Abstract
Gaze pursuit is the coordinated movement of the eyes and head that allows humans and other foveate animals to track moving objects. The control of smooth pursuit eye movements when the head is restrained is relatively well understood, but how the eyes coordinate with concurrent head movements when the head is free remains unresolved. In this study, we describe behavioral tasks that dissociate head and gaze velocity during head-free pursuit in monkeys. Existing models of gaze pursuit propose that both eye and head movements are driven only by the perceived velocity of the visual target and are therefore unable to account for these data. We show that in addition to target velocity, the positions of the eyes in the orbits and the retinal position of the target are important factors for predicting head movement during pursuit. When the eyes are already near their limits, further pursuit in that direction will be accompanied by more head movement than when the eyes are centered in the orbits, even when target velocity is the same. The step-ramp paradigm, often used in pursuit tasks, produces larger or smaller head movements, depending on the direction of the position step, while gaze pursuit velocity is insensitive to this manipulation. Using these tasks, we can reliably evoke head movements with peak velocities much faster than the target's velocity. Under these circumstances, the compensatory eye movements, which are often called counterproductive since they rotate the eyes in the opposite direction, are essential to maintaining accurate gaze velocity.