Stereopsis is perturbed by vergence error.

Fixation disparity or vergence noise produce instantaneous vergence errors. These errors are analogous to the imposition of a pedestal disparity, which is known to elevate stereothresholds. In this study, stereothresholds were measured as a function of induced vergence errors in subjects with normal binocular vision. Stereo half-images were viewed in the dark through a custom mirror haploscope. Vergence constant error and vergence variability were induced by introducing horizontal disconjugate position offsets in a pair of moveable mirrors within the haploscope, resulting in forced vergence demands of 6(Delta) base-in to 12(Delta) base-out. In addition, vergence variability was simulated by producing oscillatory disconjugate retinal image motion via motion of the moveable mirrors. The motion of the mirrors was either sinusoidal (frequency=2-6 Hz) or random, with peak-to-peak amplitudes of 0 degrees -0.5 degrees per eye. Stereopsis worsened systematically with induced or simulated VV that exceeded approximately 1.5' disparity. The results were similar regardless of whether the vergence error was induced by forced vergence or was simulated by periodic or random disconjugate retinal image motion. Stereothresholds were invariant with the frequency of disconjugate oscillation, within the range of frequencies and amplitudes tested. Hence, the simulated vergence velocity is not the essential factor that limits stereopsis within Panum's fusional area. The results indicate that the stereothreshold is elevated if the vergence error exceeds a critical value.