My day as a lab rat:   shod vs. barefoot on a Harvard treadmill

  Over the last decade there has been a lot of interest in the mechanics of human walking and running and how it evolved, with related inquiry into whether modern shoes generally help us or hurt us.  At the scientific forefront of this research is Daniel Lieberman at Harvard, where he and a team investigate these topics in the Human Evolution Biology department and collaborate with similar efforts around the globe.  A cause for concern is the evident degeneration of lower extremity health as a whole within Westernized society, ranging from sports injuries to osteoarthritis to general pain and weakness in feet, legs, and back.  Are highly padded shoes and orthotics the right answer for everyone?  We in the barefoot community think not, and such solutions certainly weren't available to hominids thousands of years ago before footwear was developed.  While every individual case is different, research into these aspects of ourselves at a basic structural level seems important to pursue.  The conclusions can help bring beneficial advice to human populations about what's really best for their own two feet.
Thus, I was delighted to learn of an opportunity to be a participant in some of these studies.  I had already read over quite a few of the papers from Dan and his colleagues, and found the work quite interesting.  Simply obtaining good, unbiased data appeared to be one of the biggest obstacles, because the seemingly simple act of walking or running is not simple at all and involves some incredibly complex interactions between bone, muscle, ligaments, nerves, skin, and the surfaces we're on.  The forces at work are fairly large and dynamic, and a lot happens in multiple directions and in a very short time with every step we take.  And we don't really think anything of it, at least after we've learned how to do all this in the first place from babyhood.  The lab studies begin with methods of how to think about it and how to capture some of those dynamics, and then how to intelligently analyze what's going on.

The early work in this area, which dwells mostly on running, can be viewed here:
which is a multi-section treatise that links out to many references.  The Nature video on the second "Why Consider Foot Strike" page is worth a watch, describes the origins of this whole investigation, and shows some of the earlier lab setups.

More recent studies can be found at
which describes projects in other areas besides locomotion.

foot strike dynamics screengrab
[Pic: Nature, vol 463, January 2010]
  So apparently the group wanted to continue these efforts by studying some long-time barefooters, both in the US and other countries around the world.  They had already worked with runners in Kenya who largely grew up and trained barefoot, as well as visiting the Tarahumara tribe in Mexico whose prominence had been greatly amplified by Chris McDougall's book Born to Run. Because these studies were mostly done in the field, sophistication of the necessary equipment may not have been what the researchers would have preferred -- but still more feasible than convincing research subjects to travel long international distances for a few minutes in a lab.  In the US itself, finding ordinary people with a substantial history of being dedicated to a barefoot lifestyle is not so easy, mostly due to our misguided social stigmas that have developed over the last half-century or so.  And that's in spite of the rise in popularity of "minimalist" and barefoot running that sprang up around the 2010 timeframe, supported by McDougall's book and some other contemporary work.

While that movement didn't catch on quite well enough to put a decisive end to the pervasive social prejudice against bare feet, it was another small step toward more public awareness of why it can be a healthy way to live.  Fortunately, many people who understand the benefits have grouped together and formed supportive affiliations on the internet, and even more fortunately, some of them happen to reside in the Boston area convenient to Harvard and the established labs with all the best gear.

[Images are linked to larger detailed copies.]
Lab with the famous instrumented treadmill Contact was established and logistical arrangements were made, and I headed into Cambridge to meet the folks behind the studies.  Nick Holowka met me and showed me into the lab, and there in front of me was that treadmill of song and story -- the very one used in those studies I'd read about, still quite functional and a key resource in the data collection process.  Feet far more worthy than mine had been on those belts, and I felt honored to just be in the presence of this apparatus!

Treadmill sensor connections Below the belts are large arrays of strain sensors, wired out through thick multi-conductor cables into the back of a bulky signal-processing box over by the desk.  Other cables carry power, motor control and speed sensing.

Treadmill sensor plates I couldn't really peek in far enough to see the actual sensors.  This particular treadmill is not quite an early prototype from the company that built it, but only one or two steps past that toward development of their actual product line.

Qualisys high-speed infrared cameras Coupled with the treadmill is an array of these mounted around the lab: high-speed self-illuminating infrared motion-capture cameras, that all tie back into a video processing system.  Made by Qualisys, a Swedish company, and likely used in a lot of other motion-capture applications like CGI for movie effects.

Wall of shoeboxes The cameras are clamped to a wire tray that goes all the way around the room, which along with the specialized gear contains a lot of the typical laboratory appurtenances including a fume hood and refrigerated storage.  Since this room is where gait biomechanics are primarily studied, another notable feature is storage for many shoeboxes.

