| The visual perception of contours by the brain is selective. When embedded within a noisy background, closed contours are detected faster, and with higher certainty, than open contours. We investigate this phenomenon theoretically by considering a set of locally coupled oscillators subject to local uncorrelated noise. Noise is needed to overcome the excitation threshold and evoke spikes and intermediate noise leads to the best result. In that sense the current work make use of the noise-induced effect known as Coherence Resonance, also referred to as Stochastic Coherence and Self-induced Stochastic Resonance in the literature. We model one- and two-dimensional structures with different types of coupling and consider the synchronization throughout them as a mechanism for contour perception, for various system sizes and local noise intensities. The model with a closed ring structure shows a significantly higher synchronization than the one with the open structure. Interestingly, the effect is most pronounced for intermediate system sizes and noise intensities. |
|