Grouping by similarity is mediated by feature selection: evidence from the failure of cue combination.

As one of the most classic gestalt phenomena, grouping is often considered to occur preattentively. Therefore, it seems natural to expect conflicting grouping cues to combine preattentively and to lead to an impression of "no grouping." On the other hand, a recent account suggests that grouping by similarity is mediated by the attentional selection of a feature. This account makes the distinctive prediction that for grouping by similarity, but not for low-level groupings (e.g., connectedness, common region), when two conflicting grouping cues are present, the grouping structure will tend to be driven either by one type of cue or by the other type of cue and will usually not lead to the impression of "no grouping." This predicted pattern of results was confirmed in the present study: the presence of two conflicting low-level grouping cues led to reports of "no grouping" in 73.0 % of the trials, but for two similarity grouping cues, "no grouping" was only reported in 25.5 % of the trials. In summary, although low-level grouping cues do indeed work together to determine the perceptual structure preattentively, grouping by similarity is likely mediated by the attentional selection of a feature.