Analytical approaches to understanding tinnitus progression and prevention patterns

Abstract

Tinnitus the ongoing perception of sound in the absence of any external source remains one of the most puzzling conditions in auditory science. Its expression is far from uniform. The wide variability, shaped by both individual perception and intricate neurological processes, continues to frustrate clinical understanding and treatment approaches. In this thesis, this novel disease called tinnitus is examined through the lens of digital health and real world data. Tinnitus, the persistent perception of sound without an external source remains a difficult condition to study and treat because of its highly variable expression across individuals. For some it is a mild background noise, while for others it becomes a constant and intrusive presence that disrupts daily life. This thesis explores tinnitus in everyday contexts by applying Ecological Momentary Assessment (EMA) (Goldberg et al., 2017) to capture fine-grained fluctuations in symptoms and evaluate a smartphone based intervention that combines structured counselling (TinEdu) with a personalised sound therapy module (Shades of Noise, SoN). The study was conducted within a multiple baseline design and analysed by applying statistical modelling such as Linear Mixed Effects Models (LMMs) and a Generalised Linear Mixed Model (GLMM). The statistical modelling revealed a gradual decline in tinnitus related distress over time and with stronger improvements during the intervention phase. Loudness patterns proved less consistent showing individual variability and weaker associations with treatment exposure. Moderators such as age, sex, personality traits and depressive symptoms were explored and relieved useful results, though depressive symptoms showed a weak link with louder symptom reports. In contrast, the clearest evidence emerged from Tinnitus Handicap Inventory (THI) , outcomes: 63% of participants achieved a clinically meaningful reduction in tinnitus handicap scores during the intervention when compared with baseline these results corresponds to answering RQ6 and RQ9 through reduction in THI scores significantly. These findings suggest that while daily symptom dynamics remain complex and idiosyncratic, structured digital interventions can deliver tangible clinical benefit. The study highlights the value of EMA for capturing symptom variability and emphasises the potential of smartphone based therapies as scalable tools for managing perceptual disorders such as tinnitus.

Date of Thesis submission	2025
Original language	English
Awarding Institution	Western Sydney University
Supervisor	Kenan Matawie (Supervisor)