by Carsten Hoever

Introduction

This is the second part of a series of blog posts about the RunLAB fullbody running analysis offered by the Swedish running shoe company Salming. The first post, which was published on The Running Swede Blog, was mostly a general description of the RunLAB and the general experience. This post will focus more on the technology used in the RunLAB, the biomedical parameters which are captured, and the outcome of my own analysis.

It is usually recommended to attend two RunLAB sessions. The first one will identify aspects of the running form which can be improved. Based on these findings the runner will be given suggestions for form changes or special exercises. The second session is typically a couple of months later to see if improvements in running form have been made. Originally, I planned to write this post after such a second visit. Unfortunately, a recent marathon resulted in acute plantar fascia issues which so far have prevented me from re-visiting the RunLAB. This means that this post has to live without an assessment of whether I have actually improved my running form or not after my initial analysis (but some information about how a friend fared is given at the end of the post). I hope to be able to provide this information in an injury-free future.

Technology

Salming currently has two permanent RunLABs in the Swedish towns of Gothenburg and Stockholm (in the US the technology was also showcased at the  2015 Running Event (more info here).

The basis of Salming’s RunLAB is a treadmill-based motion capture system. The treadmill is a professional grade model from the Swedish company Rodby. The motion capture system is from Qualisys (yet another Swedish company). In the Gothenburg RunLAB this treadmill is located in the centre of the Salming flagship store. On the walls surrounding it, there are eight special high speed motion capture cameras (see Figure 1) which can take images at 400 frames/s. Attached to each camera is a strobe which emits infrared light pulses which are invisible for the human eye but can be detected by the camera. These pulses are reflected by a total of 35 reflecting markers which are attached to the runner’s body at key locations (see Figure 2), usually at “bony” places close to joints, but for example also on the forehead. By combining the two-dimensional images from all eight cameras the motion of all 35 markers in a three-dimensional space can be recreated. With the help of the very high camera frame rate this allows a very precise analysis of full-body motion during running.

Figure 1: One of the high speed cameras.

Figure 1: Markers attached to the runner.

In a post-processing step, this data is then analysed for key biomechanical parameters such as

cadence;
step length;
ground contact time;
flight time;
pelvis height, obliquity, tilt and rotation;
knee angle;
ankle flexion;
foot rotation (pronation, supination), and contact (heel, mid foot, forefoot);
frontal/sagittal ankle path;
shoulder–pelvis flexion, lateral flexion, rotation;
elbow angle, and path;
wrist path; and
contact in relation to centre of gravity.

Figure 3: The RunLAB treadmill in the centre of the Salming flagship store in Gothenburg. While doing the analyses you can actually see a running skeleton of yourself at the big screen in front of you.

Each RunLAB session has the runner being filmed for a couple of minutes at three different speeds which are chosen based on a recent 10k race time (see Figure 3). In addition to the filming, the runner is also critically observed by a trained running coach from Salming’s RunLAB team.

After the results are compiled into sets of animations, pictures and diagrams, the coach goes through all results and explains them to the runner. As a normative reference, the data is also compared to the average results of large number of tested Swedish elite runners. Based on the data analysis, a few key areas in which running form can be improved are identified. This might also involve the runner doing another session of running or doing some other exercises under the scrutiny of the coach.

Before finishing the session, the runner is also given some hints on how to work on the problematic areas. A couple of days after the session the runner receives a link to a detailed web report, see Figure 4 for a screenshot (a full example in Swedish can be found here), which presents all the data in a very descriptive way. More importantly, the web report also contains information about the identified weaknesses and some suggestions for corrective exercises. Note that the layout of the web report will change in the coming weeks; the idea is to align it more with Salming’s so-called Running Wheel philosophy.

Figure 4: Screenshot of the Swedish web report. And yes, they got my name wrong. ;)

My case study

The most insight into the full body running form analysis is probably given by going through the results of an exemplary RunLAB session. In this case my session shall serve as such an example. Due to the extensive nature of the full analysis, it is not possible to go through all recorded parameters in this post. Instead, I will focus on a few parameters which cover the complete range from “totally normal” to “surprising” and finally, and probably most interesting, to “problematic”. In case anyone is interested in seeing my full results, this is the webreport (again unfortunately only in Swedish).

Note: the following images are taken from a PDF version of the report which was provided to me by Salming for the sake of this blog post. The PDF is neither as nice looking nor as informative as the normal web report (see linked examples above) and only includes the raw data and no further analysis, but at least it is in English (which is seen as big advantage for the sake of this post!).

Here are some of the key findings:

Stable, symmetric core motion

We will start with something very boring, a part of my running form which actually is pretty “normal”, the shoulders-pelvis angles for the (lateral) flexion and the rotation. As shown in Figure 5, my results (the blue/green lines) are pretty much in line with the Swedish elite runners results (the grey area). My flexion is probably on the edge of being too high, but not in a way that this is something to worry about. Something we will come back to later is that my shoulder-pelvis rotation is a little bit too small at lift-off.

Figure 5: Results for the different shoulders-pelvis angles. COG is centre of gravity, and LFS and RFS are left and right footstrike, respectively. The blue and green lines are my results for the different speeds, and the grey shaded areas are for the reference. Arrows added by the author for clarification.

Heel striking – Do I or Don’t I???

Popular belief often has it that the root of all running injuries and inefficiencies can be found in two causes related to the feet: (over-)pronation and heel striking. Interestingly enough, there is not only a lot of confusion about the general importance of these features, but also a wrong self-assessment by many runners who put themselves into certain footstrike or pronation-related categories with firm conviction. The RunLAB analysis gives you all the data to finally prove these beliefs…or to shatter them. For the sake of brevity, I will only focus on the footstrike in this post and leave the pronation results out.

I am what some people would call a midfoot striker, i.e. my foot touches the ground more or less flat and simultaneously with both the heel and the forefoot. There is no discussion about this, that’s how it feels when I run and that’s what the wear patterns on my shoes indicate. Only…it is not true, as is shown by the measured floor contact angle in Figure 6.

