Figure 7

The velocity-dependent parameters reveal learning according to the subject’s somatosensory stochastic signatures in each training context. (A) The empirical frequency distribution of maximum speed across subjects (inset) and the MLE of shape and scale Gamma parameters for each subject, for the fast and slow instructed speed condition. (B) The normalized maximum speed parameter (invariant to possible allometric effects due to individual differences in limb sizes) aligns participants across the line of unity with experts at the far right symmetric range of the Gamma (more predictive power in their performance). Inset shows the empirical frequency distribution across subjects for this normalized parameter. (C) The individual learning progression for novices as they performed slow and fast versions of the jab. Notice that the stochastic signatures of their speed maxima shifts towards the right for the fast condition (more predictive) in some cases, whereas in other cases it is instead the slow condition which has this effect. Notice also that given the same number of repetitions, the rate of change is very small for some subjects and very large for others. This plot captures the individual’s learning progression and unveils which training context is most adequate to make the subject’s motions more predictable.