prisma, Siemens, Erlangen, Germany) with a 64-channel phase-array head coil. All imaging studies were performed at the Medical Imaging Research Center of our hospital. Parallel imaging was employed for data acquisition. High-resolution three-dimensional (3-D) structural T1-weighted images were acquired using a 3-D magnetization-prepared rapid gradient echo (MP-RAGE) sequence with the following parameters: repetition time (TR) = 2530 ms, echo time (TE) = 2.98 ms, inversion time (TI) = 1100 ms, FA = 7◦, number of slices = 192, slice thickness = 1 mm, bandwidth = 240 Hz/Px, field of view (FOV) = 256 mm × 256 mm, and matrix = 256 × 256, resulting in an isotropic voxel size of 1 mm × 1 mm × 1 mm3.
Image preprocessing was performed with the VBM8 toolbox in the SPM8 software package (Statistical Parametric Mapping, Wellcome Department of Cognitive Neurology, London, UK) running in MATLAB (MathWorks, Natick, MA, USA). The procedures for image preprocessing have been described in detail in our previous research (28). Briefly, image processing in this work included spatial normalization using the Montreal Neurological Institute (MNI) 152 template and segmentation of the GM, WM, and cerebrospinal fluid (CSF). Only the GM images were analyzed in this study. Modulation was also applied to avoid volumetric deformation of the GM due to stretching and shrinking effects during the normalization procedure. The modulated GM images were smoothed with a 6 mm full width at half maximum (FWHM) isotropic Gaussian kernel. Finally, the smoothed GM images were resampled to a 3 mm × 3 mm × 3 mm voxel size for statistical analysis. Subjects with excessive head motion (more than 1.5 mm in translation or 1.5° in rotation) were excluded from analysis.
Construction of SCNs
First, VBM8 software was used for structural image segmentation. We put GM volume maps of patients before and after treatment in one folder and maps from HCs in another folder. The GM volume value was extracted using DIPABI software (http://rfmri.org/dpabi). Second, the automated anatomical labeling (AAL) atlas was used to divide the whole brain into 90 cortical and subcortical regions of interest (42), and each was considered a network node. Last, we used the Brain Connectivity Toolbox software (43) to construct SCNs in MATLAB version R2017a.
SCN analysis
We used binarized graphs to calculate global properties and local properties and calculate the area under the curve (AUC) for each property over the sparsity range. The global property was defined as the average inverse of the characteristic path length (44). The global properties included the clustering coefficient, global efficiency, small-world properties, and shortest path length. Local (nodal) structural alterations were evaluated based on the local efficiency, degree centrality and betweenness centrality of each region (45). Nodal efficiency measured the global efficiency of parallel information transfer in a network. Degree centrality is the number of nodes directly connected to the node, which measures the importance of a single node in the network. Betweenness centrality examines the contribution of each node to the shortest path between all other pairs of points (46). The degree centrality and betweenness centrality of nodes reflect the importance of nodes in information transfer.
Statistical analysis
Demographic data were compared through two-sample t tests and paired two-sample t tests using SPSS 19.0 software (SPSS, Inc., Chicago, IL). P values < 0.05 were considered statistically significant. Longitudinal changes in the THI score were also analyzed by using paired two-sample t tests.
For VBM data, to determine the group × time interaction effect between the two groups and the two scans, the main effects of group (the tinnitus patient group and the HC group) and time (baseline and 24-week follow-up period), two-way mixed model analysis of variance (ANOVA) and post hoc analyses were performed. An F value of VBM analysis and a P value of SCN analysis less than 0.05 were considered statistically significant (false discovery rate (FDR) corrected). In post hoc analyses, two-sample t tests and paired two-sample t tests were used to compare GM volumetric changes. To prove our hypothesis, Pearson’s correlation analyses were further conducted to investigate the relationship between the change in GM volume and the clinical characteristics of tinnitus patients (disease duration at baseline, △THI score (△THI score = THIpre – THIpost)). P < 0.05 was set as the threshold to determine significance. The GM volume results were visualized with BrainNet Viewer (http://www.nitrc.org/projects/bnv/) (47). Pearson’s correlation analysis was performed using SPSS 19 software (SPSS, Inc., Chicago, IL) between the THI scores.
For SCNs, we used two-sample t tests to construct a network of patients with tinnitus (baseline and after 24-week treatment) and a healthy group (baseline and after 24-week scan). Then, the graph theory index of the covariant brain network was calculated, and a permutation test (1000 times) was performed on the graph theory index of the two groups. We applied an FDR of 5% to adjust for the multiple comparisons of mean local and global efficiency across the between-group contrasts before and after 24 weeks of treatment.