On the costs of lag-1 sparing.

The attentional blink (AB) is a dual-target, rapid serial visual presentation (RSVP) deficit thought to represent a failure of perceptual awareness that reflects the dynamics of temporal attention. However, second target (T2) report is typically unimpaired when the targets appear within 150 ms of one another (i.e., lag-1 sparing). In addition, this sparing can be extended if more targets appear sequentially. It is thought that sequential targets are processed in the same attentional window. Here, we investigated the fate of targets processed in these windows and, specifically, the consequence for subsequent targets when an item at lag-1 is reported versus missed. The results demonstrated that target encoding in attentional windows has an all-or-none influence on subsequent item report: When comparing two- and three-target (T1 and T2 not separated by distractors) RSVP streams, there was no difference in AB magnitude for the final target when either T2 or T1 was missed in the three-target condition, but both of these conditions had significantly smaller blinks than those observed when T1 and T2 were accurately reported. A comparison of our results to a computational model of temporal attention demonstrates how structural limitations on the rate of encoding affect perception, even during sparing.