An adaptive tinnitus model based on least mean square algorithm

Tinnitus is a kind of auditory perception in the absence of external stimulus, its pathogenesis is not fully understood yet. Animal studies indicate that hearing loss through cochlear damage can lead to behavioral signs of tinnitus, however its mechanism is still unclear. Sound therapy is one of the most effective treatments for tinnitus on clinic. In this paper, we proposed a novel adaptive model for tinnitus to expound the pathogenesis of tinnitus and the clinical effects of sound therapy. This model can maintain the signal intensity in central auditory system to a stable value by employing the Least Mean Square (LMS) algorithm. When hearing loss occurs in this model, the central auditory system will adaptively magnify the auditory signals as well as the spontaneous signals. The significantly increased spontaneous signal is assumed to be tinnitus in this paper. Simulation results demonstrate that this model can not only explain the pathogenesis of tinnitus and the clinical effects of sound therapy, but also uncover the various curative corresponding to different external signals and the phenomenon of tinnitus relapse after the treatment of sound therapy.