A schematic image showing how the cyber-frog's assembled components work together.

A schematic image showing how the cyber-frog's assembled components work together.

The epoch of genetic engineering has provided curious scientists with a previously unimaginable toolkit to begin unravelling the secrets of life. Yet, tinkering with DNA can only take us so far. When it comes to understanding how tissues such as muscle propel animals through the world, researchers need other tools, which stretch the tissue while simulating the nerve signals that drive contractions, to learn how they deliver force. But these experiments are far from the real-world experiences of muscles powering limbs. ‘We can't currently measure how an isolated muscle would interact with the skeleton and other muscles’, says Chris Richards from The Royal Veterinary College, UK. However, when Richards came across a paper by Benjamin Robertson and Greg Sawicki from Georgia Tech, USA (doi:10.1073/pnas.1500702112), describing how they had coupled the calf muscle from an American bullfrog with a computer simulation of the leg's movement, Richards realised that this type of virtual reality body could help him understand how frogs jump. But first, he and student Enrico Eberhard would have to construct a cyber-frog leg.

First they needed a computer simulation of the limb's movement that could perform the calculations faster than the movement itself. Richards says, ‘The speed is crucial because the simulation needs to update quickly enough for the muscle to “feel” as if it's interacting with the simulation’. Fortunately, David Borton at Brown University had told Richards about a piece of software called MuJoCo, which calculates how a group of objects moves faster than the actual movement, so Richards knew he could simulate the movement of the frog's leg fast enough in real time. The duo based the motion of the simulated leg on the movements of red-legged running frogs (Kassina maculata) as they leapt, substituting an African clawed frog (Xenopus laevis) calf muscle for the running frog muscle in the cyber-frog leg. ‘Most difficult was getting the muscle hardware to communicate with the simulation software’, says Richards, describing how Eberhard had to tightly coordinate the electronics recording the strength of the muscle's contraction with the MuJoCo leg movement simulation.

Once all of the components had been assembled, Richards and Eberhard were ready to throw the switch to set their in vitro virtual reality frog leaping, and calculate how the limb moved in response to each tiny step in the calf muscle contraction to recreate a single 100 ms leap. In addition, they ran a series of test leaps that a real frog could never attempt, ranging from a leg with no muscle to a muscle that just behaved like an elastic band. In addition, Richards and Eberhard investigated how well the frog pushed off when the muscle only began generating force half-way through the leap, as well as slightly altering the anatomy of their simulated frog, by shifting where the top of the calf muscle attached higher up the thigh.

Not surprisingly, the cyber-frog fell flat on its face when it had no muscle to push off with and the simulation where the muscle behaved like an elastic band didn't fare much better. However, the simulated amphibian did manage a mini-leap when the muscle kicked in half-way through the launch. Impressively, the cyber-frog produced an elegant hop when the muscle contracted normally while the modified simulated amphibian with the shifted muscle performed the highest and longest bound.

‘I was extremely happy when I saw the results’, says Richards, who is also excited about the opportunities that his new virtual reality amphibian present. ‘I plan to use this to study how evolutionary changes in skeletal morphology affect performance’, he says.

References

Richards
,
C. T.
and
Eberhard
,
E. A.
(
2020
).
In vitro virtual reality: an anatomically explicit musculoskeletal simulation powered by in vitro muscle using closed-loop tissue–software interaction
.
J. Exp. Biol.
223
,
jeb210054
.