Search

We evaluated the effect of substitutive noise on contrast sensitivity within the context of linear (Fourier) and nonlinear (non-Fourier) visual processes. Orientation judgments for D6 (sixth spatial derivative of Gaussian) patterns were obtained from three visually normal subjects when random regions of the target and background were occluded by small (1.7 arc min) pixel arrays that were either all of the same contrast polarity or a mixture of equal percentages of negative and positive contrast. The target was presented either synchronously or asynchronously with the occluding elements. Our results indicate that the manipulation of noise characteristics in this way can bias performance either toward a nonlinear process that is insensitive to noise contrast polarity but sensitive to temporal asynchrony or toward a quasi-linear process that is sensitive to noise contrast polarity but insensitive to temporal asynchrony. These findings have relevance to models of the effect of spatial sampling on the visual performance of persons with retinal disease.

The purpose of the study was to determine whether spatial sampling has equivalent effects on visual performance for grating and letter stimuli, two optotypes that are frequently used in the clinical assessment of visual function. Test targets consisted of five-bar square-wave gratings, spatial D6 (sixth derivative of a Gaussian) patterns, or individual Sloan letters, presented for 255 ms in the center of an adapting field. Spatial sampling was introduced by replacing random regions of the targets with occluding elements that had the same luminance as the adapting field, so that only random samples of the test target were visible on each trial. The occluding elements consisted of 4 x 4, 8 x 8, or 16 x 16 pixel arrays (subtending 0.67, 1.33, and 2.67 arcmin, respectively), and the percentage of the test target that remained unoccluded ranged from 3% to 100%. Visual acuity and contrast sensitivity for the spatially sampled targets were measured in two visually normal subjects. Results were compared with the predicted effect of the reduction in effective contrast that was introduced by the sampling paradigm. For grating stimuli the effect of spatial sampling was equivalent to the reduction in effective stimulus contrast produced by the sampling, regardless of occluding-element size. For letter stimuli the reduction in effective stimulus contrast predicted most, but not all, of the loss of visual acuity and contrast sensitivity for sampled targets, and the deficit in performance increased when larger occluding elements were used.(ABSTRACT TRUNCATED AT 250 WORDS)