Abstract
We examined local and global effects in crowding. Crowding anisotropy was tested varying flankers' local orientation, global relations, and location. A high-contrast horizontal Gabor target was placed on the horizontal meridian at 2.5° eccentricity, flanked by two Gabor patches of same contrast, at 0°, 45°, or 90° relative to the horizontal meridian (referred to as radial, diagonal and tangential conditions, respectively). In each condition flankers were either collinear or parallel relative to each other (local orientations of 0°/90° in the radial and tangential arrangement, and 45°/135° in the diagonal arrangement). Crowding was quantified as an increase in log orientation discrimination thresholds in the presence of flankers. Crowding was insignificant with the tangential pairs ([[lt]]0.1), but was found with the diagonal (0.3 & 0.56), and the radial (0.55 & 0.12) pairs (collinear & parallel pairs, respectively). The difference between the collinear and parallel pairs' effects was independent of local orientations (0° & 45°). Thus, oriented Gabor stimuli show an 8 shape interaction zone rather than the elliptical used to describe the results of Toet & Levi (1992) with T stimuli. Additional results indicate an interaction of flankers' location, local orientations, and global relations in crowding by two flankers. The effects of these pairs, however, were found to be non-indicative of crowding produced by a variety of circular arrangements of flankers (as in Livne & Sagi 2007). We will describe results from such configurations suggesting that flankers' grouping is an important factor in crowding. Flankers modulate the effects of each other - remote flankers may exert their effect on a target via intermediate flankers, showing in some cases separation-independent crowding. By applying a quantitative model of crowding, based on basic gestalt grouping rules (proximity, similarity, continuity), to the obtained experimental results we were able to rank specific local relations (collinearity, co-circularity, etc) in respect to their combined effect.