Computational modeling of tinnitus development

Abstract

Animal models and human neuroimaging studies have shown that tinnitus is generated through pathologically altered spontaneous activity of neurons in the central auditory system. Sensorineural hearing loss has been identified as an important trigger for the development of these aberrant patterns of neuronal activity, but the functional mechanisms that underlie this process have not yet been pinpointed. Using computational models, we have investigated which neuronal plasticity mechanisms could account for the development of neuronal correlates of tinnitus after hearing loss. We could show that a model based on the principle of activity stabilization through homeostatic plasticity can explain the development of neuronal hyperactivity as observed in animal studies. Moreover, the model’s predictions of tinnitus frequencies from the audiograms of patients with noise-induced hearing loss and tonal tinnitus are close to the observed tinnitus pitch. The model thus proposes a specific mechanism for how plasticity in the central auditory system could lead to the development of tinnitus after cochlear damage. The model also predicts that central auditory structures may show increased response gain, which could explain why tinnitus and hyperacusis often occur together. Moreover, the homeostasis model is consistent with recent experimental findings from tinnitus patients with normal audiograms, and it explains why auditory deprivation through an earplug can lead to the occurrence of phantom sounds.