Motion discrimination of single targets: comparison of preliminary findings in normal subjects and patients with glaucoma.

Abstract

Luminance, global motion and flicker sensitivities are affected in patients with primary open-angle glaucoma. Although no theoretical model has been put forward to explain the observed reduction in sensitivity in this patient group, these findings have often been attributed to diffuse and selective loss of large optic nerve fibres. Movement processing was investigated using an optical projection system that generates smooth, continuous motion with control of speed, displacement and motion direction. Motion-displacement and direction-discrimination thresholds were measured in eight normal subjects and in three patients with diagnosed glaucoma. At each speed tested, targets were presented for a range of displacements and thresholds were extracted after probit analysis. The measurements were carried out both foveally and at 19 degrees in the periphery and provided the data necessary to develop and optimise a model of motion perception based on multiple time delays for the correlation of signals that map progressively more distant parts of the visual field. Our preliminary findings show that direction discrimination can be at chance level even for large displacements when motion is detected 80% of the time. Model simulations show that specific changes in the spatial sampling interval and the speed of transmission of the motion signals involved can cause the observed reduction in motion sensitivity and direction discrimination in patients with glaucoma. A model for motion detection and direction discrimination of single targets has been proposed to account for the measured functional relationship between motion displacement thresholds and target speed in normal subjects. Tested patients with glaucoma show reduced motion sensitivity and poor discrimination of motion direction. The type of degraded performance observed experimentally in glaucoma patients is also predicted by the model. Such predictions require specific changes in model parameters that may be indicative of changes in the retina caused by the disease.