Abstract
Random noise can enhance the detectability of weak signals in nonlinear threshold systems, a phenomenon known as stochastic resonance (SR). This concept is not only applicable to single threshold systems but can also be applied to dynamical systems with multiple attractor states, such as observed during the phenomenon of binocular rivalry. Binocular rivalry can be characterized by marginally stable attractor states between which the brain switches in a spontaneous, stochastic manner. The switches are thought to be driven by a combination of neuronal adaptation and noise. Here we used a computational model to predict the effect of noise on perceptual dominance durations when either low-contrast or high-contrast stimuli are presented. Subsequently we compared the model prediction to a series of three experiments where we measured binocular rivalry dynamics when noise (zero-mean Gaussian white noise) was added either to the visual stimulus (Exp. 1) or directly to the visual cortex (Exp. 2 and Exp. 3) by applying transcranial Random Noise Stimulation (tRNS 1mA, 100-640 Hz zero mean Gaussian white noise). We found that adding noise significantly reduced the mixed percept duration (Exp. 1 and Exp. 2). This effect was only present for low-contrast but not for high-contrast visual stimuli which is in line with the model predictions. Our results demonstrate that both central and peripheral noise can influence state-switching dynamics of binocular rivalry under specific conditions (e.g. low visual contrast stimuli), in line with a SR-mechanism.