For both feet initial contact is clearly with the heel, and not even close to an angle at which it could be classified as a midfoot strike (the green area in the plot). People often claim to “get on their toes” at higher speeds, but for me this is not the case: the foot contact angle gets bigger (i.e. I am more dorsiflexing/heelstriking) at higher speeds. There is an interesting asymmetry between my left and right feet: dorsiflexion at footstrike is consistently higher for the left foot, and only at the highest speed for the left one am I getting into a large negative floor contact angle (i.e. larger ankle plantar flexion). I wonder if this is the reason why, when racing, I tend to get hot spots at the tip of my toes, but only on my left foot.

Figure 6: Results for the floor contact angle. LFS and RFS are left and right footstrike, respectively, and SD denotes the standard deviation. The lines are my results for the different speeds, and the colored areas show what classically would be classified as heel striking (blue), midfoot striking (green), and forefoot striking (beige).

So, now that I am officially a heel striker, does that mean that my injury risk is larger? Most certainly not, for this it is for example much more important where my foot hits the ground in relation to the pelvis. Obviously, the RunLAB analysis also includes this data. However, I will spare you the details now (which can be found in the linked web report) – it should be suffice to say that my running form is fine in this regard.

Sitting all day

Above I talked about my relative shoulder-pelvis motion being within the norm. The same is also true for the pelvis obliquity and rotation, see Figure 7. Unfortunately, this is not true for pelvic tilt, where my pelvis is rotated way too much backwards; a very typical trait for people like me who sit too much. This messes up the whole kinetic chain and can lead to inefficiencies and potential issues such as overstriding or problems to get the leg behind the body.

Figure 7: Results for the different pelvis movements. LFS and RFS are left and right footstrike, respectively. The blue and green lines are my results for the different speeds, and the grey shaded areas are for the reference group of Swedish elite runners. For an explanation of the different lines and markers see Figure 5.

As it is often the case, the limited pelvic tilt is only the symptom of a cause which sits somewhere else. Based on his observations of me running on the treadmill and some extra exercises I had to do after the actual RunLAB analysis, Salming’s coach concluded that my limited pelvic tilt is caused by a very stiff upper back. Based on how I feel in my upper back when I do certain movements I am not surprised by this finding. It is, however, quite interesting when you are reminded that a problem which sits rather high up in the body can affect the running form in a substantial way. To increase my upper back flexibility, I received instructions for three special upper body stretches which I should include in the warmup before every run. These have helped considerably to reduce my perceived back stiffness. If this has helped with my pelvic tilt I cannot say right now as it is difficult to judge without an outside observer.

Additionally, my pelvic rotation is too high at toe off. This is the opposite of what was observed for my shoulder-pelvis rotation. This asymmetry means that the pelvis and the upper body are not really working together, leading to an unstable posture.

Excessively static arm motion

The second major critique that I got was that my arm motion is too static. This was partly related to not being relaxed enough in the shoulders and a slightly wrong chest posture. The major issue, however, is the elbow angle shown in Figure 8. Ideally, the angle should be larger when the arms are in the backswing behind the body, and smaller in front of the body. I tend to have a rather constant elbow angle irrespective of where in the swing cycle my arms are. Arm swing can affect the motion of the legs as well, so it is something important to consider.

I was told to be more aware of my elbow angle, especially at higher speeds, and that I should generally try to be more dynamic in my arm swing (but only the sagittal plane, not in the frontal plane, where everything was good). Following this advice, I can clearly say that my stride felt considerably more powerful at higher speeds. At lower speeds, however, I am struggling a little bit to find the right amount of dynamic motion. I very quickly get a little bit too dynamic which then lengthens my stride and lowers my cadence. More importantly, the correct arm motion will also help with getting my chest posture correct, which in turn will help with the position of the pelvis.

Linking things together

So, to sum up, there are actually some things which I need to fix in my running form, and somehow, they all seem to be connected:

· a correct arm swing helps to get a better chest posture and a better leg extension,

· a more flexible upper body is necessary for a better chest posture, and

· a better chest posture leads to a better pelvis posture and is necessary for the pelvis and upper body to work together in a symmetrical and stable way.

· This in turn will help me to have more control at faster paces which should also help me to get off my heels at higher paces.

Figure 8: Results for elbow angle. COG is centre of gravity, and LFS and RFS are left and right footstrike, respectively. The blue and green lines are my results for the different speeds, and the grey shaded areas are for the reference group of Swedish elite runners. For an explanation of the different lines and markers see Figure 5.

Conclusions

This post was my attempt to convey some of the parameters which are measured in a RunLAB session, and, more importantly, how these parameters can be related to more tangible aspects of the running form. I cannot stress enough that the few titbits I provided here are only a small part of the actual analysis. Furthermore, in the way I presented these few examples, there could be the impression that some few key parameters are sufficient for an assessment of a runner’s biomechanics. The reality could not be further from the truth. An assessment of the whole kinetic chain is only possible if the motion of all relevant body parts is considered. This is the great strength of the RunLAB, as it is a true full body running form analysis which looks at the runner literarily from top to bottom. Running dynamics data as delivered by wearable devices such as Garmin’s HRM-Run cheststraps or the RunScribe footpods can be very helpful, but it always only tells a (very small) part of the story.

The captured data is only a part of the reason why a RunLAB session can be such a valuable tool. The more crucial part is the expert coach who analyses and explains the data to the runner. Salming has very wisely decided not to put the data analysis in the hands of ordinary salespersons, but true experts in running biomechanics. The captured data helps the coach to make the right decision, and the visually appealing way in which the data is presented makes the analysis more accessible and more useful for the runner. The data and the coach complement each other in a perfect way: the whole is greater than the sum of its parts.

Regarding the potential benefits of the analysis I would like to bring up the case of my friend Jay, who joined me for the first RunLAB session and recently had his second session. His initial analysis revealed similar shortcomings in arm motion and pelvis position as I have. In his second session he had improved in both of these aspects; he could now could start working on the next area (glute activation) to improve.

