15Unifying Models and What They Mean for Treatment

FWhy we need more than one model

The competing theories of tinnitus are often presented as rivals, but they are better read as describing different stages of one process. A useful frame is a four-stage cascade: generation of an aberrant signal, perception of it as sound, distress attached to it, and chronification that makes it permanent. Most models explain one or two stages well and stay silent on the rest [2014].

De Ridder and colleagues made this synthesis explicit with an integrative model in which tinnitus is a single percept emerging from several separable, interacting subnetworks — an auditory network that supplies the sound, and salience, distress and memory networks that supply its loudness, unpleasantness and persistence [2014].

TGeneration and perception: gain, dysrhythmia and the DCN

The earliest stages are bottom-up. The central-gain model holds that after deafferentation the auditory system turns up its gain to restore mean firing rate, amplifying spontaneous activity and noise into a perceptible signal — Noreña framed this as a central-gain controller of neural sensitivity [2011]. The DCN/maladaptive-plasticity account locates an early generator in the dorsal cochlear nucleus, where disinhibited fusiform cells fire spontaneously and burst [2016].

The thalamocortical-dysrhythmia model explains how that activity becomes a percept: deafferentation slows thalamic firing, the resulting low-frequency rhythm escapes lateral inhibition at its edges, and a halo of pathological gamma in cortex is read as sound [1999]. These three are complementary — gain sets the volume, the DCN provides an early source, and dysrhythmia turns activity into perception.

CPerception as inference: the Bayesian-precision model

A more recent reframing treats perception as prediction. In the Bayesian brain, what we hear is the brain’s best inference combining prior expectations with sensory evidence, each weighted by its precision (reliability). De Ridder and colleagues proposed that phantom percepts resolve sensory uncertainty: when deafferentation deprives the auditory system of reliable input, the brain fills the gap with a prediction — and that prediction is tinnitus [2014].

Sedley and colleagues developed this into a sensory-precision model in which an aberrantly high-precision prior for the tinnitus frequency keeps the percept locked in, explaining why tinnitus is so stable and why attention (which adjusts precision) modulates it [2016]. This model unifies generation and perception under a single computational principle.

CDistress and chronification: Jastreboff and the triple network

The top-down half of the story explains suffering and permanence. Jastreboff’s neurophysiological model (1990) made the pivotal claim that the problem of tinnitus is not its generation but its perception and emotional evaluation: a neutral signal becomes bothersome only when limbic and autonomic systems tag it as threatening and habituation fails [1990]. This is the conceptual engine of Tinnitus Retraining Therapy.

Network neuroscience operationalised this with the triple-network framework. Vanneste and De Ridder showed that tinnitus engages overlapping subnetworks — an auditory percept network plus salience, attention/control and memory networks — and that distress in particular maps onto a salience/limbic system [2012]. The percept and the distress are dissociable: blind-source-separation work identifies a distinct ‘distress network’ over-active in bothered patients [2011]. Chronification, in this view, is the consolidation of an aversive auditory memory that keeps the loop running [2011].

COne stage model, many targets

Assembling the pieces yields a single staged model with stage-specific treatment targets. Generation (peripheral trigger, gain, DCN) argues for restoring input — hearing aids, cochlear implants, and bimodal/somatosensory neuromodulation aimed at re-balancing inhibition [2016]. Perception (dysrhythmia, aberrant priors) is the target of cortical neuromodulation (rTMS, tDCS) and, conceptually, of therapies that re-weight precision [2014]. Distress (limbic/salience) is the domain of CBT, mindfulness and counselling — the strongest-evidence interventions. Chronification argues for treating early, before the aversive memory consolidates [2014].

The clinical payoff of the unified view is that it explains why no single treatment works for everyone: a patient dominated by peripheral generation needs amplification, while one dominated by distress needs psychological therapy. Matching the intervention to the dominant stage is the goal of mechanism-based, personalised tinnitus care [2015].