Discrete movements of a limb about a single joint are a widely studied experimental paradigm for the investigation of voluntary movement. Never-the-less, the literature on this apparently “simple” task does not reveal a consensus on how such movements are controlled. Rather, a diversity of control rules exist, one of which must be selected depending upon specific aspects of the task. For example, movements that vary in distance appear to be controlled according to different rules from those that vary only in speed. Since natural movements as well as experimentally contrived ones usually vary from each other in several “parameters”, this appears to complicate the task of any central motor controller which is first, to select a control rule for a movement and then to apply it to the peripheral neuromuscular mechanisms.

In this paper, a computational procedure (program) will be defined that generates control signals for the motoneuron pools of agonist and antagonist muscles. This program can move a single-joint limb from one stationary position to another. The program accounts for moving different distances with different inertial loads and for the influence of different instructions concerning movement speed and accuracy. The model that generates these motor commands describes EMG patterns as well as force and kinematic trajectories which are consistent with much of the data found in the literature of these movements. The model is premised on the notion that kinematically defined tasks are accomplished by programming forces, based on only a few cognitively recognized kinematic and dynamic features of the task. Most of the features found in EMG and kinematic patterns can be considered consequences of the program’s algorithmic procedures rather than specifically planned features of those movements.

EMG data will be presented that illustrate the behavior of the motor system performing different kinds of single-joint elbow movements in the horizontal plane. These will be compared to the predictions obtained from the motor program equations solved on a computer. A high degree of correlation will be demonstrated between modeled and actual EMGs under diverse conditions.