Now while I am pretty positive about the RunLAB in general, there are some aspects where I see limitations or problems with the approach. The first, and maybe biggest, it the fact that the test is under “lab conditions” on a treadmill. On average I do less than ten real treadmill runs per year (not counting the few minutes at the running store trying shoes). That doesn’t mean that running on the treadmill is something I struggle with, but each time I step onto a treadmill I also have the sensation that my running form is changing slightly compared to running outdoors. In particular, I have a feeling that I heel strike more. So we have the typical science dilemma here: you are trying to apply findings from an individual lab situation (the treadmill test) to the more general case (running outdoors) which might have slightly different boundary conditions. For most runners this might be of absolutely no relevance, but there might be also other runners who run completely different on a treadmill than they do outdoors. That’s not a problem with RunLAB alone, but basically all tests done on treadmills.

Another aspect is the choice of reference group. Right now they are comparing results against a set of Swedish elite runners. Probably a good assumption as it is likely that someone who runs fast also has a good form (though there are exceptions). Still, an elite runner is also a different beast than your ordinary 30km/week runner. My tempo pace might be an elite’s easy pace, and that can have implications on their and my form at that speed. Now don’t get me wrong, I don’t think there is a generally problem with this particular choice of reference group, but it’s always good to have the context in mind.

Then obviously, the outcome of a RunLAB session might have the runner trying to change their form, with all the potential implications that entails. “If it ain’t broke, don’t fix it” is still a mantra many deem valid with respect to running form as every runner is so uniquely individual in their biomechanics. On the other hand, certain “running form rules” are surely more or less universally applicable: e.g. you will have a hard time finding a runner who benefits from overstriding. Still, there is a risk that a form change might mess up your body in an unexpected way.

I also emphasised that the analysis is nothing without the coach. Right now Salming has experts working with the RunLAB who know their biomechanics and running form and can analyse the data appropriately. This is probably a more important problem for making the technology even more accessible than just the cost of the equipment. You absolutely need to have the right people to work with it. This is probably the biggest reason why we can’t expect RunLABs to open all over the world in the near future.

Finally, to put things into perspective, motion capture analysis of running biomechanics is not really something new. However, Salming can be complimented on making the technology and the expert assessment better available to the general public. The price is not cheap, but reasonable for what you get (at current exchange rates $220 for one session, or $330 for two), and not having to go a special testing facility (like a university lab) for the assessment but just your ordinary local running store certainly also helps.

Disclaimer

The RunLAB session was provided to the author free of charge by Salming Sports AB Sweden.

A few days ago a friend on Facebook posted a video that examines neuroplasticity as it relates to riding an unusual bike. As I watched it I kept thinking about the parallels to changing one’s running form. I also kept thinking about how amazing the human brain is – watch the video below, it’s well worth it!

1. Barefoot running is different and no shoe perfectly replicates the barefoot condition. Running barefoot, particularly on a hard surface, increases the likelihood that a runner will adapt a midfoot or forefoot strike. Running barefoot will also generally result in an increased stride rate and decreased stride length.

2. Running in a shoe with no cushioning will simulate some aspects of barefoot running, but will not necessarily simulate the barefoot condition perfectly. This might in part be due to the ability of any type of sole, even one with no cushioning, to reduce friction between the foot and the ground and thus reduce plantar skin abrasion.

3. Running in a “minimal” shoe with a moderate amount of cushioning is unlikely to alter form very much, particularly foot strike. In other words, if cushioning is present, a heel striking runner is unlikely to move to a midfoot or forefoot strike.

A new study was just published on-line in the journal PLOS One that adds additional support to some of what I have written above (full text available here). Here’s the Abstract:

Comparison of Minimalist Footwear Strategies for Simulating Barefoot Running: A Randomized Crossover Study

Karsten Hollander, Andreas Argubi-Wollesen, Rüdiger Reer, Astrid Zech

Published: May 26, 2015, DOI: 10.1371/journal.pone.0125880

Abstract

Possible benefits of barefoot running have been widely discussed in recent years. Uncertainty exists about which footwear strategy adequately simulates barefoot running kinematics. The objective of this study was to investigate the effects of athletic footwear with different minimalist strategies on running kinematics. Thirty-five distance runners (22 males, 13 females, 27.9 ± 6.2 years, 179.2 ± 8.4 cm, 73.4 ± 12.1 kg, 24.9 ± 10.9 km.week^-1) performed a treadmill protocol at three running velocities (2.22, 2.78 and 3.33 m.s^-1) using four footwear conditions: barefoot, uncushioned minimalist shoes, cushioned minimalist shoes, and standard running shoes. 3D kinematic analysis was performed to determine ankle and knee angles at initial foot-ground contact, rate of rear-foot strikes, stride frequency and step length. Ankle angle at foot strike, step length and stride frequency were significantly influenced by footwear conditions (p<0.001) at all running velocities. Posthoc pairwise comparisons showed significant differences (p<0.001) between running barefoot and all shod situations as well as between the uncushioned minimalistic shoe and both cushioned shoe conditions. The rate of rear-foot strikes was lowest during barefoot running (58.6% at 3.33 m.s^-1), followed by running with uncushioned minimalist shoes (62.9%), cushioned minimalist (88.6%) and standard shoes (94.3%). Aside from showing the influence of shod conditions on running kinematics, this study helps to elucidate differences between footwear marked as minimalist shoes and their ability to mimic barefoot running adequately. These findings have implications on the use of footwear applied in future research debating the topic of barefoot or minimalist shoe running.

Methods

In a nutshell, the study authors recruited 35 runners and had them run trials at 3 speeds (12:04 min/mile, 9:38 min/mile, 8:03 min/mile if I did the math correctly) in each of four different footwear conditions. The footwear conditions (see image at top of post) were barefoot, shoe with no cushion (Leguano), shoe with moderate cushion (Nike Free 3.0), and standard running shoe (Asics 2160). They recorded the following biomechanical variables during each trial: ankle angle at footstrike, knee angle at footstrike, stride frequency, step length, frequency of rear-foot strikes.

Results

Barefoot running was significantly different from all other footwear conditions (including the shoe with zero cushion) for three of the five variables measured. When barefoot, runners exhibited reduced ankle dorsiflexion at contact (flatter foot strike), increased stride frequency, and reduced step length. Barefoot runners still exhibited a heel strike about 60% of the time, which was similar to the zero-cushion shoe condition. In cushioned shoes, heel striking was observed about 90% of the time.

