Novel evidence for cue-based retrieval of top-down sets in spatial cueing.

Task cues that correctly (vs.incorrectly) inform humans about their upcoming tasks, benefit (vs. interfere with) performance because participants can use task cues to retrieve the corresponding task set, so that targets can (vs. cannot) be processed according to the currently applying task set from target onset onwards. Here, we tested if task-associated features of peripheral cues have a similar effect. Typically, peripheral cues with a task-associated, searched-for color (i.e., top-down matching cues) capture attention: Search for targets presented at the cued position (valid condition) is faster than for targets presented away from the cue (invalid condition), even if cues do not predict the likely target location. For example, when searching for red and green targets, a red cue captures attention even if presented prior to a green target, but a blue cue does not. We know that cue-target color congruence-whether the cue has a target-similar color (congruent case) vs. a target-dissimilar color (incongruent case)-additionally expedites vs. delays search times. However, it is unclear if this congruence effect reflects feed-forward color priming of the target only; or if cue-elicited retrieval of color-specific task sets is involved. Crucially, we hypothesized that cue-based task-set retrieval should incur additional costs if the task sets for the two target colors differ more. In contrast, mere feed-forward priming should not be affected by task-set similarity between color-associated task sets. Congruence effects were indeed larger when color-associated task sets were more different. This finding indicates cue-elicited retrieval of color-associated task sets can contribute to effects of cue-target color congruence. Results are discussed in light of recent theories.