When brief blank fields are placed between alternating displays of an original and a modified scene, a striking failure of perception is induced: The changes become extremely difficult to notice, even when
they are large, presented repeatedly, and the observer expects them to occur (Rensink, O'Regan, & Clark, 1997). To determine the mechanisms behind this induced "change blindness", four experiments examine
its dependence on initial preview and on the nature of the interruptions used. Results support the proposal that representations at the early stages of visual processing are inherently volatile, and that
focused attention is needed to stabilize them sufficiently to support the perception of change.