When subjects ran in the shoe with zero cushion (Leguano) they exhibited reduced ankle dorsiflexion at contact, increased stride frequency, reduced step length, and lower frequency of heel striking relative to both cushioned shoe conditions.

The cushioned minimal shoe (Nike Free 3.0) differed from the standard shoe (Asics 2160) in that the runners exhibited increased stride rate and reduced step length. Ankle angle and frequency of heel striking did not differ between the two cushioned shoes.

Knee angle at foot strike did not differ between any of the footwear conditions (barefoot included).

Commentary

The results of this study suggest a gradation of effect of running footwear on form. None of the shoes mimicked the barefoot condition perfectly, even the minimal shoe with no cushion. When barefoot, the runners had the smallest amount of ankle dorsiflexion, a higher cadence, and a shorter stride. The zero cushion shoe yielded similar results to barefoot for foot strike, and had intermediate values for ankle angle. Stride rate decreased incrementally from the zero cushion shoe to the standard shoe, and step length increased incrementally from the zero cushion to the standard shoe.

This study by Bonacci et al. found a similar reduction in stride length and increase in stride rate in the Nike Free relative to a traditionally cushioned shoe, so it does appear that a moderately cushioned shoe may induce some amount of form change in the direction of barefoot running. That being said, retention of a heel striking gait in such shoes can lead to increases in impact forces over traditional footwear.  Although the importance of impact forces to injury risk remains a source of debate, it seems prudent to suggest that care should be taken when migrating toward shoes with moderate cushion.

One of the disadvantages of this study is that it looked at immediate change in runners who were not familiar with running in minimal footwear, so we can’t know for sure if changes from the standard shoe might become more apparent with time and additional adaptation to such shoes.

In my own research I have found that barefoot runners on asphalt are more likely to midfoot or forefoot strike compared to runners in the minimally cushioned Vibram Fivefingers. Combined with results from the study discussed here (and others), these findings support my belief that barefoot running is different from running in any kind of shoe, and that although a zero or minimally cushioned shoe can alter form in the direction of barefoot running, it may never perfectly simulate what happens when you take your shoes off. At the same time, running in a minimal shoe is different than running in a more traditional shoe, but it might take removal of most or all of the cushioning to elicit major changes in running form (at least in the short term).

In my recent review of the Nike Free 5.0 I cited a study that compared running biomechanics in the Nike Pegasus and the Nike Free 3.0. This research has been out for awhile, but I never wrote about the paper. Since I’ve been running a bunch in various Nike Free shoes lately, I thought I’d write up a summary.

The study is authored by Richard Willy and Irene Davis and is titled “Kinematic and kinetic comparison of running in standard and minimalist shoes.” It was published in 2014 in the Journal Medicine & Science in Sport & Exercise. Here is the Abstract:

Abstract

PURPOSE: The purpose of this study was to determine whether running in a minimalist shoe results in a reduction in ground reaction forces and alters kinematics over standard shoe running. The secondary purpose of this study was to determine whether within-session accommodation to a novel minimalist shoe occurs.

METHODS: Subjects were 14 male, rearfoot striking runners who had never run in a minimalist shoe. Subjects were tested while running 3.35 m·s(-1) for 10 min on an instrumented treadmill in a minimalist and a standard shoe as three-dimensional lower extremity kinematics and kinetics were evaluated. Data were collected at minute 1 and then again after 10 min of running in both shoe conditions to evaluate accommodation to the shoe conditions.

RESULTS: Shoe-time interactions were not found for any of the variables of interest. Minimalist shoe running resulted in no changes in step length (P = 0.967) or in step rate (P = 0.230). At footstrike, greater knee flexion (P = 0.001) and greater dorsiflexion angle (P = 0.025) were noted in the minimalist shoe. Vertical impact peak (P = 0.017) and average vertical loading rate (P < 0.000) were greater during minimalist shoe running. There were main effects of time as dorsiflexion angle decreased (P = 0.035), foot inclination at footstrike decreased (P = 0.048), and knee flexion at footstrike increased (P = 0.002), yet the vertical impact peak (P = 0.002) and average vertical loading rate (P < 0.000) increased.

CONCLUSIONS: Running in a minimalist shoe appears to, at least in the short term, increase loading of the lower extremity over standard shoe running. The accommodation period resulted in less favorable landing mechanics in both shoes. These findings bring into question whether minimal shoes will provide enough feedback to induce an alteration that is similar to barefoot running.

Rationale and Methods

Many runners choose a moderately cushioned shoe to transition toward more minimal footwear figuring that it is a safer approach than jumping right into a shoe with no cushion at all. Because of this, the authors were interested in determining if running in a cushioned “minimalist” shoe might more closely simulate barefoot-like running mechanics than a more traditionally cushioned shoe. To address this, they recruited 14 heel-striking runners and had them run in both the Nike Pegasus and the Nike Free 3.0. The Pegasus has about 20mm more cushion in the heel, and the cushioning in the Peg is also significantly softer than that of the Free. Each runner ran two 10-minute trials on an instrumented treadmill, one in each shoe, and biomechanical measures were recorded near the beginning and end of the trials to see if acclimation time to the shoes had any effect.

Results

Contrary to expectations, step rate (both shoes just under 170 steps/min), step length, and foot inclination at foot strike did not differ between the two shoes. Even more unexpected was the fact that ankle dorsiflexion, vertical impact peak, and vertical loading rate were all significantly higher in the Nike Free 3.0.

Comparisons from the beginning to the end of the 10 minute run did not reveal any acclimation differences between the two shoes. In both the Pegasus and the Free 3.0 runners did not alter step length or rate, foot inclination and ankle dorsiflexion decreased, knee flexion at foot strike increased, and vertical impact peak and loading rate increased over the course of the run.

