Investigating Disinhibitory Brain Mechanism in Tinnitus and Hearing Loss

SELECTION AND WITHDRAWAL OF PARTICIPANTS

Recruitment

1. Clinical routes Suitable candidates may be identified by the sites, at Ropewalk House Nottingham Audiology Services and ear, nose, and throat (ENT) services at Nottingham University Hospitals National Health Service (NHS) trust, via searches of their clinical databases or opportunistically during routine clinical appointments. Candidates will only be approached by a member of staff at the site or NIHR Clinical Research Network (CRN) staff to whom recruitment activities have been delegated by the local PI. All patient contact information will be kept confidential and undisclosed to the research team. The initial approach will be providing the participant with the information packs, that will contain (i) an invitation letter, (ii) a participant information sheet explaining all aspects pertaining to participation in the study and (iii) a reply slip. Potential participants who wish to find out more about the study will then be able to opt to be contacted by the researcher by completing a reply slip attached to the invitation letter and returning it in a pre-paid and addressed envelope. Alternatively, potential participants can contact the researcher directly via email and phone number that is written on the participant information sheet. It will be explained to the potential participant that entry into the trial is entirely voluntary and that their treatment and care will not be affected by their decision. It will also be explained that they can withdraw at any time, but attempts will be made to avoid this occurrence. In the event of their withdrawal, it will be explained that their data collected so far cannot be erased and the investigators will seek consent to use the data in the final analyses where appropriate. Posters about the study will be on display in the relevant clinical areas.
2. Non-clinical routes: Non-clinical routes will include advertising the study via

