Asymmetries in Brightness and Darkness for Assimilation and Simultaneous Contrast?

Recently, de Weert and Spillmann (1995 Vision Research35 1413 – 1419) reported a striking example of assimilation. Their stimulus was a pincushion formed by four arcs, each consisting of a number of black and white rings on a gray background (the area surrounded by the rings, see their figure 1). When the gray background is immediately surrounded by white rings, the background appears lighter and vice versa. When a subject is asked to match the luminance of a circle in a different spatial location for both the ‘lighter’ and the ‘darker’ pincushion, the matching luminance of the test is lower than the actual background luminance. This result is surprising but also counterintuitive. For example if a ‘light’ pincushion is matched with a ‘dark’ pincushion, it is expected that the luminance of a ‘light’ pincushion needs to be decreased in order to match the ‘dark’ pincushion. Conversely, the luminance of the ‘dark’ pincushion needs to be increased to match the ‘light’ pincushion. Therefore luminance values on both sides of the default background luminance are expected. We replicated their basic experiment and found the same results. In additional conditions, we had subjects adjust the background luminance of a ‘light’ pincushion compared to a ‘dark’ and vice versa. In that case the luminance values were symmetrical on either side of the default background luminance. It would seem that the method of testing is crucial here. Therefore we also tested simultaneous contrast stimuli (all rings were made black or white) using the circle-match-task as in their original experiment. We found that both values were nicely distributed on both sides of the background luminance value, indicating that de Weert and Spillmann's finding is not attributable to the test condition as such.