Two sources of performance-stabilizing synergies: An experimental exploration using finger force production

We explored the hypothesis on two sources of finger force variance in multi-finger accurate force production tasks, related to variability in the sharing of total force among finger forces and to sensory-based covariation of the finger forces. This hypothesis was explored within the space that did not affect task-specific performance variable (the uncontrolled manifold, UCM) and within the space that affected this variable (orthogonal to the UCM, ORT). Young, healthy subjects performed steady-state accurate total force production tasks with and without targets for the individual finger forces. These targets varied in size from 1.5% to 40% of the task total force level. The UCM hypothesis framework was used to quantify the two variance components, V_UCM and V_ORT, across trials and across 0.1-s samples selected from single 30-s trials at 1-s intervals. Across all conditions, V_ORT was similar for the inter-trial and single-trial analyses and across the finger force target sizes. In contrast, V_UCM (typically larger than V_ORT) was smaller during the single-trial analyses compared to the inter-trial analyses. V_UCM also decreased significantly with the drop in the target size for individual finger forces. We interpret these results as pointing at two sources of V_UCM. First, variability in the sharing of total force between the individual finger forces, based on practice, that can be seen from the very beginning of trials. Second, negative covariation of finger forces along individual trials based on visual feedback and, potentially, on loops within the central nervous system. Combining single-trial and inter-trial analyses of variance can provide information on these two sources and turn into a tool to quantify impaired control of movement stability in neurological patients.