Commentary

Contrary to expectations, runners in the Nike Free 3.0 exhibited higher impact and loading rates compared to when they wore the Nike Pegasus. Furthermore, they did not exhibit changes typical of barefoot runners such as reduced step rate and stride length, or reduced foot inclination at initial contact that is indicative of adopting a midfoot or forefoot strike. In plain terms, the Nike Free 3.0 didn’t do a good job of simulating barefoot running. Rather, wearing the barefoot-inspired shoe resulted in a more impactful stride.

So what might be going on here? The authors suggest that “In order to see a change in footstrike pattern that simulates barefoot running, the shoes may need to be as minimal as possible.” In other words, the Nike Free 3.0 has enough cushion that it does not encourage alteration of the running stride in the direction of barefoot running. However, it has less cushioning than a traditional shoe and is firmer than the Pegasus so continuing to heel strike in a shoe like this results in increased impact loading. I should point out that I have observed that almost 50% of runners in the minimally cushioned Vibram Fivefingers continue to heel strike, so it may be that no shoe really does a perfect job of encouraging a barefoot-like stride.

The authors further point out that loading rates in the Free 3.0 exceeded those reported for runners with a history of tibial stress fractures. Other research has found that transitioning into Nike Frees increases injury risk to a greater degree than transitioning into the Nike Pegasus or the Vibram Fivefingers (injury risk in the latter two did not differ). We don’t know for sure if increases in impact loading might explain the increased injury risk in Frees (it could be their extreme flexibility or some other factor), but the combined results of these studies suggest that caution is warranted when beginning to run in a flexible, moderately cushioned shoe like the Nike Free 3.0. At the very least, runners should pay attention to running form when transitioning into a moderately cushioned shoe as they will not provide the same level of protection for a runner with a high-impact stride.

In this post we’ll take a look at another study on foot strike, in this case the influence of surface hardness on foot strike type in habitually shod runners who were asked to run barefoot in the lab. The study, led by Allison Gruber, was published in 2013 in the journal Footwear Science. They were interested in trying to determine whether pain associated with a barefoot heel strike on a hard surface might trigger a shift to a midfoot or forefoot strike.

Here is the abstract (full text is available here):

Footfall patterns during barefoot running on harder and softer surfaces

Allison H. Gruber^a, Julia Freedman Silvernail^a, Peter Brueggemann^b, Eric Rohr^c & Joseph Hamill^a*

Footwear Science, Volume 5, Issue 1, 2013, pg 39-44

Abstract

It has been suggested that the development of a thick, soft midsole of running shoes over the past 30 years has been primarily responsible for the majority of runners adopting a rearfoot or heel-toe footfall pattern thus deviating from a more ‘natural’ forefoot pattern. The purpose of this study was to determine the freely chosen footfall pattern when running barefoot on a harder versus a softer surface. Forty habitual rearfoot runners performed two running conditions: barefoot over a harder surface and barefoot over a softer surface. Three-dimensional motion analysis and ground reaction force data were collected to measure the ankle angle, vertical impact peak and strike index. The kinematic and kinetic parameters were used to confirm the footfall pattern in each condition. Only 20% per cent of the participants ran with a midfoot or forefoot pattern on the soft surface whereas 65% of the participants ran with a midfoot or forefoot pattern when running on the hard surface. Out of the 80% of participants that maintained a rearfoot pattern on the soft surface, 43% of these participants ran with a midfoot or forefoot pattern on the hard surface. These results suggest that, while running barefoot, the hardness of the running surface may be a significant factor causing an alteration in a runner’s footfall pattern.

Methods

The researchers recruited 40 habitually shod runners who were confirmed to be rearfoot strikers in their typical training shoes. They had each runner run trials in the lab along a 25 m runway equipped with a force platform. Trials included two conditions: 1) concrete runway with no cushioning (the hard surface); 2) runway covered with 20mm EVA foam mats to simulate characteristics of a shoe midsole (the soft surface). Foot strike patterns and a variety of other biomechanical measurements were recorded for each subject in both conditions.

Results

On the soft surface 80% of the runners ran with their usual heel-striking pattern, 17.5% ran with a midfoot strike, and one person ran with a forefoot strike. On the hard surface, 35% of subjects ran with a heel strike, 27.5% ran with a midfoot strike, and 37.5% ran with a forefoot strike. Statistical analysis confirmed that there was a significant shift in foot strike patterns between the hard and soft surfaces.

Commentary

What I liked about this study was that it asked and tested a simple question: What happens if habitually shod heel strikers run barefoot on a hard vs. a soft surface? What they found is that runners tended to shift to a midfoot or forefoot strike on concrete, but most continued to heel strike on the cushioned surface. This suggests that the presence of cushioning allows runners to heel strike comfortably and could indicate why heel striking is commonly observed in shod runners on hard surfaces, whereas barefoot runners tend to shift to a midfoot or forefoot strike on hard surfaces. Moving to a midfoot or forefoot strike allows the Achilles tendon and calf muscles to assist in attenuation of impact forces and could make running on hard surfaces more comfortable.

The results here could also explain why habitually barefoot or minimally shod people like the Daasanach, Tarahumara, or Hadza children and adult women have been observed to heel strike while running slowly on natural sand/dirt surfaces, whereas experienced Kalenjin runners moving fast tend to midfoot or forefoot strike on such surfaces. It might also explain why approximately 50% of runners in Vibram Fivefingers have been observed to contact on the heel when running on asphalt – there may be just enough cushion present in minimal shoes to allow this to be done comfortably if the pace is not too fast.

One interesting finding here was that 20% of the runners switched from their typical heel strike in shoes to a midfoot or forefoot strike even on the soft surface. The authors reported in their Discussion section that some of these subjects had previous experience with barefoot running, and that they indicated that they were more comfortable with the MFS/FFS when running barefoot. In other words, previous experience running barefoot could have led to development of modified form in this condition (similar to the role of experience in the Hadza runners). I would also suggest that even though cushion is present in both conditions, plantar sensation when running barefoot on a mat vs. with a shoe on the foot might differ and could explain some of these changes.

