Fig. 3: Task, results (n = 36 participants), and model predictions of Experiment 1b: tracking two independent objects in the imagination.

Circles and triangles indicate mean responses for different true impact time and response order (‘1st ball’ refers to the ball participants responded to first, and ‘2nd ball’ to the ball they responded to second), error bars show standard error of the mean (SEM), colored lines show best linear fit, shaded area is 95% confidence interval (CI), and dotted line shows hypothetical perfect performance (where the subjective impact time equals the true impact time), as reference. Both models made similar predictions for the first response. For the second response, the Parallel Model predicted a minimal delay, due to perceptual noise. The Serial Model predicted a large delay for the second response, due to ending the simulation of the first object before turning to the second.