Eye-head movement coordination: vestibulo-ocular reflex suppression with head-fixed target fixation

To maintain clear vision, the images on the retina must remain reasonably stable. Head movements are generally dealt with successfully by counter-rotation of the eyes induced by the combined actions of the vestibulo-ocular reflex (VOR) and the optokinetic reflex. A problem of importance relates to the value of the so-called intrinsic gain of the VOR (VORG) in man, and how this gain is modulated to provide appropriate eye movements. We have studied these problems in two situations: 1. fixation of a stationary object of the visual space while the head moves; 2. fixation of an object moving with the head. These two situations were compared to a basic condition in which no visual target was allowed in order to induce "pure" VOR. Eye movements were recorded in seated subjects during stationary sinusoidal and transient rotations around the vertical axis. Subjects were in total darkness (DARK condition) and involved in mental arithmetic. Alternatively, they were provided with a small foveal target, either fixed with respect to earth (earth-fixed target: EFT condition), or moving with them (chair-fixed-target: CFT condition). The stationary rotation experiment was used as baseline for the ensuing experiment and yielded control data in agreement with the literature. In all 3 visual conditions, typical responses to transient rotations were rigorously identical during the first 200 ms. They showed, sequentially, a 16-ms delay of the eye behind the head and a rapid increase in eye velocity during 75 to 80 ms, after which the average VORG was 0.9 +/- 0.15. During the following 50 to 100 ms, the gain remained around 0.9 in all three conditions. Beyond 200 ms, the VORG remained around 0.9 in DARK and increased slowly towards 1 or decreased towards zero in the EFT and CFT conditions, respectively. The time-course of the later events suggests that visual tracking mechanisms came into play to reduce retinal slip through smooth pursuit, and position error through saccades. Our data also show that in total darkness VORG is set to 0.9 in man. Lower values reported in the literature essentially reflect predictive properties of the vestibulo-ocular mechanism, particularly evident when the input signal is a sinewave.