RAPID COMMUNICATION Primate Head-Free Saccade Generator Implements a Desired (Post- VOR) Eye Position Command by Anticipating Intended Head Motion

Crawford, J. Douglas and Daniel Guitton. Primate head-free nisms that distribute this command into separate but coordisaccade generator implements a desired (post-VOR) eye position nated control signals for the eye and head remain in dispute. command by anticipating intended head motion. J. Neurophysiol. The highly coupled eye-head gaze saccades of cats have 78: 2811–2816, 1997. When we glance between objects, the brain been adequately modeled by essentially parsing gaze error ultimately controls gaze direction in space. However, it is currently to the eye and head as a matter of different weightings (Galiunclear how this is allocated into separate commands for eye and ana and Guitton 1992; Guitton et al. 1990). However, addihead movement. To determine the role of desired final eye position tional position signals appear to be required to account for commands, and their coordination with intended head movement, the gaze behavior of primates and humans. These signals we trained three monkeys to make large gaze shifts while wearing may come into play in determining both the behavioral limit opaque goggles with a monocular 87 aperture. Animals eventually developed a new set of context-dependent eye-head coordination of the size of the saccade (Guitton and Volle 1987), and strategies, in particular expanding the head range and compressing determining the initial proportion of drive to the eye and the eye-in-head range toward the aperture (while wearing the goghead (Becker and Jürgens 1992; Freedman and Sparks 1997; gles) . However, when we shifted the location of the aperture to a Fuller 1996; Goossens and Van Opstal 1997). To account different subsection of the normal head-free oculomotor range (by for such observations, Guitton and Volle (1987) proposed covering the original aperture and creating a new one), eye-head that eye motor error was computed by comparing initial eye saccades failed to acquire visual targets, because they continued position to a saturating desired eye position within a gaze to drive the eye ultimately toward the now occluded original aperfeedback loop. A further position-related problem arises in ture. Even when a head-stationary saccade acquired the new apercoordinating the initial eye saccade with the subsequent vesture, subsequent head-free saccades drove the eye eccentrically tibuloocular reflex (VOR) slow phase, where the direction toward a point that anticipated the intended head movement, such of the latter depends only on concurrent head movement. that the subsequent vestibuloocular reflex slow phase brought the eye onto the location of the original aperture. Animals could only This becomes a nontrivial problem in multidimensional gaze acquire the new aperture consistently after several days of retrainshifts, because the eye and head do not rotate in exactly the ing. These results suggest that 1) eye-head coordination is achieved same direction (Freedman and Sparks 1997; Glenn and Vilis by a plastic, context-dependent neural operator that uses informa1992; Guitton and Crawford 1994). In this case, for the eye tion about initial eye/head position and intended movement to to come to rest at an intended final eye-in-head position, compute desired combinations of final eye/head position and 2) the ocular saccade would have to be ‘‘preprogrammed’’ to acquisition of these positions involves sophisticated anticipatory compensate for slow-phase eye movements linked to the compensations for subsequent movement components, akin to intended head movement (Crawford and Vilis 1991; Guitton those observed previously in complex oral and manual behaviors. and Crawford 1994; Tweed et al. 1995). A recent threedimensional (3-D) model (Tweed 1997) incorporates all of