A decade ago, runners had a method for picking shoes that was simple, scientific, and wrong. It was all about pronation, ensuring that your shoe enabled your foot to roll inward by just the right amount with each stride. But amid the upheaval of the barefoot revolution, one of the first casualties was the pronation paradigm: despite two decades of increasingly clunky pronation-controlling shoes, runners kept getting injured.
The challenge, ever since then, has been figuring out what to replace it with. The temporary solution that many people (including me) settled on was a proposal from University of Calgary biomechanist Benno Nigg—one of the original proponents of the pronation paradigm, back in the 1980s—that runners should rely on what he dubbed the “comfort filter.” The idea is that if a shoe feels comfortable, your bones and joints are probably moving the way they’re supposed to, lowering your risk of injury. It’s convenient and simple but it’s also untested scientifically, and sounds suspiciously like a cop-out: we’re out of ideas, so just run in whatever feels good.
All of this lurks in the background of a recent study that tests a refinement of Nigg’s idea. The study was actually published back in January 2020, in the journal Scientific Reports, but I just came across it recently. It’s from a group of collaborators at the German Sport University in Cologne, the University of Massachusetts Amherst, and the shoe company Brooks. It’s a long way from offering a definitive answer on how to choose the right running shoe, but it offers an important first step: experimental data.
The theory being tested in the study is called the “habitual motion path,” which is the idea that (a) each person’s joints prefer to move in a distinctive way, and (b) the shoe that enables your joints to move as closely as possible to that preferred path will minimize your injury risk. It builds on Nigg’s earlier “preferred movement path,” and it has been developed by several researchers including Steffen Willwacher and Gert-Peter Brügemann at the German Sports University, Joe Hamill at UMass, and Matthieu Trudeau at Brooks.
One way to understand the first claim about unique movement patterns is to move cadaver legs through a running motion. Here’s a graph from a 2019 paper on the habitual motion path that shows the movement of the knee joint in six cadaver legs being flexed back and forth. The graph’s three axes show three different angles describing the movement of the knee, with each (dead) subject in a different color:
Each subject’s movement is highly repeatable: the knee joint moves in exactly the same way each time it flexes and extends. But each subject’s movement is also completely different from the other subjects. The proportions of your limbs, the stiffness of your tendons, and your movement history all conspire to make your knees unique. We each have our own habitual motion path.
The second claim—that this habitual motion path tells us something useful about running shoes—gets a lot trickier. First you have to decide how to measure the habitual motion path in living people. To do this, they measure joint angles while the subjects do an unloaded half-squat. The idea is that a half-squat looks like the way you bend your knee while running but places a fairly light load on your joints, just like the typical day-to-day activities such as walking and getting up from a chair that our joints and soft tissues have presumably adapted over time to handle. You can compare the angles measured during the half-squat to the angles measured during running, and calculate a number that tells you how far your running motion deviates from your habitual motion path. The more you deviate, the theory suggests, the more you’ll be loading joint areas and tendons and other structures that aren’t used to it.
In the newest experiment, Willwacher and his colleagues had 12 volunteers complete a series of three 75-minute runs in different shoe conditions. Immediately before and after the run, they had an MRI to measure the thickness of cartilage in their knees. That length of run temporarily compresses your cartilage, so greater loads on your knee should lead to greater squeezing of the cartilage.
The main result of the study is that runners with a greater deviation from the habitual motion path measured from their half-squat had significantly greater loss of cartilage volume in three parts of the knee during the run. This was based on a comparison between the six runners with the greatest deviation (12.5 degrees on average) and the six runners with the smallest deviation (6.2 degrees on average), averaged over all three runs. It’s an intriguing hint that the idea of a habitual motion path corresponds to real biomechanical effects.
It gets even more interesting when you break down the three treadmill trials. One was in standard neutral shoes, either the Brooks Launch or the Brooks Glycerin. Another was in a Launch with little plastic tubes rammed into the midsole to make the inner side of the shoe stiffer. The third was in a Launch with tubes making the outer side of the shoe stiffer. Here’s what those two shoes looked like:
Both of the two altered shoes should change your motion path. But which is better? Well, that depends. Three of the runners had their lowest deviation, which in theory should be the best choice, in the unaltered neutral shoe. Two of them did better with the stiff outer midsole, while seven did better with the stiff inner midsole. Looking at it from the opposite perspective, six runners had their highest deviation (i.e. the worst choice) in the neutral shoe, while three each did worst in the other two shoes. There’s no apparent pattern: everybody responds differently.
But do the angle changes correspond to lighter loads on the knee? Once again, the results were encouraging. When comparing each runner’s highest-deviation shoe to his or her lowest-deviating shoe, the high-deviating shoe produced significantly greater cartilage compression in the medial-tibial area of the knee joint. Bingo: evidence that you should pick a shoe that minimizes deviation from your habitual motion path.
OK, that’s the optimistic reading of these results. The first big set of caveats to note is that this is a very small study with a large number of potential outcome measures, none of which is “Did this runner get injured?” Assuming that changes in biomechanical variables would lead to real-world outcomes like reduced injury risk is precisely what led the running shoe industry down the pronation path. So this is at best an early proof of principle, not a final answer.
There’s also a practical question. If habitual motion path is the way to go, what does the practical in-store shoe-trying process look like? Brooks has adopted the idea, dubbing it Run Signature, and they’ve set up 30 specialty running stores around the world with a tablet app that enables the sales associate to estimate your habitual motion path and then see how much it changes in different shoes. But when I asked Trudeau, a biomechanist by training who heads Brooks Running’s Future Concepts team, he was careful not to overhype the tablet system, which he sees as the first step in a longer journey: “The scientist in me is hesitant,” he said.
Even if the lab, measuring joint angles properly is tricky. It usually involves affixing a bunch of markers to the runner, sometimes drilling holes in their shoes and clothing, and using a big array of carefully placed cameras. Getting sufficient accuracy with a tablet is a big ask, so Trudeau is more interested in the future prospects of markerless gait analysis technology. But he’s also cautious for deeper reasons. Estimating habitual motion path with a half-squat is “a good first try,” he says. But there may be better ways to monitor joint motion in a wider variety of common tasks like climbing stairs, standing from a chair, and so on, perhaps using wearable sensors to get a more holistic picture of how your joints prefer to move.
And the biggest question of all is whether shoes really matter. Nigg himself has argued that 80 percent of running injuries are caused by what he calls training errors: too much, too soon, in essence. Leaving the specific number aside, I think that sentiment is largely true. Trudeau doesn’t disagree either. He figures shoes are probably just a small factor in running injuries—but, he adds, they’re a factor we can control. So whatever happens with this idea, it’s important not to oversell it. Even a shoe that perfectly matches your habitual motion path won’t make you invincible.
Still, as preliminary as these results are, I find them really interesting. Maybe that’s because almost all of us wear shoes, so we need some method beyond color and brand for choosing between them. According to Trudeau, even Nigg himself sees the comfort filter idea as “a crutch until we can figure out how to use biomechanics.” It’s nice to know there’s some progress on the biomechanics, and the prospect of more to come. In the meantime, I guess I’ll stick with comfortable shoes.
For more Sweat Science, join me on Twitter and Facebook, sign up for the email newsletter, and check out my book Endure: Mind, Body, and the Curiously Elastic Limits of Human Performance.