Abstract
In the Craik–O’Brien–Cornsweet effect (COCE), a central edge of an opposing pair of luminance gradients makes the adjoining regions with identical luminance appear to differ in brightness. Some have suggested that the COCE is mediated by cortical propagation processes pertinent to perceptual filling-in. While the COCE can be affected by contextual cues such as lighting direction, there is evidence that low-level mechanisms are involved in this illusion. The COCE requires a duration of 0.1 s (Dakin & Bex, 2003), or, if no backward masks were applied, 0.05 s (Masuda et al., 2011). We investigated the temporal characteristics of the COCE and how the induced brightness might be used to track apparent motion. Stationary gratings consisting of the COCE-inducing luminance pattern along a circle were presented frame by frame. The phase of each subsequent grating was shifted by a quarter, so that the apparent differences in brightness could determine the direction of rotation, whereas the physical edges of the luminance gradients couldn’t. Immediately after the stimulus presentation, the stimulus region was covered by a stationary random-dot pattern as a mask. Participants were asked to report whether the perceived direction of the rotational apparent motion was clockwise or counterclockwise. The results demonstrated that the participants correctly judged the direction of the phase shift as the duration of the COCE stimulus increased. The direction was correctly judged when the duration of each frame was as brief as 0.06 s at optimal numbers of frames. The rate of correct responses increased with increasing number of frames up to 8 frames. These results indicate that the COCE involves a fast process taking less time than previously estimated on the basis of the filling-in hypothesis. We will discuss the perception and the temporal characteristics of the COCE in the context of visual motion processing.