Abstract
Illusions where a luminance border gives rise to illusory brightness percept have been taken as evidence that local border contrast plays a dominant role in perceived brightness. One of the key issues has been whether this is achieved by an active, temporally extended filling-in mechanism that propagates the ‘brightness signal’ initiated by the border responses.
Here, a reverse correlation technique was used to investigate the spatiotemporal characteristics of edge-induced brightness. We used a contrast polarity discrimination task for a 1-D Craik-Cornsweet-O'Brien stimulus in which the edges induce the perception of a 2.6 degrees wide bar appearing either brighter or darker than the background. The target stimulus was shortly flashed (4 frames, 67 ms) added in the middle of a dynamic white noise sequence of 36 frames. A fixation mark was used to indicate both the location and the duration of the target stimulus.
The resulting spatiotemporal classification image shows that brightness perception is associated with two spatiotemporally distinct responses: (1) a spatially local and temporally short border response and (2) a weaker response that corresponds spatially to the illusory brightness percept and is temporally more extended than the border response. Properties of this latter ‘brightness’ response resemble the output of a spatiotemporal low-pass linear filter. We did not find evidence for significant delays related to the propagation of the filling-in signal: the brightness response does not seem to be markedly delayed when compared to the border response.