Abstract
Purpose. Selecting and tracking a moving visual target requires coordination between pursuit and saccades. Because eye movements are preceded by shifts of attention, a likely mechanism for this coordination is the common role of attention in target selection by the two systems. Most studies have examined target selection using suddenly appearing stimuli, likely confounding attention-based selection with stimulus-evoked changes in attention. To eliminate this problem, we used a novel task that required target switching during steady-state tracking with minimal changes to the visual display. Methods. Human subjects (n=3) initially fixated a central fixation cross. Horizontal dot arrays (50 0.3 dg gray squares spaced 0.5 dg apart) were presented above and below fixation at one of three eccentricities (0.15, 0.3 or 0.6 dg). After presenting a color cue specifying the target, the arrays changed color (one to the cued color, both 34 cd/m^2) and began moving in opposite directions at 14 dg/s. On 50% of the trials, the arrays switched color during the trial but were otherwise unchanged. The subject's task was to track the array of the cued color; when the arrays switched color, subjects had to switch tracking to the other array. Results. The latency of pursuit to the target-switching averaged 233, 257 & 271 ms for the 3 subjects. The switch in tracking to the other array could occur in the complete absence of saccades, and the frequency of saccades accompanying the target-switching (25, 60, & 90%) depended on the vertical separation between the two arrays (0.3, 0.6 & 1.2 dg, respectively). Conclusions. Switching tracking to a new target does not require a saccade. The long latency for target-switching by pursuit likely reflects the slow time course associated with endogenous shifts of attention. The shorter latencies historically associated with pursuit may reflect the speeding of attentional shifts caused by visual transients typically included in pursuit paradigms.