Tinnitus is a hearing disorder characterised by sounds heard without a physical source. The classical explanation for this experience involves hearing loss-triggered excessive neural noise within the central auditory pathway. Since not everyone with hearing loss experiences tinnitus, there must be additional features to the explanation with the neural noise-cancellation mechanism hypothesising those with tinnitus exhibit not only excessive neural noise, but also a reduced ability to “cancel out” this noise. The surgically induced unilateral deafness (UD) group seem to offer a rare opportunity to explore the noise-cancellation mechanism of tinnitus in a relatively well-controlled way, due to a homogeneous pattern of complete hearing loss in the surgery ear with only a portion experiencing tinnitus. The current thesis is a collection of four electroencephalography (EEG) studies designed with this overarching aim.

Studies 1 and 2 compared the cortical auditory evoked potential (CAEP) and the auditory brainstem response (ABR) between UD individuals with associated tinnitus (UD+T; n = 13) and without (UD-T; n = 8), and binaurally hearing control group (n = 13). Stimuli were delivered in quiet and in noise respectively in each study. The noise-cancellation process is thought to work on cortical structures, hence group differences should manifest at the cortical level only. Study 1 showed precisely this pattern. The UD ABR wave III/V ratio was higher than the control group while only the UD+T group showed higher CAEP N1 amplitudes, consistent with their perception of tinnitus. The ABRs show an additive effect (sensitive to both stimulus and noise levels) and the CAEPs show a non-additive effect (sensitive to noise only) with mechanisms behind these phenomena also implicated in the tinnitus-related neural noise-cancellation. In study 2, both UD groups showed ABR additivity, but CAEP non-additivity was only seen in the UD+T group. The collective results of studies 1 and 2 point towards inadequate neural noise-cancellation system in the UD+T group.

The thalamic reticular nucleus (TRN) is the hypothesised structure involved in neural noise-cancellation and its non-invasive biomarker involves sleep spindles (a sleep EEG hallmark). The exploratory study 3 examined the hypothesis that if UD+T individuals show inadequate neural noise-cancellation, this would be reflected in sleep spindles. However, no group differences in a range of sleep spindle parameters were observed, suggesting similar TRN activity regardless of the presence of tinnitus. However, a dual trend towards increased sleep spindle amplitude and decreased density in the UD+T group was observed, which might warrant further exploration. Lastly, study 4 used resting-state EEG to compare the brain rhythms of study groups. The UD group showed higher temporal region gamma power, which is associated with deafferentation induced maladaptive neuroplasticity in the auditory cortex, along with stronger phase amplitude coupling in the ventral anterior cingulate cortex relative to the control group. Amongst future plans, it will be interesting to note whether these trends are relevant in clinical and applied research determining the causal mechanisms of tinnitus.