The contrast dependency of the Munker-White illusion

In the Munker-White illusion, a grey bar placed on the black stripes of a black-and-white grating appears brighter than the one place on the white stripes despite having the same luminance. We investigated the mechanisms underlying this illusion with an asymmetric matching paradigm. Participants viewed a screen displaying a point at the center and a matching bar on one side and a test pattern with a bar embedded either in a black or a white stripe of a black-and-white square wave grating on the side of the display. Participants were instructed to compare the brightness of the test with the matching bar and indicated which bar seemed brighter. In each run, the luminance contrast between the test bar and the background was either a positive or a negative constant (brighter or darker than the background respectively) while the contrast of the test bar varied from trial to trial with the participants’ response and a one-up-one-down staircase algorithm. The test contrast, polarity and location all had a significant effect on the strength of the illusion measured as the difference between the test and the matching contrast). The illusion reduced quickly as the test bar contrast increased when the test was embedded in a stripe of the opposite contrast polarity. However, the illusion was enhanced first before decreasing with the test contrast when the test was embedded in a stripe of the same contrast polarity. The test contrast effect is inconsistent with the high-level theories of Munker-White illusion, such as the belongingness theory, which would suggest no test contrast effect as the stimulus configuration remained the same regardless test contrast. This effect is qualitatively consistent with the contrast theory, which predicted an increased illusion with higher test bar contrast. However, the polarity effect suggested an involvement of a nonlinear process.