Lillian Hauser¹, Timothy Sweeny¹; ¹University of Denver

Redundancy masking studies have found that observers frequently underreport the number of items (e.g., lines and shapes) in a repeating pattern shown in the periphery. This phenomenon is speculated to compress the complexity of the visual world, simplifying processing demands. The cost of losing awareness of a few items may be tolerable for sets of geometric features, but it may be less desirable when perceiving socially important features like faces. Here we compared the strength of redundancy masking with four types of stimuli: upright faces, inverted faces, rectangles, and lines. We hypothesized that faces would be less susceptible to redundancy masking than these other features. On each trial, three to six radially arranged, evenly spaced (edge-to-edge: 1°) identical stimuli were presented in either the left or right periphery, centered 39° from fixation. Observers reported the number of stimuli they perceived in each set, ranging from zero to nine. Observers underestimated the number of rectangles the most (M = -0.39), followed by lines (M = -0.25) and inverted faces (M = -0.01); conversely, observers overestimated the number of upright faces in a set (M = 0.09). This distinction between upright and inverted faces is especially notable because it cannot be attributed to differences in visual complexity. Further, the upright faces' greater resistance to redundancy masking (compared with inverted faces) only occurred when sets were viewed in the left visual field, consistent with the right hemisphere’s dominance in face processing. These findings indicate that specialized facial processing systems may help combat redundancy masking.