On lateral inhibition in the auditory system.

Suga (1994) has presented arguments supporting the existence of lateral inhibition in the auditory system. We developed a computational model of a lateral inhibition neural network possibly taking part in auditory stimulus processing. The behavior of the model under several hypothetical auditory stimuli matches experimental results. We have shown that lateral inhibition can be the cause of tinnitus in some conditions. 95% of subjects have phantom perceptions in the absence of auditory stimulation (acoustically shielded chambers). We suggest that the spontaneous activity in the auditory nerve (a Poisson-like neural noise), processed by lateral inhibition, is responsible for this phenomenon. Activity is generated in the output layer of the network under stimulation with a Poisson-like noise. In the presence of acoustic stimulation this neural noise is masked by the stimulus or the ambient wideband acoustic noise. The wideband noise is effectively suppressed in lateral inhibition. The shape of the probability density function of the noise determines whether it will be effectively suppressed. The edge effect has been demonstrated to be also a possible consequence of lateral inhibition.