Abstract
The Rotating Snakes is a dramatic illusion of motion perceived in a static figure (Kitaoka & Ashida, 2003), which is comprised of luminance-defined micropatterns: black, dark-gray, white, and light-gray. Murakami et al. (2006) demonstrated that the illusion had a positive correlation with fixation instability and suggested the model that the biphasic nature of the temporal impulse response function is a prerequisite for the illusion. Previously, we tested this model by changing the retinal illuminance, thereby indirectly changing the shape of the temporal impulse response function from biphasic to monophasic (Hisakata & Murakami, 2007 VSS). The strength of illusion decreased with decreasing retinal illuminance. In this study, we aimed to ensure the expected change in the impulse response with retinal illuminance. We used the double-pulse method (e.g., Burr & Morrone, 1993) to measure the temporal impulse response at several retinal illuminances. The two successive Gabor patches (horizontally oriented, 1 cpd, sigma 1.79 deg) were presented at 12 deg to the left. For 16 SOA conditions and two Gabor-phase conditions (in-phase or out-phase), contrast sensitivities were measured. Based on the measured sensitivities, temporal impulse response functions were estimated by a standard protocol. As a result, the shape of the impulse response indeed changed from biphasic to monophasic within the range of illuminances we previously used (2007 VSS). We also confirmed that the visibility of high spatial-frequency components in the illusory figure was not compromised under dark illuminances, and that intermediate contrasts (as in dark-gray or light-gray regions of the figure) were detectable. We conclude that some transient temporal processing system is necessary for the illusion, and argue that the same mechanism may also be involved in the processing of retinal-image motions with fixation instability. We will discuss the validity of several models for the Rotating Snakes or other peripheral drift illusions.
This work was supported by Grant-in-Aid for Scientific Research 18203036.