Abstract
Because our retinal images are continuously changing, we often have only short time to accurately perceive a new stimulus. It is still unclear how we overcome this problem. For motion perception, we recently showed that temporal synchrony between a moving stimulus and a static reference is important for rapid and accurate perception. Specifically, we found that only for short stimulus presentation, the coherence-signal necessary to discriminate motion in a dynamic random-dot field was halved when a static surround composed of static dots was presented in synchrony with the random-dots (Linares, Motoyoshi & Nishida, 2012). From one perspective, this surround facilitation is counterintuitive. A transient surround adds motion energy in all directions, which, if integrated with the dynamic random-dot field in the center, would reduce the signal to noise ratio, and decrease motion sensitivity. Since the surround instead increases motion sensitivity, the surround might be somehow segmented from center, rather than integrated with it. Segmentation might be based on the difference in temporal structure between the static surround and the dynamic center. If so, adding dynamic dots in the surround should impair rather than facilitate motion sensitivity. From a different perspective, one could consider that the additional energy of the surround could facilitate motion perception due to an accelerating non-linearity of the underlying neural response. According to this view, adding dynamic dots in the surround should facilitate motion perception. We found that a synchronous surround composed of dynamic dots impairs rather than facilitate motion sensitivity in agreement with the first prediction. This finding suggests that center-surround mechanisms integrate the signal of elements sharing temporal structure, while segment and enhance the signal of elements with different temporal structure. In another experiment, we found that the signal enhancement by a static synchronous surround also improves minimum motion detection.
Meeting abstract presented at VSS 2013