The Guidance of Attentional Selectivity in Visual Search Is Always Feature-Based: Behavioral and Electrophysiological Evidence From Feature and Conjunction Search Tasks

Abstract

It is generally assumed that the guidance of attention in visual search operates in a feature-based fashion, but this conclusion is mainly based on the results of search tasks with feature-defined targets. Here, we investigated whether attentional templates may contain integrated object-based representations in tasks where targets are defined by feature conjunctions. Attentional load was manipulated by instructing participants to search for a single or one of two different possible targets. Search targets were either defined by a particular color or shape (feature search tasks) or by a color/shape combination (conjunction search task), and load effects were compared between these two types of tasks. Load effects on search performance as well as on electrophysiological markers of attentional guidance (N2pc components) and the number of attentional templates activated in working memory (CDA components) were much more pronounced during conjunction search. This suggests that attentional templates are always feature-based, even when object-based templates are in principle available to reduce memory load.