Nonlinear Neural Dynamics of Mutual Inhibition Circuit in a Real-Life/Computer Model Hybrid System

Abstract

To process ambiguous and noisy images, often experienced in our daily life, the neural system has to actively select and organize the input signals. For a percept to emerge it has been assumed that there are selection processes of competing neural pools. Theoretical research assumed a mutually inhibiting neural circuit underlying the competition and successfully modeled bi-stable perception that occurs in response to ambiguous images. We developed an experimental system to record two real life-pyramidal neurons (in vitro) connected by modeled mutual inhibition circuit (in silica). We show that simultaneous stimulations of the two pyramidal neurons in this hybrid system evoked bi-stable activity. Furthermore, the effect of adding noise and changing stimulus strength showed similar characteristics known from bi-stable perception, suggesting a fundamental role of the non-linear dynamics in perceptual organization.