Task relevance changes the impact of salient items on attention

Studies examining the influence of salience on attention typically use paradigms where the salient item is not beneficial for the task to examine salience-only processing. In the additional singleton paradigm and learned suppression paradigms, the salient item is never the target (0% validity). In the contingent capture paradigm, the salient item is sometimes presented at the target location (25% validity). However, the impact of relevance of the salient item has not been systematically manipulated within these paradigms. Taking all of these into account, we created a task like the additional singleton paradigm, in which we manipulated both the search mode (type of search array: singleton detection or feature search), and the relevance of the salient item. The salient item was present in 60% of trials, but we manipulated the proportion of valid vs. invalid salient trials between subjects (0, 25, 50, 75 or 100% validity). This manipulation allowed us to observe both costs of salient distractors and benefits of salient targets, and how they are affected by relevance and search mode. We observed lower costs and benefits for singleton detection mode than feature search mode, leading to main effect of search mode. We found RT costs and benefits increased with increasing relevance. Finally, there were larger increases in costs and benefits for feature search mode than singleton detection mode. Simple main effects analyses showed that the impact of relevance was present for both feature search and singleton detection mode. These results suggest that there is not a simple distinction between bottom-up and top-down guidance of attention. The bottom-up factors can be weighted flexibly, based on the relevance to the task.