Perception, 24(suppl.):26-28. 1995. [ECVP 1995; Tuebingen, Germany.]

Several recent investigations (Grimes, in press; McConkie and Currie, in preparation) report that large changes in images of natural scenes can remain unnoticed if these are made during saccades. We show here that similar massive effects can be obtained without synchronization to saccades. This is done via a "flicker" technique in which an original and an altered image (each of duration 240 ms) are repetitively alternated, with a blank field (duration 27 or 290 ms) between each display. One of four kinds of change (color, left-right reflection, translation, or appearance/disappearance) were made in the foreground or background of each scene. Many of these changes were difficult to detect, even over long periods of observation (35 seconds). We believe that this is due to the spatially-distributed transient induced by the blank field, which swamps the localized flash that would otherwise draw attention to the changes; observers were therefore forced to rely on higher-level (probably non-iconic) representations of the scenes to detect the change. Our results indicate that the failure to notice scene changes during saccades is not due to saccade-specific mechanisms, but rather, involves more general mechanisms of visual attention.