Control processes underlying elbow flexion movements may be independent of kinematic and electromyographic patterns: experimental study and modelling.

Using a non-linear dynamic model based on the lambda version of the equilibrium-point hypothesis, we investigated the shape and duration of the control patterns underlying discrete elbow movements. The model incorporates neural control variables, time-, position- and velocity-dependent intrinsic muscle and reflex properties. Two control variables (R and C) specify a positional frame of reference for activation of flexor and extensor motoneurons. The variable R (reciprocal command) specifies the referent joint angle (R) at which the transition of net flexor to extensor active torque or vice versa can be observed during changes in the actual joint angle elicited by an external force. The variable C (coactivation command) surrounds the transition angle by an angular range in which flexor and extensor muscles may be simultaneously active (if C > 0) or silent (if C < or = 0). An additional, time-dimensional control variable (mu command) influences the dependency of the threshold of the stretch reflex on movement velocity. This control variable is responsible for the reflex damping. Changes in the R command result in shifts in the equilibrium state of the system, a dynamical process leading to electromyographic modifications and movement production. Commands C and mu provide movement stability and effective energy dissipation preventing oscillations at the end of movement. A comparison of empirical and model data was carried out. A monotonic ramp-shaped pattern of the R command can account for the empirical kinematic and electromyographic patterns of the fastest elbow flexion movements made with or without additional inertia, as well as of self-paced movements. The rate of the shifts used in simulation was different for the three types of movements but independent of movement distance (20-80 degrees). This implies that, for a given type of movement, the distance is encoded by the duration of shift in the equilibrium state. The model also reproduces the kinematic and electromyographic patterns of the fastest uncorrected movements opposed in random trials by a high load (80-90% of the maximal) generated by position feedback to a torque motor. The following perturbation effects were simulated: a substantial decrease in the arm displacement (from 60-70 degrees to 5-15 degrees) and movement duration (to about 100 ms) so that these movements ended near the peak velocity of those which were not perturbed; a prolongation of the first agonist electromyographic burst as long as the load was applied; the suppression of the antagonist burst during the dynamic and static phases of movements: the reappearance of the antagonist burst in response to unloading accompanied by a short-latency suppression of agonist activity. These kinematic and electromyographic features of both perturbed and non-perturbed movements were reproduced by using the same control patterns which elicited a monotonic shift in the equilibrium state of the system ending before the peak velocity of non-perturbed movements. Our results suggest that the neural control processes underlying the fastest unopposed changes in the arm position are completed long before the end of the movement and phasic electromyographic activity. Neither the timing nor the amplitude of electromyographic bursts are planned but rather they represent the long-lasting dynamic response of central, reflex and mechanical components of the system to a monotonic, short-duration shift in the system's equilibrium state.