Abstract
Anticipatory smooth pursuit eye movements are faster when cues to future target motion are consistent with Newtonian mechanics (Ladda et al., 2007; Badler et al., 2010; Santos & Kowler, 2017), analogous to the role of Newtonian mechanics in other visuomotor behaviors (Deeb et al., 2021). We examined the role of cues to an impending collision in a design that allowed an assessment of learning. Displays consisted of a centrally-located target and a “launcher” in a randomly chosen corner. The launcher moved toward one of two points of collision with the target. The post-collision path of the target depended on the point of collision. The direction of target motion after the collision was either consistent with (Newtonian condition) or opposite to the Newtonian prediction (non-Newtonian condition). A third (neutral) condition tested the same target paths while the launcher remained stationary. Conditions were run in separate blocks allowing equivalent opportunity for accurate prediction and learning of the target’s path on the basis of the launcher path. Pursuit in both Newtonian and non-Newtonian conditions showed anticipation. Eye velocity was higher in the Newtonian condition, and the direction of anticipatory pursuit corresponded to the direction predicted by Newtonian mechanics Anticipatory pursuit was also faster in the non-Newtonian than in the neutral condition, showing a role for learning in the absence of physically-realistic cues. The results show that both natural physical cues and learning can facilitate anticipatory pursuit. Cues that are compatible with normal physical events are more effective. Results also show that encoding of the launcher path is precise enough to allow pursuit to discriminate between two post-collision paths of the target. Future work will investigate finer discriminations and more complex paths in order to understand the limits on the representations of physical principles that are able to guide anticipatory smooth pursuit.