Abstract
Remembering unseen layout beyond the edges of a photograph (boundary extension; BE) is thought to facilitate integration of successive views. In prior research, however, the briefest retention interval tested was 1 second (Intraub et al., 1996). Might a veridical representation be retained for a few hundred ms following offset? On each trial, 3 photographs of unrelated scenes were presented for 325 ms each in rapid succession at center screen. They were followed by a masked retention interval and repetition of one picture (the first, second, or third). The repeated picture remained on the screen while the observer rated it on a 5-point scale (“same [0]”, “a little closer [1]”, “a lot closer [2]”, “a little wider-angle [−1]”,or “a lot wider-angle [−2]”). The retention interval was either 1000, 625, 250, or 100 ms (between-subjects variable: total N=144). On 1/3 of the trials the test picture appeared in the same spatial location as the stimulus. Anticipating that an early visual buffer might maintain a veridical representation, on the other trials, the test item was shifted to the left or right, requiring an eye movement - a situation more comparable to visual scanning. Surprisingly, results showed significant BE at all retention intervals, irrespective of serial position or spatial position at test. Beginning with the briefest interval, mean scores were .56, .44, .41, and .41. Spatial extrapolation occurred extremely rapidly for briefly glimpsed pictures, rapidly enough to facilitate integration of successive views.