Analysis of the synergies underlying complex hand manipulation.

Coupling of actuators into motor synergies has been observed repeatedly, and is traditionally interpreted as a strategy for simplifying complex coordination problems. This view implies a small number of task-independent synergies. We have shown that optimal feedback control also gives rise to synergies in the absence of any simplification; the structure and number of such optimal synergies depends on the task. To compare these hypotheses, we recorded hand postures from a range of complex manipulation task. The structure of the synergies we extracted (via PCA) was task-dependent, and their number significantly exceeded previous observations in a simpler grasping task. Our results lend support to an optimal control explanation rather than a "simplicity" explanation.