Spatial Tuning of Visual Responses to Symmetries in Textures

Symmetry is a biologically significant visual feature that relies on the perceptual grouping of spatially separate elements. Symmetry has been shown to play a role in numerous domains of visual perception in both humans and other animals. Brain imaging studies have revealed that several regions in the visual cortex exhibit robust and precise responses to symmetry. Here we explored the spatial mechanisms that mediate symmetry perception by measuring Steady-State Visual Evoked Potentials using high-density EEG. Our stimuli were taken from a class of regular textures, known as wallpaper groups – a set of 17 unique combinations of symmetry types that represent the complete set of symmetries in 2D images. We focused on groups PMM, which contains reflection symmetry, and P4, which contains four-fold rotation symmetry, and generated exemplars from each group based on log-domain band-limited random noise patches. This approach allows us to manipulate both the spatial frequency content of the exemplars, and the scale of the lattice structure that is repeated to tile the plane in all wallpaper groups. Across 8 conditions, we varied spatial frequency between 1 to 8 cycles-per-degree (centre frequency of the noise patch), and the lattice scale between 1/12 and ½ (relative size of lattice to overall wallpaper). Consistent with previous studies we found that symmetry-specific responses were weaker overall for rotation compared to reflection. However, responses also exhibited clear evidence of spatial tuning, with low spatial frequency and small scale lattices generally producing the biggest responses for both wallpaper groups. Interestingly, reflection (PMM) and rotation (P4) symmetry elicited clearly distinct response patterns across the spatial frequencies and lattice scales, suggesting that the two symmetry types rely on distinct cortical mechanisms. Future studies will relate these findings to responses in distinct areas along the human visual processing hierarchy.