抄録
Diverse types of chaos have been confirmed at several hierarchical levels in the real neural systems. The role of chaos in actual perceptual processes should be investigated energetically. In this paper, we present a perception model of ambiguous patterns based on the chaotic neural networks with the refractoriness and the time-hysteresis, from the viewpoint of bottom-up approach. Computer simulations reproduce the Gamma distributions obtained in psychophysical experiments. This is difficult for the stochastic activity to attain to. In addition, we refer to a dynamic picture regarding the chaotic trembling mode as the small involuntary perceptual movement.
