Abstract
Filling-in occurs when a visual feature is perceived in a particular region of visual space though that feature actually is absent from the region and present instead in a nearby area. In the particular case of color, a classical example of filling-in is the Boynton Illusion, in which the area contiguous with a black squiggly contour is filled-in by the color from a yellow nearby region. In this case, the squiggly line serves as a luminance contour that bounds the area filled-in by color.
PURPOSE: The present study investigated whether color filling-in is contained also by illusory contours.
METHODS: This study evaluated different kinds of contours: real (luminance-contrast edge) and illusory (Kanizsa square from solid pacmen, Kanizsa square from “bull's eye”; pacmen, and horizontally phase-shifted vertical lines). For all stimuli, a yellow square on an achromatic background was presented within one type of contour. In one condition (a control), the yellow square physically abutted the contour. In two other conditions, the square was not touching the contour: in one condition it was approximately 90% of the original square size and in another condition about 80%. These two sizes addressed whether filling-in occurred. Stimuli were presented at five different durations: 1, 2, 3, 4.5, or 8 sec. The subject indicated whether the yellow square appeared to be touching the contour (filled-in color) via a button press. The proportion of times that filling-in occurred at each duration and the response time of filling-in was measured.
RESULTS & CONCLUSIONS: Filling-in occurred for both the real contour and the three types of illusory contours. Thus contours defined by both local luminance contrast and object-based boundaries are capable of constraining filled-in color.
Supported by NIH grant EY-04802.