RESUMO
Six normal humans experienced yaw axis steps of velocity at 120 degrees /s in the dark. During the post-rotary period, subjects either had a null-task (do nothing); an ocular motor task (forced convergence: crossing the eyes); or a visual task (fixating a head-stationary target against a background of 10 degrees light/dark bars). Tasks started at 3 s post-rotation, and lasted either 2, 5, 10, or 15 s. Ocular motor and visual tasks were tested on different days. Five repetitions of each task duration were recorded for each subject. A mean VOR gain of 0.52 was observed, which did not vary with experimental conditions. Both convergence and fixation markedly suppressed nystagmus; in fact, the VORs obtained with the two different tasks are superficially similar in appearance. However, mean null-task time-constants were 9.4 s for convergence days, but 8.4 s for fixation days, and there was a small but significant reduction in overall null-task VOR amplitude on fixation days. Also, post-convergence slow-phase velocities were slightly enhanced, while post-fixation velocities were significantly reduced. The time-constant of velocity storage was found to be 10.1 s for convergence responses and 8.2 s for fixation responses. These differences can be understood in terms of modifications in central velocity storage during visual fixation which do not occur with convergence. The mean fixation data were analyzed in the context of a VOR model well-established for monkey data. With appropriate choice of parameters, this model accurately reproduces most features of the human data. An estimate for the human cupula time-constant of 3.3 s is obtained. Compared with the monkey, fixation suppression is greater and post-fixation velocity reduction less. Retinal slip alone accounts well for this; "velocity dumping" by an integrator shunt must be slight if present at all. The model correctly represents the post-fixation VOR for all durations of fixation.
