Walking dynamics are symmetric (enough)

What is it about?

The human musculoskeletal system is approximately bilaterally (left--right) symmetric, but to what extent do our walking dynamics display this symmetry? We found that the walking dynamics of healthy human participants (N=8) exhibited statistically significant asymmetries and yet ignoring these asymmetries produced a more consistent and predictive model of human walking. These results provide a clear example of when neglecting evident characteristics of a system (in this case, asymmetry) can lead to a model that is not only more parsimonious, but also a better description of the data.

Featured Image

Why is it important?

Symmetry is a basic feature of life, but life is never perfectly symmetric. Are we wrong to say “hey, that flower is symmetric!” when it is really, at the micro-scale, clearly not symmetric? What we showed is that for human walking, regular people “limp” (for lack of a better term), i.e. they show some dynamical difference between how they step from left-to-right, compared to how they step from right-to-left. But when you are watching most able-bodied people walk, it would be hard to see that difference with the naked eye. In fact, what we then showed is that it is better from a modeling point of view to simply ignore the fact that people are a little bit asymmetric. You get a better, more accurate model, if you embrace the “almost” symmetry, even though, you know that platonically speaking your symmetric model is wrong! This is weird and counter-intuitive: you can prove walking is asymmetric, but nevertheless you get a better model by assuming it is symmetric, all using the same data set. From a human health point of view, thus study could be a first “step” toward understanding how symmetry breaks down in patients with neurological disorders. At what point to they “limp” so much that we can no longer assume symmetry? What might this mean for rehab and treatment?