Bita Manouchehri¹, Stephanie M Shields¹, Kathryn Bonnen¹; ¹Indiana University

Humans are able to predict and follow the trajectories of moving targets. Previous research has studied manual tracking of targets moving in unpredictable Brownian trajectories, trajectories with abrupt changes in direction, and highly predictable sine wave trajectories. Here, we test the hypothesis that participants’ ability to follow a target depends systematically on the predictability of the target’s trajectory. We generated 3D pink noise (1/f) trajectories and manipulated their predictability by applying a bandpass filter with a fixed lower frequency and a variable upper frequency. We presented targets dichoptically using a PROPixx projector (VPixx Technologies), we recorded responses with a LeapMotion device (UltraLeap), which allowed participants to move a cursor in 3D by moving their hand. Their task was to follow the target’s motion with that cursor. We evaluated tracking performance by calculating cross-correlograms (CCGs) of the target and response velocities in the horizontal, vertical, and depth dimensions. We then quantified response latency by finding the latency of each CCG’s peak correlation. Overall, we found that participants’ response latency decreased as predictability increased, suggesting that more predictable trajectories were indeed easier for participants to follow. For the most predictable trajectories (with near-sinusoidal motion), response latencies were close to zero (0.038 seconds). Consistent with existing literature, we found that performance tended to be worse for motion-in-depth than for horizontal and vertical motion: Peak correlation values tended to be lower, and latencies tended to be greater. Notably, however, the motion-in-depth latency of multiple participants decreased to near-zero, essentially catching up to the horizontal and vertical latencies. Such an effect could potentially result from depth tracking operating over a longer temporal integration window. Hence, our results support the hypothesis that tracking performance improves as target trajectories become more predictable and suggest interesting directions for future research.