Animals have a lot of muscles, and many of them do similar things. To an engineer, the multitude of muscles poses a conundrum. When a frog wants to kick its leg back, extending its hip, it could use either the gracilis muscle or the semimembranosus. As they both produce forces that tend to extend the hip, how does the animal decide which muscle to use? And if the animal needs to make a complex movement, how does it manage to sort through the plethora of muscle activation patterns in a reasonable amount of time, when many of the patterns produce roughly the same motions?

Experimental results point to a way out of the conundrum. When researchers measure muscle activation in many muscles, they often see that certain muscles– like the gracilis and semimembranosus – tend to come on together. These patterns, called `synergies', could mean that the frog doesn't need to make a decision. When it decides to extend its hip, maybe it just activates the `hip extending synergy' that turns on both the gracilis and semimembranosus.

Synergies could solve both the decision problem – which muscles to turn on – and also the sorting problem – making the decision in a reasonable amount of time – by restricting the number of choices. Rather than choosing an activation level for each muscle individually, the animal would only have to choose between a much smaller number of synergies.

But doesn't the reduction in the number of choices come at a cost? Wouldn't the animal end up turning on some muscles unnecessarily, because they were linked together in a synergy, and therefore end up wasting energy?

According to a recent study in PNAS by Matthew Tresch and his colleagues at Northwestern University, the answer is `no' – provided one chooses the synergies properly. In fact, the group found that only five synergies are sufficient to produce efficient motions in a mathematical simulation of a frog leg.

The researchers developed a computer simulation of the frog hindlimb and 13 muscles. Then they used two different sets of criteria to estimate synergies for the limb. First, they aimed to produce synergies that worked best with the natural biomechanics of the limb, which they called `natural dynamics synergies'. Second, they calculated synergies that would allow the animal to produce torques in any direction around each joint, called `joint torque synergies'.

They put their simulated frog leg through a test, asking it to move its leg to six different points. Initially, they calculated the best-case scenario:the most efficient motion, using the smallest muscle activations for each of the 13 muscles individually that would move the leg accurately to each point. Then they compared this best-case pattern to the closest motion produced using the two sets of synergies, rather than activating the muscles individually.

The joint torque synergies ended up wasting a lot of energy, but the natural dynamics synergies did well, producing muscle activations that were about 85% similar to the best-case situation. In support of the idea that synergies could produce the same movements but with fewer choices, they were also able to reduce the simulation to only five natural dynamics synergies– as if there were only five muscles in the limb, not 13 – and still keep more than 80% similarity to the best-case situation.

The benefit of using a few well-chosen synergies was clear, the researchers stated: solving the full, best-case scenario optimization problem took many hours on the computer. But, by using the natural dynamics synergies, they could find a solution in seconds that was nearly as good. Maybe that's how animals do it too.

Berniker, M., Jarc, A., Bizzi, E. and Tresch, M. C.(
). Simplified and effective motor control based on muscle synergies to exploit musculoskeletal dynamics.
Proc. Natl. Acad. Sci. USA.