• Invitation emails to NIHR Nottingham Hearing participant database. This database includes everyone that previously stating interest to takes part in tinnitus and hearing loss study (including healthy participants and hearing-related patients).
• Newsletter articles and announcements published by the relevant patient and professional organisations.
• Posts on social media channels (Twitter (@hearingnhir, @TIN_ACT @UoNPIBeacon) and Facebook. In all cases, fully informed consent will be sought by explaining the study's procedures and requirements following the standard script to interested potential participants. The potential participants will have ideally a minimum of 24 hours in which to decide whether or not to take part in this study. If the participant agrees to take part in the study, the consent form will be signed, during COVID-19 time this will be done electronically. In all cases, informed consent will be obtained and the consent form will have to be signed prior to any assessment or intervention. Expected duration of participant participation The study is expected to run between August 2021 until April 2022. The expected total duration of participation in the study from the time of informed consent is likely to range from one week to eight weeks. This prolonged time estimate is due to possible organisational difficulties in arranging the audiology assessment and the MRI scans, especially during the COVID-19 pandemic. There will be two visits for each participant, where each visit will last for around 1.5 hours. Removal of participants from therapy or assessments/Participant Withdrawal Participants may choose to withdraw from the study at any time without giving a reason. They will be withdrawn from the study if they are unable to be scanned (due to uncontrollable movement), or if they are unable for any reason to complete the audiology assessment. Participants will also be removed in the event of pregnancy, taking GABA enhancing or anti-depressants medication, having a cochlear or any metal implant that is not compatible with MRI scan, or having metal implant(s) during the time that they first agree to participate and the booked date for scanning, or withdrawal of consent. In the unique case relating to COVID-19, if participants report symptoms of the virus before the study appointment, it will be discussed if they still want to continue participating in the study, and while waiting for their 14 days quarantine the investigators will reschedule their appointment. However, in the case where participants are confirmed to have coronavirus before the study appointment, participants will be removed from the study. Participants will be made aware (via the information sheet and consent form) that should they withdraw the data collected to date cannot be erased and may still be used in the final analysis. If a participant withdraws from the study, the investigators will record the reason for and date of discontinuation. They will be replaced, such that at the end of the study, the number of completed scans, not the number of enrolled participants defines the end. Abrupt termination of enrolment in the study will not affect participant safety. The participants will be made aware that this will not affect their future care. STATISTICS Methods The investigators will use three tests to analyse the data: primary, secondary, and exploratory test. + Primary test Univariate group comparison between-group differences in imaging outcomes, this includes a comparison between: Auditory cortex GABA, GABA/Glx, and Cho level, Auditory cortex cerebral blood flow (CBF), Local functional connectivity density, Interhemispheric auditory cortices functional connectivity, Cross-modal functional connectivity between auditory and visual cortex, and Auditory cortex blood oxygen level-dependent (BOLD) response to visual attention task. Between-group test for differences in correlation, between: GABA and hearing loss, and Auditory cortex functional connectivity and hearing loss. + Secondary test: Within-group correlation analysis of imaging markers of inhibition (such as GABA, functional connectivity, CBF and BOLD response to attention) with audiometric results indexing sensory deafferentation. The investigators will also do regression analysis of imaging metrics with tinnitus severity. + Exploratory: Multivariate prediction model of tinnitus severity and regression of brain network metrics with affective tinnitus phenotypes. The findings will be evaluated by the investigator team. Due to the complex statistic question and strategy, the investigators will seek advice throughout the analysis from both experienced statistician and senior researcher. The analysis will take place on the University of Nottingham (UoN) computers, using R statistical software package and Matlab, and backed up to the UoN servers. Sample size and justification Approximately 60 subjects in total (30 per group). For details, please see below. All tests were done using the R statistical software package and STATA SE16. GABA Previous studies investigating GABA alteration in tinnitus and hearing loss patients are very limited, with the discrepancy between them. In Gao's study with a total sample of 36 subjects, reported around 50% (r=-0.57) of total GABA variance is explained by hearing loss in the elderly cohort, with the assumption that half of their subjects have tinnitus (tin+) and the rest of the half did not have tinnitus (tin-), while in Sedley's study with a total sample of 28 subjects, reported significant GABA concentration reduced (median 1.12 vs 1.28 mM/L) in right auditory cortex for tin+ > tin- respectively. Using these values, the investigators can estimate their effect size using: Cohen's d = (M_1-M_2)/(SD pooled) Also, Cohen's d = 2r/√(1-r^2 ) Where effect size for Gao and Sedley's study respectively are, r=-0.57 (correlation between hearing loss and GABA) and d=-3.7 (GABA effect size between patient and control). One might think of this as a promising effect, but the investigators are aware of the low power of these studies considering their small sample size compared to the population, and thus these effect sizes might be inflated and not represent the true effect size. The investigators then propose to recruit a bigger sample size, and with the time constraint and the limitation due to the COVID-19 pandemic, the realistic number of subjects that the investigators are pursuing is 30 per group (N total=60). With this sample size and based on knowledge from previous studies, the investigators estimated effect size of d=0.7 (two-sided test) and r=0.4 using 0.05 alpha value and 80% power. The investigators will also use a relative measurement to quantify GABA such as GABA/Cr ratio that presumably has a lower variance, and thus will power the effect size, compared to the absolute measure (i.e., mM/L). Previous pilot measurement has shown that GABA measurement was good enough in terms of repeatability and showing the right trend but failed to reach significance in within-subject measurement. Therefore, the other strategy is to involve and inter-relate the GABA measurement with other functional measures (resting-, visual task, tonotopy scan). The analysis will not be limited to only between-subject analysis, but the investigators will also do within-subject analysis that has shown to have more robust results. Functional connectivity In fMRI, the combination of a large number of dependent variables (i.e. huge number of voxel comparison) and a relatively small number of observations (subjects) resulted in underpowered studies with grossly inflated effect size and thus poorly represent the actual effects in the full sample/population. However, using meta-analysis, one can study whether reported findings from selected pooled studies (with a small sample) are really significant or not. Similar to the GABA approach, the investigators then estimate the effect size from a meta-analysis study of functional changes between tin + and tin - and compared it with the estimated effect size. r = z/√N Cohen's d = 2r/√(1-r^2 ) The investigators estimate to be able to detect medium effect size, d=0.6 (N total= 30, whole brain, 0.05 alpha value, and 80% power). As the investigators are fully aware that the results might not represent the full picture of the true effect, the proposed study will still benefit tinnitus field research and add more power to validate previous findings. It is also noted that the meta-analysis study was using a whole-brain measurement (i.e. ReHo), where high variance is expected due to the big number of voxel comparison. Instead, the study proposes to focus on region-of-interest (ROIs) analysis to lower the variance, thus increasing the effect size. Defining the ROIs will be guided by an independent condition such as visual task and tonotopy scan, using Independent Component Analysis (ICA) analysis, and/or using anatomical maps such as Brodmann area. Some functional measures that will be applied are regional connectivity density, interregional correlations, and fractional amplitude of low-frequency fluctuations (fALFF). Cross-validation with GABA measurement, both between-subject and within-subject is expected to enrich the current knowledge between inhibition and neural activation changes in tinnitus patients. Procedures for missing, unused and spurious data In the case of missing or spurious data, the participants whose data is in question will be excluded from that particular analysis.