Abstract
Clutter impedes object identification in peripheral vision. “Crowding zone” refers to the spatial extent surrounding a peripheral target within which the presence of a flanker impedes target identification. Several methods are commonly used to measure the crowding zone, but these methods do not necessarily measure the same quantity. When luminance contrast threshold is measured as a function of target-flanker spacing, a common approach is to fit the threshold vs. spacing function with a pair of lines: a line of negative slope intersecting with the horizontal line that represents the contrast threshold for identifying an isolated target. The target-flanker spacing at the intersection is taken as the spatial extent of crowding. When accuracy is a dependent variable, a comparably defined “full” spatial extent of crowding is the target-flanker spacing associated with the upper elbow of the psychometric function – spacing beyond which identification accuracy is the same as that for an isolated target. This, however, is not the common practice. Rather, most studies that use accuracy as a dependent variable define the crowding zone as the spacing associated with a specific level of accuracy somewhere at the middle of the psychometric function (e.g. 75% correct). This “partial” crowding zone is related to the “full” crowding zone by the slope of the psychometric function. We tested 11 subjects, each at 5 retinal locations and 2 eccentricities, for up to 6 different combinations of letter-like targets and flankers. We found while the full and partial spatial extent of crowding did correlate, there were a great deal of variability between the two, indicating that the slope of the psychometric functions varied significantly across conditions. We derived an efficient method for estimating the full crowding zone based on the adaptive psi method (Kontsevich & Tyler, 1999; Prins, 2013), rendering any reliance on partial estimation unnecessary.
Meeting abstract presented at VSS 2015