Abstract
Goodwin & Fender (1973a,b) studied smooth pursuit of trajectories composed of a predictable sinusoid and an unpredictable noise signal. When these signals are applied in orthogonal directions, the sinusoidal component of the motion is pursued with a latency approaching 0 (perfect prediction). Here we examine how this decomposition of the trajectory is influenced by the spatial content of the supporting pattern, to provide insights into low-level motion computations. Square-wave plaid patterns (1 cpd, 50% contrast) were viewed through a dual-Purkinje eye-tracker. The eye-tracker's stimulus deflectors were used to apply a sine/noise trajectory to the pattern, which moved behind a stationary circular aperture (diam. = 10 degrees). Subjects attempted to track the pattern, and pursuit latency was computed by correlogram analysis (50 trials per condition). Eight conditions were run, consisting of 4 trajectory orientations (0,45,90,135) crossed with 2 plaid orientations (0,45). We replicated the results of Goodwin & Fender, finding predictive pursuit of the sinusoidal component (latency 10–40 msec), with longer latencies for the unpredictable component (110 msec). Some subjects reported that in the aligned condition the components were less likely to “cohere,” often appearing to slide over one another. No effect of plaid/trajectory alignment or motion direction was observed in the latencies to the noise component, but predictive latency for an aligned plaid was approximately 10 msec faster than that for either an unaligned plaid or a simple spot. A larger effect was found for direction of motion, with predictive pursuit in the vertical direction having a latency about 20 msec shorter than the horizontal or oblique directions. The results suggest that pursuit is driven by pure “pattern” motion with little or no influence from “component” motion signals. The shorter predictive latencies for vertical motion may be related to the relatively weak reflexive pursuit in the vertical direction.
Supported by the Airspace Operations Systems (AOS) project of NASA's Airspace Systems program, and EY-RO1-12986 to SBS.