Abstract
Two visual search experiments tested the effects of visual crowding using traditional feature singleton search, by varying the spacing between lure objects and the lure-target similarity. Traditionally, search for a feature singleton has been characterized as being independent of set size and produces flat search slopes < 10 ms/item (Wolfe, 1998). However, recent results from our lab suggest search for a feature singleton has a logarithmic relationship between reaction time and set size that is modulated by the lure-target similarity (Buetti et al., submitted). These results have been interpreted as resulting from the first stage of visual processing that is exhaustive, unlimited-in-capacity and resolution limited. Items sufficiently dissimilar to the target are rejected by stage-one processing and items sufficiently similar to the target are inspected with focused attention. Here we ask if the limitation in resolution is a result of crowding, and what the effect of crowding is in the displays producing these logarithmic search slopes. In two experiments, we tested two possible display arrangements differing in the crowding they produced: in the Grid arrangement, spacing was independent of eccentricity and crowding was likely, whereas in the Concentric arrangement, crowding was minimized by increasing inter-item spacing as a function of eccentricity in accordance with Bouma's Law (Pelli, 2008). The results from both experiments converge on the same pattern of results: reaction times increased logarithmically with set size and the logarithmic slope was modulated by lure-target similarity for both spatial arrangements. Further, in both experiments, the Concentric displays produced faster reaction times than the Grid displays, consistent with an additive effect of crowding on reaction times. No interactions were found. These results provide strong evidence against the argument that crowding is a limiting factor in feature search nor is crowding responsible for the logarithmic nature of the reaction time slopes.
Meeting abstract presented at VSS 2016