3-DIMENSIONAL EYE, HEAD, AND CHEST ORIENTATIONS AFTER LARGE GAZE SHIFTS AND THE UNDERLYING NEURAL STRATEGIES

1. The fixation orientations adopted by the eye, head, and chest were examined when all three were allowed to participate in gaze shifts to visual targets. The objective was to discover whether there are invariant, neurally determined laws governing these orientations that might provide clues to the processes of perception and motor control. This is an extension of the classical studies of eye-only saccades that determined that there is only one eye orientation for each gaze direction (Donders' law) and that the rotations necessary to take the eye from a reference orientation to all other orientations adopted are about axes that lie in a plane (Listing's law). 2. The three-dimensional orientations of the static eyes, head, and chest were measured after each gaze shift to a visual target, the targets having been fixed at positions ranging from 0 to 135 degrees to the left and right of center and 45 degrees up and down. These measurements were taken of seven human subjects by means of the search coil technique with coils attached to the sternum, head, and right eye. Orientations were plotted as quaternion vectors so that those orientations obeying Donders' law formed a surface and those obeying Listing's law formed a plane. 3. The orientations adopted by the eye, head, and chest were found to be a small subset of those possible under the biomechanical and task-imposed constraints. Thus there is a neurally implemented restriction, specifically of the rotation of the eye relative to space (i.e., the orientation variable e(s)) and to the head (e(h)); also of the rotation of the head relative to space (h(s)) and to the chest (h(c)), and the rotation of the chest relative to space (c(s)). Plotted as quaternion vectors, the data for each orientation variable formed a characteristic surfacelike shape. In the case of e(s), h(s), and h(c) these were twisted surfaces, whereas for e(h) the surface was planar and for c(s) it was nearly linear. Thus to a first approximation each of the orientation variables conformed to Donders' law. 4. The eye adopted a pointing (gaze) direction that has the ratio of vertical to horizontal components generally greater than one when fixating each of the corner targets. The chest, by contrast, moved almost entirely in the horizontal direction, whereas the head performed an intermediate role. 5. The e(s)-, h(s)-, and h(c)-fitted surfaces and c(s)-fitted lines were tilted remarkably little from the vertical axis (i.e., the gravity direction) despite larger tilts being possible. This may reduce high inertia rotations and minimize energy expenditure against gravity. 6. The fixation orientations of the eye relative to head were found to be confined to a plane (Listing's plane) with no large, consistent departure. This suggests that Listing's law may be a strategy related to minimizing eye muscle stretch, a quantity more closely associated with e(h) than e(s). 7. We compared head orientations relative to space and to the chest. The standard deviations of the h(c) surface fits were lower than those of h(s), indicating that the former follows Donders' law more closely. The shape of the surfaces fitted to h(c) were more consistent between conditions with and without chest participation than were those of h(s). 8. Because of the simplification inherent in following Donders' law and having a surface shape of a single type, the comparison of h(s) and h(c) surfaces suggests that head orientations are calculated in a chest rather than spatial frame of reference. Utilizing a chest frame may be advantageous because h(c) can be more easily related to muscle stretch and the biomechanical limits of head movements than can h(s). 9. The eye relative to space does not closely adhere to Donders' law. Although approximately similar to a Fick surface, on closer examination e(s) shows significant, systematic variation oblique to the surface for a given gaze direction. Expressed as quaternion vectors, the e(s) orientations adopted when viewing a single target lay along a line. A similarly directed line can be calculated by plotting the e(s) orientations corresponding to an eye with the same mean gaze direction that is rotated by varying amounts about the line of sight. This indicates that, unlike eye orientations relative to the head, e(s) is not directly constrained by the CNS.