Use of the uncontrolled manifold (UCM) approach to understand motor variability, motor equivalence, and self-motion.

Abstract

We review the uncontrolled manifold (UCM) approach to the analysis of motor variance. In that approach, variance in multi-degree of freedom systems is decomposed into variance that leaves task variables invariant (UCM) and variance that does not (orthogonal to UCM). Larger variance within the UCM than orthogonal to it is interpreted as evidence for a task-specific solution of the multi-degree of freedom problem. The extent to which UCM measures depend on the choice of variables and coordinate systems has been a topic of controversial discussion. We clarify these issues and explain the sense in which the UCM approach is geometric in nature. We embrace a combined approach in which the geometrical perspective is retained but complemented by an assessment of the correlations in multi-degree of freedom movement data. We then review the problem of motor equivalence, in which deterministic rather than stochastic perturbations probe the organization of multiple degrees of freedom. We argue that motor equivalence requires a UCM perspective because the effect of perturbations on the task variable cannot be compared to their effect on task-irrelevant dimension of the system unless both are embedded in a shared space. The geometrical interpretation of UCM is thus critical for this extension of the UCM approach. We finally briefly review the concept of self-motion that is likewise based on the geometrical view of UCM.