Motion sensor at greater trochanter Greater trochanter location on skeleton
My efforts to spread the word through the barefoot community had not been for naught; the lab had already lined up several subjects for this new investigative phase.  It turned out that two of us Boston-local barefooters had been scheduled for that same afternoon, and I offered to be second in line for testing so I could get pictures of the process applied to someone else.  Nick and the assistants were okay with me documenting the day, since part of the idea is for more of this research to be generally accessible and understood -- so I was going to try and do my part.

Now the reason they wanted us to wear "compression shorts" aka spandex became clear, along with keeping the shirt tucked in: a video sensor was applied right at the bump of the greater trochanter, the bone mass at the top of the femur, and had to stay in place and be visible.  That pretty much defines the leg/pelvis pivot point.  Along with this, height from the floor to that point [e.g. leg length] was recorded.  More sensors would get stuck onto us later.

Ultrasonic epidermal depth measurements But first, Nick wanted to measure our pad thickness.  Routine barefooters develop much thicker epidermis on the ground-contact areas of the foot, and those of us who continually challenge ourselves on rough terrain effectively grow our own shoes [as nature indeed intended, I'd argue].  The idea here was to gauge the distance in to the dermis, the next [and more vulnerable] layer of skin down, at a couple of uniform points of maximum epidermal thickness and toughness.

Viewing ultrasound image on tablet Ultrasound boundary layers
The probe is applied gently through a layer of ultrasound gel for better coupling, and getting a good read of the successive layers underneath it takes a bit of positioning skill.  The tablet displays the image in realtime, so the probe can be adjusted for best read on the fly before capture.  The strong white line is the skin surface; the more subtle line where the green arrow points is considered the boundary between epidermis and dermis, and flight time of the sound and echo in theory shows the depth.  The rest of the image is interior flesh of the foot, and in our case probably well-developed fat-pad and musculature for a good way in.

I really wanted to see what a "tenderfoot" would look like here, but didn't think to ask if one of the (shod) lab workers was willing to get measured...

Fitting FiveFingers shoes The order of the three test phases was randomized for each of us -- regular running shoes, Vibram FiveFingers minimal shoes, and barefoot.  First up was the Vibrams, which both of us found are actually fairly hard to get into.  Toes aren't used to finding their way into specific channels in a glove, and it took a while to figure out if these were actually fitting right or not.

Ready to go But eventually they felt tight and stable enough to be ready for testing, and in the meantime the rest of the sensors had been applied.  "Starting the treadmill in 3 .. 2 .. 1 .." came from the lab tech running the motion-control computer, and the belts started to move.  We quickly found that the 2 meters per second that had been previously set is a *very* brisk rate of over 4 MPH, so a bit of adjustment was needed to crank it down to a reasonable walking pace.  [Multiply m/s by 2.24]

Run 1: VFFs Treadmill from front with test subject
It took a couple of minutes to obtain a comfortable, stable gait; more on that later as I discovered the personal subtleties of that for myself.  There's also the minor caveat that the feet really should land clear of the centerline on each side to be on the reaction sensors, which for some people is a little unnatural.  I noted that in the Nature video, some of the prior running tests had the subject stay on one belt so their feet could track more in-line, but that works less well for walking.

Fitting running shoes Next up: the terrifying running kicks!  Stiff, heavily padded, with all kinds of internal "support" and a deliberately wide sole and forced "toe spring" up-curve at the front.  These gleaming white foot-prisons from Asics would provide no ground feel, and only a vague hope of stability through a rigid mechanical layout which could not possibly be right for every situation.

Basketball and ACE bandage not included.

Running shoes under test It again took a bit to settle into a stable gait in these, especially for someone this *not* used to being in them; a good laugh was had all round over just how odd it could be for certain individuals.

Computer displays from treadmill and cameras While our first subject was patiently suffering through this, I was watching the traces -- *those traces* just like I'd seen in the papers -- go by in realtime on the display, realizing that a deliberate gait change could probably show up really well here.

Barefoot test Ahhhh, much better without shoes at all.  But maybe this wouldn't really look so radically different on the traces.

As I pointed out during some discussion later, many barefooters will adopt a somewhat heel-strike gait on flat, uniform surfaces just because it lengthens the stride a bit and does feel a little more efficient.  Some do that, some don't.  But go ahead and dump that box of Legos onto the head of the treadmill, and you'd immediately see completely different movement dynamics!  Perhaps, in fact, in a way not so different from the gait we would invariably adopt a couple of weeks after this during our "Gravel Grind" challenge outing in Burlington.  A mid/forefoot placement allows more time to compensate and shift weight off of pointy things, and is simply what we're wired to do naturally on more hostile terrain.  Except that it's not simple at all, and we may never understand all the physiological subtleties of how that works.

