Detecting glaucomatous damage: evaluation with contrast independent tasks

Abstract

Several studies have indicated that contrast-dependent tasks, such as detecting threshold stimuli, may be more effective in discriminating glaucoma from normal groups if external noise is added to the stimulus. We sought to determine if the same were true for contrast-independent tasks. Subjects were 15 patients with glaucoma and a group of 20 normals with the same mean age. We employed two contrast-independent tasks, orientation defined texture and dot numerosity discrimination. The stimulus was presented on a computer controlled video monitor. One side of the display contained a standard and the other contained a non-standard target. For each task, noise was added by perturbing the main feature of the display, dot number or line orientation, by a Gaussian distribution truncated at 2 standard deviation units. There were four noise levels for each task. Subjects viewed the stimulus display and made a spatial two-alternative forced choice judgment. Subjects judged the side of the orientation texture which contained a sub-region with a different angle and the side of the dot figures which had the larger number of dots. Glaucoma patients performed more poorly than normals in discriminating the orientation texture (p < 0.05) and in judging dot numerosity (p < 0.05). This was true even in the absence of added external noise. Adding external noise did not increase the differences between glaucoma patients and normals. Unlike contrast-dependent tasks in which the differences between normal and glaucoma patients are increased when external noise is added, contrast-independent tasks show maximal differentiation between the two groups without added noise. Tasks such as texture discrimination and dot numerosity may be useful in detecting glaucoma.