Figure 3.

Effect of different training interventions on compound acceleration vector. A, Group data (n = 10) showing the mean compound acceleration vector for the extension/flexion direction before and after training. Before training, the compound acceleration vector of each subject was aligned such that all vectors pointed into the extension direction. After PhysPract, the mean compound acceleration vector came to point into the flexion direction, corresponding to the practiced direction. ObsPract_opposite decreased the length of the mean compound acceleration vector, whereas ObsPract_toward did not change the mean compound acceleration vector. ^*p < 0.01. Open bars, Before training; filled bars, after training. B, Time course of changes of the mean compound acceleration vector as a function of training intervention. At the end of 30 min, the mean compound acceleration vector was smaller in PhysPract than in ObsPract_opposite or ObsPract_toward and with ObsPract_opposite than with ObsPract_toward. Error bars represent SEM. pre, Before training; B1, B2, and B3, results obtained after the first, second, and third training block, respectively.