On the origin of variance along the uncontrolled manifold: Its effects on anticipatory and spontaneous changes in performance-stabilizing synergies in force production tasks

We used the framework of the uncontrolled manifold (UCM) hypothesis to explore the origin of inter-trial variance within the UCM, which by definition does not affect the salient performance variable, during accurate two-finger force production. Specifically, we tested several hypotheses on two main sources of variance within the UCM, variability in the sharing patterns between the fingers across trials and covaried variability in finger forces within individual trials. We also explored effects on unintentional changes in the structure of variance during preparation for a quick force change and during force drift without visual feedback. Fourteen young healthy participants performed two-finger force production trials with feedback on total force only or individual finger forces. In one part, they were instructed to produce a quick force pulse in a self-paced manner. In the other, they tried to keep finger forces unchanged without visual feedback. All manipulations led to significant changes in variance along the UCM, but not the ORT, space orthogonal to the UCM. Namely, the relative variance along the UCM dropped prior to quick force production and was not different from the ORT after the force drift. Changing the initial magnitude of variance along the UCM was reflected in its magnitude during anticipatory synergy adjustments prior to the force pulse and following the unintentional force drift. We interpret the results assuming a hierarchical control with two commands, reciprocal and coactivation. The results support the scheme with two contributing factors to variance along the UCM, likely associated with feed-forward and feedback mechanisms.