One other interesting finding was that 35% of subjects continued to heel strike when running barefoot on concrete. The authors suggest that the lab runway may not have been long enough to induce discomfort necessary to trigger a change in foot strike. However,  I have observed a similar proportion of barefoot runners heel striking on asphalt during a race, so length of the runway may not have been the only factor. It may be that a move to a less prominent heel strike could make barefoot heel striking comfortable enough for some runners, or it could be that some runners need more experience before adopting a different foot strike when barefoot on a hard surface.

As with most studies, this one provides some answers but also raises new questions. For example, I wonder if there is an individually specific threshold combination of speed and surface hardness that triggers a shift in foot strike to attenuate the impact forces associated with a heel strike? Also, why do some people shift but others do not? How much experience is required for a shift to occur, and are there some people who will never shift their foot strike? What role does plantar sensation or abrasions with the ground surface play in determining foot strike? To me, the continual search for answers to questions like these is what makes science so much fun.

In this post we’ll take a look at another study that examined foot strike patterns in a population that habitually wears minimal footwear. This study, led by Herman Pontzer, looks at Hadza hunter-gatherers from Tanzania, Africa. Like the traditional Tarahumara huaraches, the Hadza tend to mostly wear minimal sandals made out of repurposed tire rubber.

The Hadza are an interesting group in which to study running form since adult males and females exhibit a division of labor. Adult females spend the day gathering foods like berries and tubers, whereas adult males hunt game and collect honey. Both groups walk a lot, and though running is not a common practice, it is presumed that males engage in running more frequently during hunts. This difference allows for some insight into the role of running experience in contributing to variation in running form.

Here’s the abstract to the study (full-text available here):

Foot strike patterns and hind limb joint angles during running in Hadza hunter-gatherers

Herman Pontzer, Kelly Suchman, David A. Raichlen, Brian M. Wood, Audax Z.P. Mabulla, Frank W. Marlowe

Journal of Sport and Health Science

Volume 3, Issue 2, June 2014, Pages 95–101

Abstract

Background

Investigations of running gait among barefoot and populations have revealed a diversity of foot strike behaviors, with some preferentially employing a rearfoot strike (RFS) as the foot touches down while others employ a midfoot strike (MFS) or forefoot strike (FFS). Here, we report foot strike behavior and joint angles among traditional Hadza hunter-gatherers living in Northern Tanzania.

Methods

Hadza adults (n = 26) and juveniles (n = 14) ran at a range of speeds (adults: mean 3.4 ± 0.7 m/s, juveniles: mean 3.2 ± 0.5 m/s) over an outdoor trackway while being recorded via high-speed digital video. Foot strike type (RFS, MFS, or FFS) and hind limb segment angles at foot strike were recorded.

Results

Hadza men preferentially employed MFS (86.7% of men), while Hadza women and juveniles preferentially employed RFS (90.9% and 85.7% of women and juveniles, respectively). No FFS was recorded. Speed, the presence of footwear (sandals vs. barefoot), and trial duration had no effect on foot strike type.

Conclusion

Unlike other habitually barefoot populations which prefer FFS while running, Hadza men preferred MFS, and Hadza women and juveniles preferred RFS. Sex and age differences in foot strike behavior among Hadza adults may reflect differences in running experience, with men learning to prefer MFS as they accumulate more running experience.

Methods

The researchers traveled to Tanzania and recruited 26 Hadza adults (15 men, 11 women) and 14 juveniles (mean age = 8.6 years) to run trials while being recorded by a high-speed camera. The subjects ran along a hard-packed sand/silt trackway cleared of shrubs and loose rocks. Ankle, knee, and plantar foot angles were measured at initial foot contact from the video sequences. Some of the adults ran barefoot, others ran in minimal sandals. All of the juveniles ran barefoot.

Results

For adult males, 86.7% were midfoot strikers while running, and 13.3% were heel strikers. For adult females, 9.1% were midfoot strikers, and 90.9% were heel strikers. For juveniles, 14.3% were midfoot strikers, and 85.7% were heel strikers.

No significant differences in foot strike patterns were found between barefoot and minimally shod adults (though non-significant, heel striking was actually more frequent in the barefoot condition). Speed was not found to significantly influence foot strike (though the sample was small for this subset of data). No forefoot striking was observed in any of the trials. Joint angles did not differ between the groups.

Commentary

What I like about this study is that it may provide some evidence that running experience has an influence on the development of certain aspects of running form, in this case foot strike. Hadza children overwhelmingly tend to heel strike when barefoot, as do adult Hadza women. In contrast, adult Hadza males tend to midfoot strike when running. The authors write the following as a potential explanation of this pattern:

“This pattern of foot strike usage suggests running experience may be important in developing foot strike preferences. As children learn to walk and their gait matures, RFS develops as a normal part of the walking gait cycle;²⁰ thus RFS is the behavior learned first. As the musculoskeletal system and motor control develop further during adolescence, experience running barefoot or minimally shod may lead to a preference for MFS or FFS during running, perhaps in response to the high impact forces²¹ experienced when running with RFS. Individuals who rarely run might not have the same accumulated experience of high impact forces due to RFS, and thus never switch from RFS to MFS or FFS for running.”

They point out that the only two adolescents who were classified as midfoot strikers were the two oldest boys in the group, and suggest that:

“…the change in foot strike behavior by Hadza men may develop as they learn to hunt and track wild game. While Hadza men do not typically engage in endurance running, it is likely that they run more often as they learn to hunt than their female counterparts do in learning to gather plant foods. Indeed, our measurements of travel speeds used while out of camp on forays, taken using wearable GPS devices,¹⁶ indicate that men use running speeds approximately twice as often as women.”

These results have implications with regard to the interpretation of two other studies of foot strike patterns among habitually barefoot Africans. In Daniel Lieberman’s well known study of foot strike patterns among habitually barefoot Kenyans he found that the overwhelming majority exhibited midfoot or forefoot strikes while running. The Kalenjin subjects that he filmed were experienced runners, and it’s worth noting that they were running quite fast. In contrast, habitually barefoot Daasanach adults tend to heel strike at slower running speeds (though frequency of midfoot/forefoot striking increased at faster speeds). Unlike the Kalenjins, the Daasanach rarely run. Thus, in addition to speed differences, differences in running experience could partially explain the differences in foot strike patterns observed between the two groups.

