Fig. 6.

Experiment 1 sensory subexperiment. A: sensory subexperiment example subject. Before and after the adaptation block, 18 subjects estimated the location of the P target by pressing buttons on a keypad rather than reaching to it. The change in these estimates (filled circle, Δŷ_P) reflects sensory realignment of the proprioceptive estimate of target hand position (ΔŶ_P) rather than adaptation of the motor command to the reaching hand. B: sensory subexperiment group data. For the 11 subjects with a significant P endpoint shift (first two bars), mean subexperiment Δŷ_P (solid bars) was 18.6 mm, or 52% of the P endpoint shift (open bars), which includes motor adaptation. This difference was significant (P = 0.002 by Wilcoxon rank sum test), suggesting that the P endpoint shift we measured in experiment 1 had approximately equal sensory and motor components. To verify that subexperiment Δŷ_P is not the same no matter what the P endpoint shift is, we also looked at the seven subjects whose P endpoint change was not significant (second two bars) and found that subexperiment Δŷ_P was not significantly different from zero or from P endpoint change. Furthermore, we found a significant correlation between P endpoint shift Δŷ_P and subexperiment Δŷ_P for these 18 subjects (correlation r = 0.54, P = 0.021, slope of the fit line: 0.77), supporting the idea that the P endpoint shift in experiment 1 comprised both sensory and motor components across subjects. Error bars represent SEs. *Includes a motor adaptation component in addition to sensory realignment.