Figure 6.

IQR ratio. (a–c) We calculated the IQR of final hand positions along movement extent for each condition. Here, we show the IQR ratio between conditions (y-axis) for participant data (coloured) and best-fit model simulations (dark grey) for each experiment. Hollow circles represent individual data. Solid circles represent the mean IQR ratio. Box and whisker plots represent 25th, 50th and 75th percentiles. An IQR ratio greater than one (dashed grey line) indicates greater movement variability (a,b) along the task-redundant condition or (c) the 80% probability of reinforcement condition. Participants displayed an IQR ratio greater than one in (a) Experiment 1 (light blue) and (b) Experiment 2 (dark blue). Thus, they had greater variability in their final hand positions during the task-redundant conditions compared to the task-relevant conditions. Participants in (c) Experiment 3 (purple) did not display an IQR ratio significantly greater than one. This suggests that participants showed similar levels of movement variability between the 80 and 20% probability of reinforcement conditions. Our best-fit model (Model 4, dark grey) replicated the observed data in (a) Experiment 1, (b) Experiment 2 and (c) Experiment 3.