Dependence of brightness induction on the contrast polarity of a disk-annulus stimulus

Classical theories of brightness and lightness induction have largely been stated in terms of two opposing influences from the surround: contrast and assimilation. In fact, these two effects are part of a more complex but reproducible and mathematically well-defined statistical process in which both low-level neural constraints and top-down cognitive judgments contribute to the measured percept. Here we test a particular model of this process based on edge integration theory with contrast gain control. We employed a brightness matching technique based on disk-annulus stimuli and showed that we can account for our results with a quantitative model in which the contrast polarity of the stimulus plays a critical role.