Retroreflective sensor ball This is what the video sensors look like.  They're not flat discs as in earlier tests; they are small spheres wrapped up in retroreflective glass-bead tape with an attachment flange, and can thus be clearly seen and captured from a wide array of angles.  The motion-capture system wants distinct points of input, without interference from other objects.  In fact, a small reflective label on the subject's stretch shorts had to be taped over because it was coming up on the cameras as an unknown point!

Motion-capture realtime analysis vectors During the runs, images of the four motion points were transformed into a set of vectors in 3-space and time-correlated with the footfall data.  There's a lot to this, and I'm guessing that a lot of pre-calibration had to happen as the cameras were installed and aligned.

  And that was it for the round of testing -- skin thickness measurements, and then the walking capture with the different footwear.  Now it was my turn, and I already felt comfortable knowing that I could go ahead and geek out pretty hard on this stuff!  The other test subject had a few minutes to stay around, so I handed him my camera and said he should nab whatever looked interesting, and was quite pleased with what he captured.

H* foot ultrasound Nick had a little trouble getting a good thickness read on my feet.  With all the rough-trail hiking I'd been doing that same spring and early summer, my dog-pads were probably in peak condition and almost felt to me like attached shoes when I crunched my toes down.

Thick epidermal ultrasonic profile Eventually he ferreted out some of the inner boundaries; a respectable distance to the dermis, and I think he was on my heel at the time so the shorter line farther in was probably the calcaneus or fascia around it.

Applying motion-capture sensors Sensors were applied in the same anatomical spots, for capture consistency.

H* in VFFs under test First up was the Vibrams, which felt okay after a little thrashing around to get into them.  I had never worn these before!  I also realized that with the very low heel of this particular model of FiveFingers, my toe dorsiflexion would probably have them popping off at the back all the time if I had to wear them regularly.  There's a sort of token drawstring back there, so I pulled it tighter.
This wasn't the only weirdness, though.  I haven't spent much time on treadmills in general, and found this a bit disorienting at first.  The problem is lack of important visual cues, particularly the ground about a second and a half ahead moving toward us that we use to predict where and how to step.  That and the fact that the rest of the room isn't sliding by in peripheral vision.  So I found myself weaving a bit, having to concentrate on placing my feet to either side of the centerline and sort of brute-force myself to stay straight.  It didn't help that it felt like the ends of the middle-toe pockets in the VFFs were kind of flapping around, because I couldn't get seated all the way into them.  I think they're built for people with longer toes.

After a couple of minutes I probably had something as close to a normal gait as I was going to get, and they could fire away on data capture.

H* barefoot under test Barefoot was next for me, which of course made stability easier in general.  By now I had figured out that deliberately backing away from the handrail a little allowed me to use a little slice of visual motion at the front rollers of the treadmill as feedback.  I didn't want to use the handrail at all, to avoid skewing weight distribution, so I could produce honest walking traces as though I was on normal ground.  I quickly learned to for the most part ignore the floor ahead of that and the static room around me, but it definitely took a bit of mental shift.

H* in awful running shoes under test The mainstream kicks came last, and here's your rare money shot: Hobbit in overdesigned running shoes!  They felt completely clumsy as expected, and that awful toe elevation at the front made stepping off quite a chore.  I'm accustomed to my toes sort of gripping the ground right in place and adding that power to the launch; these forced my entire foot to roll forward in a really useless-feeling way, like the toes were doing absolutely nothing except maybe digging into the insole a little with no outward effect I could perceive.  None of the natural springiness of my foot could come into play.  I'm at a loss as to how wearers find these comfortable to spend all day in.

  And that was it for me too; the lab folks made sure everything was saved, and shut things down.  Nick and I had a little time to chat about some of my other observations afterward.

One thing I wanted to touch on was some of the temporal dynamics of running.  Not that I run very much myself, but the rare times I'm in a sort of dog-trot on a rough surface I've definitely noticed a lot of interesting processing going on.  The timeframe for compensating for terrain lengthens to multiple groups of steps, in a really interesting way that also utilizes body inertia.  It's a little hard to describe, but paying more attention to what my autonomous system is doing coupled with some video analysis could probably help nail down more about what we could observe.  Our feet have a lot of true three-axis reaction force sensing built in, and we use it best when we're barefoot.  At any gait and on any surface I feel a lot of subtle detail about the propulsive, twisting, and frictional forces at work through my soles, which is immediate and essential feedback.  I don't think we yet have an external instrumented equivalent with the same kind of granularity, which can capture force vectors *within* the area of a footfall.  In the absence of that, I suggested that while it's only qualitative data, it may still be useful to have some in-depth chats and outdoor sessions with routine barefooters who also love to analyze things.

The more I learn about musculoskeletal systems and what controls them, the more fascinating the subject becomes.  It is simply mind-blowing that the same basic cell structures, present in almost all creatures, can *repeatably* mutate to specialize into all the components that make this stuff work pretty reliably.

_H*   180726

Read more barefoot advocacy