For me, the most surprising result in this study was the pattern observed for the Hadza children. If you asked me how I thought a group of kids who are typically barefoot or shod only in a pair of sandals would strike the ground when running barefoot, I would have guessed an overwhelming majority would be mid- or forefoot strikers. This study demonstrated the exact opposite – almost all were heel strikers. If you think heel striking is abnormal or bad, just think about that for a second. It’s not as simple as saying that barefoot runners forefoot strike, or that kids never land on their heels. It’s more likely that foot strike is determined by a number of factors, among which are experience, speed, surface properties, footwear, fatigue, and so on.

I’ll finish by saying once again that I don’t believe there is a perfect or “natural” running form that applies under all circumstances. Rather, form is determined by properties of the individual and the characteristics of their immediate running environment. There are things that many people could do better (e.g., reduce overstriding, better hip stabilization), but to say that every person should run a certain way simply does not sync with the range of variation we see, even in populations that are habitually barefoot or minimally shod.

Personally, I view foot strike as one aspect of running form that varies with a range of factors. These factors would include running speed, running experience, surface hardness, surface incline, fatigue, footwear type or the lack thereof, etc. For example, I think we tend to see a shift more toward the forefoot side of the spectrum when experienced runners run fast on hard surfaces, particularly if barefoot. We see a shift more toward the heel strike side when inexperienced runners run slowly on soft surfaces or in heavily cushioned footwear. Mix and match these factors and we can get a variety of different foot strike types, and it’s unlikely that any one set of conditions will yield identical responses in all runners exposed to them (because people are variable).

As an example of the influence of both external variables on running foot strike and individual variability within the same condition, I thought it might be interesting to take a look at a research study that was published last year. Authored by Daniel Lieberman, the study takes a look at foot strike patterns among traditionally shod and conventionally shod Tarahumara Native Americans from Mexico (full text is available here). The running prowess of the Tarahumara was made famous by Christopher McDougall in his book Born to Run. That book, published in 2009, was largely responsible for kicking off the form and footwear debates that ensued over the following years, and the Tarahumara thus make for a great running form case study.

Here’s the abstract of the study:

Strike type variation among Tarahumara Indians in minimal sandals versus conventional running shoes

Daniel Lieberman

Journal of Sport and Health Science, Volume 3, Issue 2, June 2014, Pages 86–94

Abstract

Purpose

This study examined variation in foot strike types, lower extremity kinematics, and arch height and stiffness among Tarahumara Indians from the Sierra Tarahumara, Mexico.

Methods

High speed video was used to study the kinematics of 23 individuals, 13 who habitually wear traditional minimal running sandals (huaraches), and 10 who habitually wear modern, conventional running shoes with elevated, cushioned heels and arch support. Measurements of foot shape and arch stiffness were taken on these individuals plus an additional sample of 12 individuals.

Results

Minimally shod Tarahumara exhibit much variation with 40% primarily using midfoot strikes, 30% primarily using forefoot strikes, and 30% primarily using rearfoot strikes. In contrast, 75% of the conventionally shod Tarahumara primarily used rearfoot strikes, and 25% primarily used midfoot strikes. Individuals who used forefoot or midfoot strikes landed with significantly more plantarflexed ankles, flexed knees, and flexed hips than runners who used rearfoot strikes. Foot measurements indicate that conventionally shod Tarahumara also have significantly less stiff arches than those wearing minimal shoes.

Conclusion

These data reinforce earlier studies that there is variation among foot strike patterns among minimally shod runners, but also support the hypothesis that foot stiffness and important aspects of running form, including foot strike, differ between runners who grow up using minimal versus modern, conventional footwear.

Methods

Lieberman traveled to Mexico and collected kinematic data from 20 Tarahumara. Twelve of these individuals wore only traditional huarache sandals (see photo at top of post for image of huaraches), and eight wore conventional running shoes. After a five minute warmup, subjects were filmed while running 15 meters along a flat, natural surface free of grass and rocks. Each subject ran the 15 m trials until three usable videos were recorded for each. A variety of biomechanical variables were measured from the video sequences.

Results

Minimally shod Tarahumara that ran their trials in huarache sandals exhibited the following foot strike patterns: 40% midfoot strike, 30% forefoot strike, 30% heel strike. In contrast, Tarahumara in conventional running shoes landed on the heel 75% of the time and on the midfoot 25% of the time. No forefoot strikes were observed in the conventionally shod runners. Of the other biomechanical variables measured, only overstride angle differed – minimally shod runners tended to land with the ankle in a position more underneath the knee at ground contact.

Commentary

These results demonstrate two things to me:

1. Footwear does have an influence on foot strike. Those Tarahumara who typically wear conventional shoes and ran their trials in this type of shoe tended to heel strike 75% of the time. This is consistent with other studies that have found that 75% or more of traditionally shod runners land first on the heel.

2. Foot strike patterns are variable in Tarahumara runners wearing huarache sandals. Almost a third of them heel strike, and there is no overwhelmingly predominant pattern exhibited by this group. Lieberman himself writes: “…minimally shod Tarahumara runners appear to be best characterized as midfoot strikers who also employ forefoot and rearfoot strikes. These data therefore partially support but also modify earlier anecdotal reports that Tarahumara runners who use huaraches primarily FFS (forefoot strike).” As the source of these anecdotal reports, Lieberman cites McDougall’s book Born to Run and Scott Jurek’s book Eat and Run.

So here we have two levels of variability. Variability among groups due to the type of footwear typically worn, and variability within groups wearing similar kinds of footwear. It’s not as simple as saying that all huarache-wearing Tarahumara forefoot strike, and this suggests that there is no “perfect” or “natural” running form or foot strike that is or should be used under all circumstances by all people. Rather, running foot strike is dictated both by external influences (footwear here) and individual variation in response to a given set of conditions.

To me, this combo of adaptability and variability is what makes the study of human running form so fascinating – we are amazing and complex animals, and this is exhibited by both our incredible running ability as well as how we run.

If you enjoyed this post, you might also like this one on foot strike patterns in Hadza hunter-gatherers.

