Abstract
Nelson and Palmer (2007) claimed that attention is automatically drawn to figures based on faster detection and discrimination responses for targets shown on the familiar side of a small set of bipartite, equal-area displays, each shown multiple times. We tested this “figural advantage” hypothesis using 8-region equal-area black and white alternating convex and concave displays, in which convex regions are highly likely (∼89%) to be seen as figure (Peterson & Salvagio, 2008). In Experiment 1 we used 40 displays, each repeated 16 times. Participants reported whether a masked target shown at variable SOA (0, 150,250, 500) for 80 ms on one of the regions adjacent to fixation was an “X” or a “Y”. At all SOAs discrimination tended to be better for targets on concave than convex regions, p [[lt]]0.02, indicating a perceptual advantage for the groundside rather than the figure side. In Experiment 2, we presented 24 displays, each once only, and included a set of 8 control stimuli with black and white equal-area regions delimited by straight edges. Participants reported whether an unmasked 100-ms target on one of the regions adjacent to fixation was a square or circle. Different groups saw targets appearing at 80-ms or 120-ms SOA. In both groups, subjects' reaction times were approximately the same for targets appearing on convex versus concave regions, although participants in the 80-ms SOA group were faster at reporting targets on concave regions than straight-edge regions in control displays, p [[lt]].03, again suggesting a ground advantage rather than a figure advantage. In an additional experiment, observers' discrimination responses were faster for targets on concave regions (grounds) rather than convex regions (figures). Thus, we obtained no evidence that attention is drawn to a convex figure near fixation. We suggest that Nelson and Palmer's results depend upon familiarity.
NSF BCS 0425650 & 0418179.