Posture-based or trajectory-based movement planning: a comparison of direct and indirect pointing movements.

Various models have been proposed in the literature to explain the control of human arm movements. To make a quantitative comparison between the predictions of various models, we tested subjects for movements to targets on a vertical screen in various conditions. Subjects were asked to move directly from one target to another, or to move by a via-point, at various movement velocities and in a condition with a weight of 0.6 kg attached to the forearm. This set of experimental data was used for comparison with the predictions by various posture-based and trajectory-based models on 3-D movement planning and control. Small but significant effects of starting position and path towards the target were found on the torsion of the arm at the end of the movement. No effects of movement velocity and weight attached to the forearm were found. The experimental results differed significantly from the predictions by any of the models considered. Of the models considered, Donders' law best predicts the experimental data. Our data indicate that future tests of models for motor control (1) should compare the predictions of not just one, but several models to a data set, and (2) should include not only planar, but rather 3-D movements in such a comparison.