Perception, 27(suppl.): 36. [ECVP 1998; Oxford, England.]

Attention is needed to detect change under "flicker" conditions, where an original and a modified image are continually alternated, with a brief blank field separating them (Rensink et al, 1997 Psychological Science 8: 368-373). To determine the efficiency of attentional selection, flicker displays were created using arrays of rectangles, with half the items black and the others white, and half the items horizontal and the others vertical. All combinations were equally likely. In half the trials, displays differed by a feature change in one of the rectangles; in the remainder, displays remained the same. Observers were asked to detect in each trial whether a change was occurring between the two alternating images.

The efficiency of selection for contrast polarity was determined by comparing search speed for unconstrained orientation change against that for orientation change where the target was always black. For display on-times of 80 ms and off-times of 120 ms, efficiency was close to 100%.

The efficiency of selection for orientation was similarly determined by comparing search speed for unconstrained polarity change against that for polarity change where the target was always vertical. Here, however, a different pattern emerged: for on-times of 80 ms and off-times of 120 ms, efficiency was only 30%. This was not due to slower processing, for increasing display on-times to 320 ms caused efficiency to fall to zero. Instead, these results suggest that attentional selection for orientation is limited to only one object at a time.