Fig. 3.
Results and simulations of the first experiment. Top row, metric model: A) mean final heading for humans and metric model as a function of the number of perturbed neighbors. Shaded regions represent the 95% CI of the human data. The error bars in plots A) and D) represent the 95% CI for the model. B) High-density condition: mean time series of heading for human data and metric model; curves represent the number of perturbed neighbors. C) Low-density condition: same. Bottom row, visual model: D–F) same data with visual simulations.