Emma Neto¹, Akilesh Sathyakumar¹, Isabella Crescenzi¹, Jessica Lim Tung Tseung¹, Allison B. Sekuler^1,2,3, Patrick J. Bennett¹; ¹Department of Psychology, Neuroscience & Behaviour, McMaster University, ²Rotman Research Institute, Baycrest, ³Department of Psychology, University of Toronto

Relative phase discrimination is performed using even-symmetric and odd-symmetric visual mechanisms (bar and edge detectors, respectively; Field & Nachmias, 1984). In central vision, even- and odd-symmetric mechanisms are approximately equally sensitive, but it is unclear how sensitivity varies with increasing stimulus eccentricity. Studies that measured discrimination thresholds with compound gratings (Bennett & Banks, 1987) found that sensitivity of odd-symmetric mechanisms decreased dramatically in peripheral vision; however, studies that used isolated bars and edges (Morrone et al., 1989) found that sensitivity of even- and odd-symmetric mechanisms was nearly equal across the visual field. These different results could be reconciled if lateral masking in the periphery was much greater for edge discrimination than bar discrimination. We tested this hypothesis by measuring bar and edge discrimination thresholds for various inter-feature distances. Stimuli were modified versions of the even- and odd-symmetric stimuli used by Morrone et al. and consisted of 1, 2, 4, or 8 light/dark bars or right/left edges. Discrimination thresholds were measured by varying stimulus contrast in a 2-IFC task. In four observers, increasing the number of features (i.e., decreasing the inter-feature distance) had virtually no effect on edge and bar discrimination thresholds in central vision. However, when stimuli were presented at an eccentricity of 7.6 deg, increasing the number of features dramatically increased edge discrimination thresholds but had virtually no effect on bar discrimination thresholds. These results are consistent with the hypothesis that peripheral edge discrimination is especially sensitive to lateral masking.

Acknowledgements: NSERC