Perceptual grouping allows for attention to cover noncontiguous locations and suppress capture from nearby locations.

A salient stimulus may interrupt visual search because of attentional capture. It has been shown that attentional capture occurs with a wide, but not with a small attentional window. We tested the hypothesis that capture depends more strongly on the shape of the attentional window than on its size. Search elements were arranged in two nested rings. The ring containing the search target remained fixed, while a salient color singleton occurred either in the same or in the other ring. We observed that color singletons only disrupted search when shown in the same ring as the search target. It is important to note that, when focusing on the outer array, which presumably required a larger attentional window, singletons on the inner array did not capture attention. In contrast to the original attentional window hypothesis, our results show that attentional capture does not always occur with a large attentional window. Rather, attention can be flexibly allocated to the set of relevant stimulus locations and attentional capture is confined to the attended locations. Further experiments showed that attention was allocated to search elements that were perceptually grouped into "whole" or "Gestalt"-like objects, which prevented attentional capture from nearby locations. However, when attention was allocated to noncontiguous locations that did not form a perceptual Gestalt, nearby locations elicited attentional capture. Perceptual grouping could be based on a combination of color and position, but not on color alone. Thus, the allocation of attention to Gestalt-like objects that were jointly defined by similarity and proximity prevented attentional capture from nearby locations.