Abstract
Goal. Contrast-discrimination thresholds are much higher in the presence of high-contrast distractors, consistent with a max-pooling model of sensory selection (Pestilli et al., 2011). We investigated whether this is a general property of sensory selection. Methods. Observers viewed peripheral gratings, restricted to circular apertures left and right of fixation (6째 eccentricity, 5째 diameter). Stimuli were presented simultaneously in both apertures, in two intervals (600 ms stimulus intervals, 200 ms ISI). In the contrast-discrimination experiment, observers judged whether target contrast was higher in the first or second interval, while the distractor contrast was unchanged across interval. In the orientation-discrimination experiment, observers judged whether target orientation changed clockwise or counterclockwise, while the distractor orientation was unchanged across interval. A post-cue (400 ms after the second interval) indicated target location. In both experiments, target and distractor pedestal contrasts were systematically varied (all pairings of 10, 20, 40 and 80% contrast, excluding from the contrast-discrimination experiment conditions with identical target and distractor contrasts). Conditions were presented in randomly shuffled order, and target and distractor pedestal orientations were randomly varied across locations and trials. Results. Contrast-discrimination thresholds were much higher in the presence of high-contrast distractors e.g., relative to the effects of lower contrast distractors, 80% contrast distractors overwhelmingly disrupted performance for 10, 20 and 40% contrast targets. For orientation discrimination, however, disruption from high-contrast distractors was limited to the lowest target contrast (10%) condition, and was evident for only some observers. Conclusion. The max-pooling model posits that the largest sensory responses dominate perceptual decisions. This model can explain how sensory signals are pooled across spatial locations for 2IFC contrast discrimination, but not for an orientation-discrimination task with identical stimuli. This might reflect fundamental differences between encoding and/or visual short-term memory (across ISI in a 2IFC task) for contrast and orientation. Pestilli et al., Neuron, 72: 832-846, 2011.
Meeting abstract presented at VSS 2014