Tinnitus Is Gain, Not a Phantom Sound

01The evidence

• An integrative model of tinnitus based on a central gain mechanism controlling neural sensitivity accounts for the condition as a homeostatic response to reduced input rather than as a lesion-generated signal.Neuroscience and Biobehavioral Reviews
• Bottom-up models that place the generator in the cochlea have difficulty explaining why tinnitus persists after section of the auditory nerve, which is the observation that favours a central gain account.Bottom-up tinnitus models

02The tension

Tinnitus is treated as a sound to be masked, and masking is the mainstay of management. It gives relief while the masker is on and does not change the underlying condition, which is what one would expect if the percept were a gain artefact rather than a signal.

03The theory

When input from a frequency region falls, the central auditory system increases gain to restore the expected level of activity. Gain increases amplify spontaneous activity, which is then perceived as sound.

If this is right, the percept is not a signal at all but the noise floor raised to audibility. That predicts that anything which lowers central gain, such as attention, sleep, or pharmacological reduction of excitability, should reduce the percept, and that masking works only because it occupies the gain control.

The prediction that distinguishes this from the cochlear-generator account is that the tinnitus frequency should not correspond to a lesion edge but to the region of maximum gain, and that interventions which restore input, such as cochlear implantation, should reduce tinnitus more reliably than masking.

The clinical consequence is that treatment should target the gain control rather than the percept, which is a different set of interventions and a different set of outcome measures.

04What this is not

05Predictions

• P1Tinnitus should persist after auditory nerve section, as a gain phenomenon.
• P2Interventions that lower central gain should reduce the percept without masking.
• P3Restoring input, for instance by implantation, should reduce tinnitus more reliably than masking.

06What would refute it

Tinnitus abolished by sectioning the nerve, or unaffected by gain manipulation, would restore a peripheral generator.

07What it would change

It would change treatment targets from masking to gain control, and change trial endpoints.