Abstract
Purpose: Studies of the central performance drop (CPD) have shown improved discrimination of textures, comprising left and right oblique lines, as the disparate texture is moved away from the fovea (e.g. Kehrer, 1987). One account of this effect is that processing is faster in the periphery than at the fovea. Therefore, when performance is limited by a backward mask, segmentation processes may have more time to operate in the periphery than at the fovea, leading to impaired central performance (Kehrer, 1987, Spatial Vision). A second account explains the CPD in purely spatial terms (Gurnsey, et al., 1996, JEP:HPP). The “processing speed” account argues that a backward mask is critical to the CPD whereas the spatial account does not. Morikowa (2000, Vision Research) showed that the CPD was not elicited when performance was limited by a simultaneous mask, thus seeming to support the processing speed proposal.
Methods: We employed a different method of degrading performance without a backward mask. Textures always comprised line segments having mean orientations of 45 and 135 degrees. Texture discrimination was made more difficult by introducing variability about these mean orientations. We refer to this variability as orientation bandwidth. In one task we determined threshold orientation-bandwith as a function of eccentricity. In a second task we measured performance in a four-alternative forced choice task, for a fixed orientation bandwidth, at a range of eccentricities.
Results: In both conditions sensitivity increased significantly as the target was moved from the fovea to about 5.0 to 8.5 degrees.
Conclusion: Consistent with the spatial account of Gurnsey et al. (1996), the CPD can be elicited without a backward mask. We believe that Morikowa's failure to find a CPD with a simultaneous mask reflects the fact that his mask was simply less effective at the fovea than in the periphery.
This project was supported by NSERC Grants to Rick Gurnsey