The effect of binocular eye position and head rotation plane on the human torsional vestibuloocular reflex

AbstractWe examined how the gain of the torsional vestibulo-ocular reflex (VOR) (defined as the instantaneous eye velocity divided by inverted head velocity) in normal humans is affected by eye position, target distance, and the plane of head rotation. In six normal subjects we measured three-dimensional (3D) eye and head rotation axes using scleral search coils, and 6D head position using a magnetic angular and linear position measurement device, during low-amplitude (∼20°), high-velocity (∼200°/s), high-acceleration (∼4000°/s2) rapid head rotations or ‘impulses.’ Head impulses were imposed manually and delivered in five planes: yaw (horizontal canal plane), pitch, roll, left anterior-right posterior canal plane (LARP), and right anterior-left posterior canal plane (RALP). Subjects were instructed to fix on one of six targets at eye level. Targets were either straight-ahead, 20° left or 20° right from midline, at distance 15 or 124cm from the subject. Two subjects also looked at more eccentric targets, 30° left or 30° right from midline. We found that the vertical and horizontal VOR gains increased with the proximity of the target to the subject. Previous studies suggest that the torsional VOR gain should decrease with target proximity. We found, however, that the torsional VOR gain did not change for all planes of head rotation and for both target distances. We also found a dynamic misalignment of the vertical positions of the eyes during the torsional VOR, which was greatest during near viewing with symmetric convergence. This dynamic vertical skew during the torsional VOR arises, in part, because when the eyes are converged, the optical axes are not parallel to the naso-occipital axes around which the eyes are rotating. In five of six subjects, the average skew ranged 0.9°–2.9° and was reduced to <0.4° by a ‘torsional’ quick-phase (around the naso-occipital axis) occurring