Abstract
Aim: The role of color in mirror-symmetry perception is controversial: while some reports indicate that mirror-symmetry is color selective, others suggest that symmetry is merely sensitive to color correlations. Here we test between the two ideas as well as examine the role of attention in color-based symmetry perception. Method: Stimuli consisted of a fixed number of colored Gaussian blobs arranged either mirror-symmetrically or randomly. Stimuli contained either two (red, green), three (red, green, blue) or four (red, green, blue, yellow) colors, with respectively 50%, 33% and 25% blobs arranged symmetrically in the symmetric condition, the remaining blobs being randomly positioned and drawn equally from the remaining colors. Stimulus conditions were: (1) ‘segregated’ – symmetric blobs were just one color that was the same throughout the session; (2) ‘random-segregated’ – symmetric blobs were just one color, but the color was randomly selected on each trial; (3) ‘non-segregated’ – symmetric blobs were of all colors and in equal proportions; (4) ‘anti-symmetric’ – symmetric blobs were of opposite color across the symmetry axis. Perceptual conditions were: (1) the observer knew beforehand the color of the symmetric blobs in the segregated condition and was required to attend to that color; (2) the observer did not know the symmetry color and no attention was required. Results: We found: (a) no significant superiority in performance between the random-segregated and non-segregated conditions, giving no support to the idea that mirror-symmetry is color selective; (b) highest performance in the attention-to-segregated condition, suggesting that symmetry perception can benefit from color-based attention, and (c) near-chance levels for the anti-symmetric condition, suggesting that symmetry perception is sensitive to color correlations. Conclusion: Mirror-symmetry detection mechanisms, while sensitive to color correlations and subject to the benefits of attention-to-color, are not color selective.
Meeting abstract presented at VSS 2015