Passive Categorisation Training for Tinnitus Management

Abstract

Background: While sound therapy (ST) is commonly used for tinnitus relief, tinnitus heterogeneity necessitates advancements in such conventional methods. This thesis evaluated tinnitus replica generation and role in tinnitus management, specifically within a novel effortless auditory categorisation-based ST. Aims: 1. Review/understand mechanisms of auditory categorisation training and its role in tinnitus management. 2. Assess the use of different sounds for appropriate/accurate tinnitus synthesis. 3. Evaluate an effortless personalised categorisation training approach relative to conventional ST using broadband noise (BBN). Methods: Three narrative reviews were undertaken regarding: 1. Current knowledge/theories surrounding tinnitus. 2. Psychoacoustic assessment of tinnitus. 3. Auditory categorisation and its potential role in tinnitus perception. Pilot study (n=22): Evaluated eight tinnitus replicas based on their emotional affect and ability to replicate participant’s tinnitus. Effortless auditory categorisation study (n=29): Compared the effects of an effortless auditory categorisation training (CT) paradigm to conventional white noise (WN) ST. Results: Narrative reviews: 1. Tinnitus heterogeneity necessitates a shift to tailor-made, machine-learning informed management methods. 2. Psychoacoustic measures are vital to tinnitus evaluation. 3. Categorisation-based perceptual training paradigms present a promising means of tinnitus management. Pilot study: Puretone-based replicas had greater negative affect (cf. noise-like replicas). Avatars and pure-tone and localisation replicas accurately capture the tinnitus experience. The uncanny valley phenomenon was not substantiated in tinnitus matching. Effortless auditory categorisation: WNST resulted in significant improvements in tinnitus impact independent of counselling; CT did not. Acute WN resulted in changes within the default mode network (DMN). Acute CT resulted in overall reduction in activity at the prefrontal cortex (PFC), and a decrease in gamma-band activity within the auditory cortex. Conclusions: To effectively address tinnitus heterogeneity, personalised management using psychoacoustic assessments and employing machine-learning may be needed. Tinnitus replica assessment revealed that both simple and complex copies accurately capture tinnitus, indicating that replica complexity should reflect its intended/overall purpose. Reductions in tinnitus complaint were greater with WNST than the experimental CT, however the CT approach showed acute changes in brain areas associated with tinnitus, different from those altered by WNST. This evidence supports further investigation into CT in tinnitus management. While tinnitus relief can be achieved solely using WNST, neural changes within the DMN (following WN) and PFC (following CT) are encouraging as they are directly correlated with tinnitus distress.