Binocular Function in Early Glaucoma

Abstract

This study investigated whether certain binocular mechanisms are disrupted in early glaucoma. Glaucoma patients, suspects, and normals were tested on a battery of psychophysical tests consisting of flicker sensitivity (5 and 34 Hz), temporal cut-off frequency (CFF), contrast sensitivity (Pelli-Robson chart), and stereoacuity. Monocular channels were evaluated with tests of monocular flicker performance and spatial contrast sensitivity. Binocular summation on spatial and temporal tests was used to reflect integrity of binocular neural interactions. Stereoacuity was taken as a measure of performance of disparity processing mechanisms. The groups differed in terms of binocular flicker sensitivity at both temporal rates, binocular and monocular peak contrast sensitivity, and stereoacuity. Binocular summation of both spatial and temporal sensitivity was normal. The glaucoma suspect group was distinguishable from the age-matched normal group on binocular contrast sensitivity and stereoacuity. The binocular mechanisms that mediate stereoacuity appear to be heavily disrupted, whereas the binocular mechanisms that mediate central neural interaction of monocular inputs are normal. Although monocular spatiotemporal abilities are disrupted, the binocular processes combine the monocular input normally. In addition, our results suggest a benefit of binocular testing for routine assessment of glaucoma patients. The profound disruption of stereoacuity appears to result from disorder in the spatial sampling array at the ganglion-cell level similar to the disorder reported in the normal periphery and the central retina of strabismic amblyopes. These and previous findings were reviewed to evaluate the supposition of preferential M-pathway disruption in early glaucoma. Such a model can not be reconciled with the present findings. We conclude that measurements of temporal modulation sensitivity fit well with such a model, but that the current evidence of spatiotemporal contrast sensitivity disruption is less supportive of such a model.