Figure 1.

Conceptual Overview of the Model. The core of the model is the process of perceptual inference, driven by a descending auditory prediction (yellow arrow; initially of silence) and by spontaneous activity in the auditory pathway that constitutes a sensory prediction error (red arrow) and, in effect, a ‘tinnitus precursor’. The precision of ascending prediction errors and descending predictions is denoted by arrow width, where the tinnitus precursor has an inherently low precision and therefore makes little or no contribution to perception or posterior beliefs. Various factors (blue boxes) – alone or in combination – can increase the precision of the tinnitus precursor which, if it becomes sufficiently high, results in the perception of tinnitus (orange). Even if this increase in precision is reversed, tinnitus can be perpetuated by two mechanisms (orange arrows): learning to expect tinnitus, and reinforcement via attention.