Abstract
False Pop Out (FPO) arises when a distractor stimulus poses as a target in a visual search display with a singleton target among homogeneous distractors. Previous demonstrations of FPO focused on odd-quadrant displays wherein grouping of items in the display yields emergent features such as symmetry that are ‘broken’ by a distractor, leading to the incorrect identification of that distractor as the target, i.e., as the odd item. For example, in a display of 3 elements like XOX, observers are unlikely to identify one of the Xs odd, but in the equivalent display ()(, they seem more likely to pick the final "(" as odd, because the first two elements group together, leaving the third perceptually isolated. Our current research demonstrates FPO with novel configurations, including displays in which proximity between stimuli is decreased to reduce grouping, displays with rotated quadrants, and linear displays with 3 items arranged in a row). Here FPO is still found to result from configural disruption rather than from any differences in basic featural properties. We propose a metric for quantifying FPO based on non-uniform error distributions across the distractors, such that one of the distractors receives far more responses than any other distractor or even than the actual target. For that distractor, FPO is calculated as a proportion of total error: FPO = ErrorRatedistractor / ErrorRatetotal Our results suggest that conventional pop out does not result from basic feature differences between a target and its distractors but rather from the ‘bustication’ (disruption) of salient emergent features of the entire, global display. We conclude by demonstrating pop out in heterogeneous displays; despite the fact that all of the items in the display are different, one pops out, and it is the one item that breaks the symmetry or other salient feature of the entire display.
Meeting abstract presented at VSS 2012