Abstract
Contrast-contrast phenomena have been explained by lateral interactions among early cortical neurons or by summation within large receptive fields with antagonistic center-surround mechanisms. The extent of contrast pooling in these has been measured by varying the distance between center and surround. Here we have studied the effects of a gap between center and surround on contrast perception when the spatial content of the gap is varied. Center and surround were horizontal sinusoidal gratings, (spatial frequency 4 cpd), with equal contrasts (0.1). In the first experiment the contrast of a 1 deg test patch was matched in contrast to a patch of varying diameter (0.5 – 12 deg). In the second experiment there was always a 7.2 min gap between center and surround. The gap consisted of a narrow ring of vertical 4 cpd grating with contrast of either zero or equal to that of center and surround. Surround width was 12 deg, and width of the center was varied between 0.5 and 8 deg. As in Experiment 1, a 1.0 deg test patch was matched in contrast to the center. Increasing the patch diameter in the first experiment decreased the apparent contrast of the grating, although the magnitude of the effect differed among observers. Adding the vertically-modulated annulus with no additional surround had no effect on apparent contrast. Introducing an unmodulated gap (0 contrast) with a surround increased the suppression of apparent contrast relative to the no-gap condition of Experiment 1, and the diameter of the center patch ceased to have any effect. When gap contrast of vertical modulation equalled that of center and surround, apparent contrast was enhanced with small center sizes and suppressed with larger center sizes. These results suggest that contextual interactions at low levels of contrast processing are more complex than previously thought, and will require more complex models than have been thus far proposed.