Abstract
Preattentive visual representations have been studied with visual search, and it is known that both primitive features and configurations of features can be represented preattentively. Rensink and Enns (1995) presented evidence for preattentive grouping, in which the groups act as a single functional unit. They proposed that proximity allows for grouping which is insensitive to contrast polarity (CP). Later on, Gilchrist, Humphreys, Riddoch & Neumann (1997) found evidences of CP-insensitivity only for colinearity but not for proximity. In this study, two experiments were carried out to address this issue. In Experiment 1, participants searched for a half-arrow among distractors formed by the target and another opposite pointing arrow, with either its same or opposite CP, separated by a small gap. Searches were efficient when distractor arrow pairs had the same CP, but were inefficient when the CP was different. As colinearity cues were present in distractors, this result contrasts with previous suggestions. In Experiment 2, arrows were replaced with circles and rectangles; results were consistent with Gilchrist et al., rather than Experiment 1. Another experiment was carried out to investigate CP-sensitivity in containment groups. In Experiment 3, participants were asked to search for: A: a short line among long lines; B: a long rounded rectangle among short ones, containing an equal length line; C: a short line among long lines, surrounded by equal length rounded rectangles. The rectangles and lines were of either the same or opposite CP. Search was efficient for conditions A & B but not C, implying that the lines and rectangles were grouped; unlike in Experiments 1 & 2, search efficiency did not vary with same or different CP. Taken together, these experiments suggest that preattentive grouping is CP-sensitive when defined by proximity, CP-insensitive when defined by containment, and the effect of colinearity may be dependent on stimulus forms.