Shifting fixation to another pursuit target: selective and anticipatory control of ocular pursuit initiation.

Abstract The retinal input variable controlling angular velocity of smooth ocular pursuit is velocity of image movement. If fixation is shifted between targets with different visual velocities, the pursuit velocity increment required to stabilize the new target's image with respect to foveal retina equals the angular velocity of its movement over peripheral retina before the shift. Full stabilization of the new target's image could be achieved more promptly if presaccade image velocity were used to set postsaccade pursuit velocity, but pursuit initiation could be effectively controlled from peripheral retina in this manner only if conflicting velocity information from other parts of the retina were excluded. Eye movements were recorded as subjects watched a display containing both stationary and moving objects: The recordings showed that each time the subject shifted fixation from a stationary to a moving target, a substantial fraction of the final pursuit velocity had already been attained with 20 msec after termination of the saccadic movement. Since the input-output lag of the pursuit control system is of the order of 100 msec, it follows that the input used to set the new pursuit velocity had impinged upon the retina before the saccade. This pattern of pursuit initiation was not altered when another, contrasting stimulus-velocity was added. The results indicated that a small nonfoveal region of the retina may provide the dominant controlling input for saccade-associated pursuit initiation even in the presence of conflicting velocity information from the fovea or from other parts of the retina.