Abstract
In a feature-based visual search, response time increases logarithmically with set size when searching in parallel and linearly when searching serially. Previous work in our lab used stimuli with a center-surround two-color structure to investigate search efficiency in color X color conjunction search. Specifically, participants searched for the red-center/green-surround target in displays containing only one type of distractors that varied in color from the target (reversed center-surround colors, same-center/different-surround, same-surround/different-center, and different center-surround colors). The results showed that search performance was neither linear nor logarithmic in these conditions. Further experiments suggested that participants might be utilizing a singleton-search strategy, rather than a feature-search strategy, so that instead of using template-based guidance with two colors defining the target in a specific spatial arrangement they were searching for the oddball item in the display. Two experiments also showed that with sufficient training, participants can search for the target in parallel, revealing that color x color conjunctions can be processed in parallel. In the present study, we tested whether the use of heterogeneous displays would also allow participants to use a feature-based strategy and search in parallel. Participants searched for the red-center/green-surround targets among two types of distractors that varied in number in a fully crossed manner (e.g., 1, 3, 6, 11 of reversed center-surround colors distractors and 1, 3, 6, 11 of same-center/different-surround distractors). The results showed that response times increased logarithmically as a function of the two different types of set size, suggesting that participants had no problem using a search template consisting of a color x color conjunction. In follow-up experiments, we tested displays containing three different types of distractors simultaneously. In conclusion, by using heterogeneous displays, participants predominantly choose to use template-based guidance and can discard in parallel distractors that are defined by color-color conjunctions.