Abstract
Visuomotor tasks often require estimating how fast one object moves relative to another. To study such judgments, and the eye movement strategies, subjects saw two discs moving toward a common meeting point (MP) on a diamond-shaped "traffic circle" (overhead view). Discs started moving either from the top or from below the bottom of the diamond. Performance in a baseline velocity discrimination task (reporting which disc would arrive at the MP first) was precise (6-7%), with no bias toward either disc. A more challenging "collision" task required a 3-category judgment: top disc first vs. bottom disc first vs. collision. Collisions were defined as any overlap of the discs at the MP. Overlap occurred when disc velocities differed by < 6.25%. Velocity discrimination in the collision task was as precise as in the baseline. However, category boundaries were inaccurate, with large biases to report "collision". Collisions were reported for velocity differences up to 10%, despite feedback on each trial. Results were similar when the task required a motor action, namely, stop the bottom disc if a collision seemed likely. Reaction times were longest in the collision task due to the additional response categories. Eye fixation remained near the MP in all tasks. Smooth pursuit was rare until after the decision was made, when the discs moved as a pair. These results show that visual system can support more complex motion judgments, such as adding response categories and evaluating the magnitude of velocity differences, without loss of precision. The observed biases to favor "collision" judgments in the 3-category task may have been a response to high levels of uncertainty when velocity differences were close to the collision boundary. Such conservative criteria may be useful in real-world equivalents of the task.
Meeting abstract presented at VSS 2018