ECVP 2002 -- CONFIRMATION

Your abstract has been succesfully saved.

Primary keyword: Modelling

Secondary keyword: Binocular Vision

------------------------------------------------------------------------

Variability and adaptability in the visual system

T. Yanagawa, F. Taya¶, K. Mogi# (Tokyo Institute of Technology, Department of Computational Intelligence and Systems Science, &, Sony Computer Science Laboratories , Takanawa Muse Building , 3-14-13 Higashi-Gotanda, Shinagawa-ku, , Tokyo 141-0022 Japan , ; ¶ Department of Physiology, Osaka University Medical School; # Sony Computer Science Laboratories, ; email: tyanagaw@hotmail.com)

The visual system is highly tuned to adapt to the changing visual environment. Setting priorities right and generating appropriate explicit representations of visual features in real time is not a trivial task. This is particularly true in cases of multi-stable perception such as ambiguous figures and binocular rivalry. Number of studies (e.g. Logothetis et al. 1996, Kovacs et al. 1996, Taya & Mogi 1999) have presented evidences that higher features and saliency crucially affect the ocular dominance pattern in binocular rivalry. Here, we present a model of binocular rivalry. Recurrent connections between the early visual area (V1) and form and motion area (IT and MT, respectively) is assumed to play an essential role. Simulation results suggest that the ability of the visual system to adapt dynamically to visual features and saliency is contingent on a certain level of variability of neural firings within the visual system even in the absence of dynamic visual features. This

result leads to a novel functional interpretation of the universal existence of firing variability observed in the central nervous system, and puts visual perception within the broader framework of fluctuation-related biological phenomena such as stochastic resonance (e.g. Marino et al. 2002).