Patients and questionnaires
Among patients (aged 19–80 years) who visited our hospital with a chief complaint of tinnitus, we selected those who had chronic non-pulsatile tinnitus for 3 months or longer, were capable of smooth communication in Korean, and who agreed to participate in the study. We excluded those patients with frequent noise exposure as a part of their professional or hobby activities and those who were expected to have difficulty operating a VR program based on a head-mounted display (HMD). Finally, 19 patients (9 male,10 female) with tinnitus participated in this study.
The medical history, demographic information, medical examination, and physical examination, including vital signs (blood pressure, heart rate, body temperature, and breathing rate), weight, height, and daily life competency of the participants were evaluated accordingly. Moreover, after confirming the treatment history and underlying diseases related to tinnitus except for other otological, neurological, and psychological problems before the study, audiograms and tinnitus symptoms were also examined for these patients before and after the experiment to exclude unnecessary bias-related cases.
The pre- and post-experimental patient statuses were evaluated through questionnaires on tinnitus itself, including the Tinnitus Handicap Inventory (THI), Tinnitus Handicap Questionnaire (THQ), and visual numeric scale (VNS) related to the severity of tinnitus. Questionnaires about symptoms associated with tinnitus, such as the Pittsburgh Sleep Quality Index (PSQI), WHO-QoL (Quality of Life Assessment), Profile of Mood States (POMS), Hospital Anxiety and Depression Scale (HADS) for depression, anxiety, and sleep disorders accompanying tinnitus, respectively. Further, the simulator sickness questionnaire (SSQ) was administered to evaluate the post-experimental patient symptoms that might occur after usage of the VR system.
Experimental protocol
First, a tinnitus avatar was created to mimic the subjective tinnitus of each patient by matching the frequency and loudness. This indicates that a fusion process can occur between subjective tinnitus and the matched stimulus presented in the contralateral ear19,20. A tutorial session was conducted before the main treatment session. The tinnitus avatar was designed to produce spatial tinnitus sounds that were implemented with a head-related transfer function (HRTF) of the Google resonance sound software development kit (SDK). The participants learnt how to move in a VR setting and perform tasks to dispose of the tinnitus avatar (Figure 1). Using HRTF, the tinnitus avatar produced 3D tinnitus sounds of five types (whistling, hissing, roaring, humming, and ringing sound). The participants were required to use their hearing to locate the tinnitus avatar. The participants could move in the virtual environment by pushing the upper and lower parts of the trackpad on the left ViveTM controller. The participants were able to see the tinnitus avatar once they had approached it within a certain distance. When the tinnitus avatar was held by pressing the trigger button on the right ViveTM controller, the color of the avatar changed and vibration feedback was generated on the right ViveTM controller, informing the participants that they had caught the tinnitus avatar. The participants then moved the tinnitus avatar to the tinnitus disposal site in the virtual environment scene (Figure 2). When the tinnitus avatar was discarded, the tinnitus sound rapidly diminished. For the tinnitus VR treatment sessions, two sets of virtual environments were developed for each city street, restaurant, living room, and bedroom scene. The tinnitus disposal sites were placed in the noisy scenes in each virtual set (i.e., at the entrance to the bedroom, beside a television in the living room, at the ordering counter in a restaurant, and in a car hood in a city street), creating the cognitive illusion of absorbing tinnitus sounds into much louder environmental noises. The participants performed three rounds of tinnitus avatar disposal tasks during each treatment session.
The patients visited the hospital four times and experienced VR tinnitus treatment every 1–2 weeks (Figure 4). Participants who agreed to participate in the experiment underwent an endoscopic eardrum examination and a hearing/tinnitus examination, followed by questionnaires including the THI, PSQI, WHO-QoL, POMS, and HADS at their initial visit. Participants experienced the VR tinnitus treatment system from the first to the third visit, for a total of three sessions. During the first visit, the participants underwent EEG tests (the detailed EEG recording method is described in the supplementary information) to determine the status of their brain waves before experiencing the VR tinnitus treatment system. The first treatment session (city street) was followed by a tutorial session on the VR tinnitus treatment system. Participants performed both the second (restaurant scene) and third (living room scene) treatment sessions on their second visit; they performed the fourth treatment session (bedroom scene) on their third visit. The treatment sessions were arranged to decrease the volume of environmental noise with progression of the sequences (Figure 3). After the fourth treatment session, the participants received an EEG test to determine the changes in the brainwaves and answered questionnaires including THI, VNS, PSQI, WHO-QoL, POMS, HADS, and SSQ. Our experimental protocol is summarized in the Supplementary Video.
Source localization using sLORETA
We used the sLORETA software, which allows the analysis of intracerebral electrical sources from scalp-recorded activity based on EEG data. We preprocessed the EEG data and prepared for 30 epochs per participant, as mentioned in the Supplementary section. These epochs were analyzed for six frequency bands (delta, 1–4 Hz; theta, 4–8 Hz; alpha, 8–12 Hz; low beta, 12–18 Hz; high beta, 18–30 Hz; and gamma, 30–55 Hz). In sLORETA, the source images were spatially modeled as a collection of 6239 voxels (size 5×5×5 mm). These layers were obtained from the amygdala, hippocampus, and cortical gray matter. The sLORETA data were based on the digitized Montreal Neurological Institute (MNI) 152 coordinates reformed to Talairach coordinates21,22.
We selected 10 ROIs from prefrontal regions based on the previous literature on tinnitus3,23-25. Each ROI comprised one or two Brodmann areas (Table 1). We also compared the current source density for each ROI using SPSS.
Outcomes
The primary outcome measure was the questionnaire results of pre- and post-experimental patient status and symptoms associated with tinnitus. We used the THI, THQ, and VNS for patient status and the PSQI, WHO-QoL, POMS, and HADS scores for symptoms. The secondary outcome measure was the current source density for 10 ROIs3,23-25. Additionally, another measure was the questionnaire results regarding VR sickness using the SSQ.
Statistical analysis
Descriptive statistics were used to analyze the outcomes. Improvements in the THI score and grade of tinnitus were reported using the change in difference with respect to measurements before treatment. The pre-treatment and post-treatment differences were reported using absolute values. The statistical significance of the differences found in the questionnaire was evaluated using the Wilcoxon signed-rank test. Absolute values before and after treatment were paired for each participant, and statistical significance was considered at p less than 0.05. All data were analyzed using SPSS (version 20.0; SPSS, Chicago, Illinois, USA). Furthermore, we performed Wilcoxon signed-rank test between EEG data of before-treatment and after-treatment and Mann-Whitney U test in subgroups separated by THI score regarding differences in EEG data between before-treatment and after-treatment on ROI analysis (THI<0: n=12 , THI≥0: n=7)21,22. We set a threshold of P<0.01.
Ethical approval
The trial was registered at https://cris.nih.go.kr under the registry number KCT0006242 (the full date of first registration, 10/06/2021). The study was approved by the ethics board of the Korea University Ansan Hospital (No. 2020AS0010), and all patients provided written and informed consent. All methods were performed in accordance with relevant guidelines and regulations.