Handling distractor interference in mixed and fixed search

When searching for a target, humans can reduce interference from a colour distractor by using different strategies, such as down-weighting the distractor dimension or the most likely distractor location. Less clear is how the heterogeneity among the nontarget features, and the similarity of target and distractor modulate distractor interference and thereby distractor location learning. In three tasks, we examined this question by analyzing target reports and selection errors. In the mixed-feature task, participants searched for a singleton shape among homogeneously shaped nontargets, forcing participants to use singleton-detection mode. In two fixed-feature tasks, participants needed feature-search mode to find the target among heterogeneously shaped nontargets, which were either gray (gray fixed-feature task) or coloured (colour fixed-feature task). As expected, the results in the mixed-feature task showed evidence for distractor interference and distractor location learning with suppression of the high-probability location below a baseline level. In the fixed-feature tasks, the results differed: We observed distractor interference in the gray fixed-feature task, but a slight distractor-presence benefit in the colour fixed-feature task, possibly indicating up-weighting of the target on the featural level. Distractor location learning occurred in the first session of the colour fixed-feature task, but results in the gray fixed-feature task indicate that the participants relied on other strategies to handle distraction. Interestingly, we found overnight learning between sessions in all three tasks, which was more pronounced in the fixed-feature tasks. These results indicate that participants can flexibly adjust their attentional strategy based on the nontarget features to reduce distractor interference.