Rodent locomotion is more complex than we thought
A new analysis of trajectory unpredictability enables the quantitative study of non-steady state locomotion in a way that is relevant to evolution, ecology, and robotics.
The paper in Nature Communications is here: http://rdcu.be/vAuh
When I first arrived as a PhD student with Andy Biewener in 2010, he told me that Kim Cooper, at the time a postdoc in the Harvard Medical School and currently a professor at UCSD, was hoping to learn more about the biomechanics of the bipedal jerboas that she was rearing at the Concord Field Station for her studies of limb development. Andy has a long career of characterizing the locomotion of many different types of animals, and we both thought that this would be a rather straightforward study to get me a jump start on my PhD. However, when Kim and I started running the jerboas on a straight racetrack, we kept finding surprising results.
Click here for a youtube playlist of all supplementary videos
Most animals with multiple gaits are cursorial, meaning that they change gaits at predictable speeds to minimize the metabolic cost of locomotion, just like a horse goes from walking to trotting to galloping. Although jerboas use three gaits, they transition between these gaits frequently and throughout their entire speed range. Kim and I thought it might be interesting to let these animals move in a less restricted area, and we discovered that they are really difficult to catch, because of their zig-zag motions and jumps straight up. Therefore, we decided to study how their locomotion affects their predator evasion ability, rather than their locomotor economy.
Kim and I traveled to Xinjiang, China, where jerboas and jirds (in the gerbil family) live together, sharing food and predators. We captured both groups of rodents in the wild and compared the locomotion of bipedal hopping jerboas to the quadrupedal scurrying jirds as they evaded simulated predation. Ram Vasudevan, who was at the time a postdoc at MIT, and is now a professor at the University of Michigan, had the insight to quantitatively measure the entropy of their trajectories, which tells us how difficult it is for predators to predict future prey behavior and plan strikes.
Click here for the video of the quadrupedal jird in the field
Click here for the video of the bipedal jerboa in the field
We found that the bipedal jerboas were significantly less predictable than the quadrupedal jirds, meaning that they should be less susceptible to predators. We wanted to go one step further, and show that our measurement of trajectory unpredictability really does correspond to predator evasion ability. After scouring the behavioral ecology literature, I found that small terrestrial animals face a tradeoff: every moment they spend in an open area provides the opportunity for discovering more food, but also presents the danger of being noticed by a predator. Thus, anxiety in open areas varies inversely with predator evasion ability. Kim suggested performing classical behavioral experiments of open-field anxiety on bipedal and quadrupedal rodents as an indirect way to tie locomotion to predator evasion ability. I completed this experiment 3 months before defending my PhD thesis in 2016. We found that the bipedal jerboas have no problem being in open areas, whereas the quadrupedal jirds spend much of their time hiding.
Click here for video of bipedal jerboa light-dark box experiment
Our research provides a mechanistic explanation for the ecological hypothesis that bipedal and quadrupedal desert rodents use different microhabitats within the same ecosystem. The evolution of bipedality increases the unpredictability of jerboa locomotion, making them less susceptible to predation and less anxious in open areas, which allows them to limit competition for limited food resources and happily coexist quadrupedal rodents that forage in covered areas near bushes. The entropy approach also enables the quantitative study of non-steady state locomotion, which is rarely characterized in comparative biomechanics, and can help engineers design systems with more life-like variation.
The paper in Nature Communications is here: http://go.nature.com/2gFU1yF