Abstract
Our visual system can prioritize target-relevant features and also suppress features related to nontargets during visual search. Although prioritization and de-prioritization using attentional templates have been found in color (Arita et al., 2012; Reeder et al., 2017) and shape (Tanda & Kawahara, 2018) domains, the paradigms used to demonstrate these findings have raised concerns that such apparently feature-based (de)prioritizations might be achieved indirectly by location-based templates. Specifically, participants might encode where items containing specific features (e.g., color or shape) are located and then construct spatial (de)prioritization maps in which search items are arranged spatially. The present study bypassed this concern by using object-based arrangements of search items and examined whether nonspatial (de)prioritization facilitates visual search. Participants identified the color of a target cross among nontarget crosses distributed along the edges of two overlapping geometric shapes (e.g., circle and triangle). This arrangement discouraged space-based (de)prioritization strategies. Every trial was preceded by a positive, negative, or neutral cue. A positive cue indicated that the target would appear on the cued shape, a negative cue indicated that it would appear on the non-cued shape, and a neutral cue indicated that the cued shape would never appear during the trial. Among the three conditions, the mean reaction time was shortest under the positive cue condition and longest under the neutral cue condition, indicating that object-based prioritization and deprioritization occurred. An identical pattern of results was found in a subsidiary experiment in which an outlined shape that carried half of the search items along its edge rotated clockwise, and the other shape carrying the remaining items rotated counterclockwise, further discouraging space-based inhibition. These results suggest that the visual system can prioritize and deprioritize multiple search items based on nonspatial, shape-based features.