Abstract
Observers often miss the second of two targets presented in rapid serial visual presentation at a stimulus onset asynchrony (SOA) of less than 500 ms (the attentional blink, AB; Raymond, Shapiro & Arnell, 1992). However, when a distractor that shares a visual feature (e.g., color) with the targets precedes T2, the AB is substantially reduced, suggesting that there is a delay in initiating processing of T2 that is overcome by the cue (Nieuwenstein, Chun, Hooge & Van der Lubbe, submitted). Here, two questions are addressed: (1) does cuing occur when the cue matches the category instead of a visual feature of the targets, and, if so, (2) does cuing depend on whether observers notice the cue. To this end, I used a three-target procedure in which the third target (T3) was presented 67, 134, or 202 ms after T2, which thus served as a cue for T3. T2 could appear either during or after an AB triggered by T1 (i.e., at SOAs of 268 or 670 ms). In a control condition, T2 was omitted. Performance for T3 was analyzed as a function of T1–T2 SOA, T2–T3 SOA, and the success of T2 identification. The critical comparison was between T3 performance when T2 was omitted versus when it was presented but not reported. The results showed that T3 report was unaffected by the AB when it was preceded by an unreported T2 (e.g., T3 report increased from 48% when T2 was omitted to 78% when T2 was presented but failed to be reported). When T2 was identified, however, T3 performance was found to decrease across the T2–T3 SOA, indicating that identification of T2 triggered a new AB. The results confirm an important role for cuing in overcoming a sluggish shift of processing from a first to a subsequent target, even when the cue is an unreported intervening target.