Abstract
A common assumption in several visual search theories is that the similarity between items in the display and the target template should impact or even determine the order in which those items are scrutinized during visual search. Here, we tested that assumption and failed to find support for it. We selected two different types of "candidate" stimuli (i.e., distractors that are sufficiently similar to the target as to require attentive scrutiny): (i) rotated Ls and (ii) rotated Ls with one of the two lines shifted inward so as to create a T-junction between the two lines. The target was a rotated T, thus the rotated L stimuli were less similar to the template than the t-joined L stimuli. The similarity relationship was verified by measuring the search slopes for displays only containing one of the two types of L distractors, and by a multi-dimensional scaling analysis. Furthermore, we varied whether or not the display contained lure stimuli (i.e., distractors that are sufficiently dissimilar to the target as not to require attentive scrutiny). Supporting previous findings from our lab, participants successfully "screened out" the lures, and the cost to do so was unaffected by the type of L-distractors in the display. Thus, participants can guide focused spatial attention towards candidate stimuli. However, the results also indicated that participants did not guide attention preferentially towards more similar candidate stimuli (shifted Ls) than the less similar ones (rotated Ls). These findings are consistent with our proposal that target-distractor similarity does not play a role in the ranking of to-be-scrutinized distractors in a scene (Buetti et al, submitted). The results also indicate that once peripheral stimuli are sufficiently similar to the target, search slopes reflect template-comparison processes, rather than differences in attentional guidance.
Meeting abstract presented at VSS 2016