Abstract
In the McCollough effect, adaptation to red-vertical and green-horizontal gratings causes achromatic vertical and horizontal gratings to appear greenish or reddish, respectively. Compared to other visual aftereffects, the color biases are unique and enigmatic for their remarkably long persistence. The stimulus properties required to engage these long-term color biases remain poorly understood. We are examining these requisite properties by titrating the McCollough stimulus in order to identify the conditions under which the long-term aftereffect fails. In this study we report on the consequences of varying the spatial profiles of the adapting and test gratings. We varied the phase spectra of the squarewave gratings typically used to induce and test the effect, so that the patterns instead had random phase or roughly triangular-wave profiles. All patterns had the same amplitude spectra. For each profile observers adapted to the red or green chromaticity of the monitor phosphors paired with horizontal or vertical, and were then tested for all profiles for achromatic stimuli displaying both orientations. As expected, the squarewave patterns induced strong and persistent color-contingent aftereffects, and these aftereffects also showed strong transfer to the phase-varied profiles. In contrast, the random or triangular waveforms produced little if any color aftereffects, on the same or different profiles. These results are consistent with the suggestion that the McCollough Effect may depend on edge color components (Broerse and O’Shea, 1995), and specifically may require adaptation of mechanisms sensitive to the color of asymmetric luminance edges.