Two eccentricity-dependent limitations on subjective contour discrimination.

Eccentricity-dependent sensitivity losses in spatial discrimination tasks can often be overcome by scaling stimuli at each eccentricity by a factor F=1+E/E(2). However, because there may be more than one eccentricity-dependent limitation at play in a particular task a single scaling function may be insufficient to explain all sensitivity losses as stimuli are moved from foveal to peripheral retinal locations. We propose a method explicitly designed to determine whether a single scaling factor is sufficient to capture all eccentricity-dependent sensitivity losses in a task. The methodology was applied to subjective contour stimuli that varied in aperture size (sigma) and carrier wavelength (omega). For a range of stimulus configurations [2(-0.5)log(sigma/omega)] we measured threshold scale [2(-0.5)log(sigma omega)] and fit data at each eccentricity to rectangular parabolas that expressed sensitivity limitations arising from aperture size and carrier wavelength. Although a single scaling factor (E(2)) explains much of the variability in the data there are systematic sources of variance in the residuals (i.e., deviations of the data from the best fitting functions). Our analysis shows that two scaling factors are required to capture all eccentricity-dependent limitations in the data.