Studies on unimpaired humans have demonstrated that the central nervous system employs internal representations of limb dynamics and intended movement trajectories for planning muscle activation during pointing and reaching tasks. However, when performing rhythmic movements, it has been hypothesized that a control scheme employing an autonomous oscillator — a simple feedback circuit lacking exogenous input — can maintain stable control. Here we investigate whether such simple control architectures that can realize rhythmic movement that we observe in experimental data. We asked subjects to perform rhythmic movements of the forearm while a robotic interface simulated inertial loading. Our protocol included unexpected increases in loading (catch trials) as a probe to reveal any systematic changes in frequency and amplitude. Our primary findings were that increased inertial loading resulted in reduced frequency of oscillations, and in some cases multiple frequencies. These results exhibit some agreement with an autonomous oscillator model, though other features are more consistent with feedforward planning of force. This investigation provides a theoretical and experimental framework to reveal basic computational elements for how the human motor system achieves skilled rhythmic movement.

This content is only available via PDF.
You do not currently have access to this content.