DOI:10.1068/p3035 The role of iconic memory in change-detection tasks

Many recent experiments have shown that people are unable to detect large changes in visual scenes. This`change blindness' is dramatic and robust and has been demonstrated in a number of ways (for review see Simons and Levin 1997). It occurs when two successive scenes are separated by a blank interval The techniques used to induce change blindness have varied, yet all the methods include an event separating the successive pictures. When the two pictures are not separated by such an event, an observer's ability to detect change is often nearly perfect. What do all of these intervening events have in common that could explain the resulting inability to notice a visual change? One possible answer is that each manipulation generates multiple transients which occur at the same instant as the to-be-detected change. O'Regan et al (1999) showed that introducing multiple transients was sufficient to produce change blindness, even when these transients did not overlap the location of the change. In their experiment, the original and altered picture were presented in succession with no interstimulus interval (ISI). Observers failed to notice an alteration of the picture when it occurred simultaneously with the appearance of six ink blotches which were scattered around the scene. The addition of multiple transients (the blotches) which occur with the to-be-detected change induced change blindness. The importance of visual transients for the detection of visual change has long been known. Phillips and Singer (1974) first reported that changes become difficult to detect when a blank ISI is inserted between the two scenes. In their experiment, they found that the longer the blank ISI, the worse people did at change detection. To explain this,