In this post we’ll take a look at another study that investigated the effects of a self-directed cadence training protocol on runners. Here’s the Abstract

Scand J Med Sci Sports. 2015 Feb 4.

In-field gait retraining and mobile monitoring to address running biomechanics associated with tibial stress fracture.

Willy RW1, Buchenic L, Rogacki K, Ackerman J, Schmidt A, Willson JD.

Abstract

We sought to determine if an in-field gait retraining program can reduce excessive impact forces and peak hip adduction without adverse changes in knee joint work during running. Thirty healthy at-risk runners who exhibited high-impact forces were randomized to retraining [21.1 (±1.9) years, 22.1 (±10.8) km/week] or control groups [21.0 (±1.3) years, 23.2 (±8.7) km/week]. Retrainers were cued, via a wireless accelerometer, to increase preferred step rate by 7.5% during eight training sessions performed in-field. Adherence with the prescribed step rate was assessed via mobile monitoring. Three-dimensional gait analysis was performed at baseline, after retraining, and at 1-month post-retraining. Retrainers increased step rate by 8.6% (P < 0.0001), reducing instantaneous vertical load rate (-17.9%, P = 0.003), average vertical load rate (-18.9%, P < 0.0001), peak hip adduction (2.9° ± 4.2 reduction, P = 0.005), eccentric knee joint work per stance phase (-26.9%, P < 0.0001), and per kilometer of running (-21.1%, P < 0.0001). Alterations in gait were maintained at 30 days. In the absence of any feedback, controls maintained their baseline gait parameters. The majority of retrainers were adherent with the prescribed step rate during in-field runs. Thus, in-field gait retraining, cueing a modest increase in step rate, was effective at reducing impact forces, peak hip adduction and eccentric knee joint work.

© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd.

What was unique about this study was that the researchers screened runners beforehand and only included those who exhibited high baseline impact loading rates. The value that constituted “high” corresponded to the average value for a cohort of runners with a history of tibial stress fracture. Thus, they were looking at runners who had impact loading rates that might put them at risk of injury and therefore might benefit most from a protocol that increases step rate.

Methods

The researchers measured a variety of biomechanical variables both before and after a cadence training protocol. The protocol consisted of 8 runs, during a portion of which subjects ran at a cadence target that was 7.5% higher than their baseline. The retraining utilized a faded-feedback design in which cadence feedback was provided during runs 1-3, 5, and 7. The rationale for this was that it might help subjects to internalize cadence changes if feedback was gradually removed.

Cadence feedback was accomplished by using a Garmin footpod that displayed real-time cadence on a GPS watch. A control group went through the same protocol but was not instructed to increase their cadence (it was not displayed on their watch). All subjects returned again 30 days after the end of the 8-run protocol to see if any changes would be retained in the absence of continued cadence training.

Results

After the 8 run protocol the gait retraining group exhibited significantly increased stride rate. Stride rate for this group was on average 166.5 steps/min at baseline, 180.8 steps/min post-training, and 180.6 steps./min 30 days post-training. This shows that the subjects had retained the new, higher stride rate even 30 days after feedback was removed.

The experimental group also exhibited significant reductions in instantaneous impact loading rate (-18.9%), average vertical loading rate (-17.9%), peak hip adduction (about 3 degrees less), and eccentric knee joint work (-26.9%). These changes were all retained 30 days later.

No differences between baseline, post-training, or 30 days post training were observed in the control group.

Comparisons between the two groups revealed that stride rate was significantly higher, and loading rates were significantly lower in the experimental group both post-training and 30 days post-training. Peak hip adduction and eccentric knee joint work did not differ between the groups.

Comments

I like this study because it provides additional evidence that runners can make lasting changes to their cadence outside of the lab/clinic setting, thus saving both time and money. Consistent with other studies, this one also shows that increasing cadence can alter biomechanical variables that have been associated with risk of injuries like tibial stress fractures, patellofemoral pain, and iliotibial band syndrome. I also like that they pre-screened subjects to include only those who exhibited baseline loading rates consistent with a population of runners with a history of stress fractures. Thus, they were looking at a population of runners who might benefit most from such an intervention.

Given the increasing number of devices that are capable of displaying running cadence (including both phone apps and GPS watches), the protocol here seems like a logical option that should be accessible to most runners. I routinely utilize the cadence function on my Garmin 620 (mainly out of curiosity), and the iSmoothRun app will give you this information on a phone. I haven’t actively attempted to retrain my own cadence, but if I were going to try I think I’d prefer going this route over using something like a metronome or music with a target tempo.

The missing link, however, continues to be the lack of interventional studies that take runners suffering from one of the injuries mentioned above through a gait retraining protocol like the one studied here. This will go a long way toward bridging the gap between gait retraining, resulting biomechanical changes, and their potential role in injury resolution. In the meantime, attempting to increase cadence a bit might be a worthwhile experiment if you suffer/have suffered from tibial stress fracture, PFPS, or ITBS.

One of the challenges with gait retraining methods employed in clinical settings is that the equipment required is not always accessible to runners outside of the clinic (e.g., accelerometers, force treadmills). This can require repeated, often costly appointments, especially if they are not covered by insurance. Cadence training has the advantage of being easy to do outside of the clinical setting using smartphone apps, metronomes, or fitness watches that can measure cadence in real time.

Questions that arise are whether runners can make lasting modifications to their running cadence on their own, whether those modifications might yield the same loading changes observed acutely in the lab, and whether there are any negative consequences to cadence change such as a loss of running efficiency. A recent study led by Jocelyn Hafer of the University of Massachusetts addressed each of these issues. Published in the Journal of Sports Sciences, the study looks at how a 6-week cadence training protocol effects running biomechanics and efficiency.

Here is the abstract of the study:

The effect of a cadence retraining protocol on running biomechanics and efficiency: a pilot study

Jocelyn F. Hafer, Allison M. Brown, Polly deMille, Howard J. Hillstrom & Carol Ewing Garber

Journal of Sports Sciences, Volume 33, Issue 7, 2015

Abstract

Many studies have documented the association between mechanical deviations from normal and the presence or risk of injury. Some runners attempt to change mechanics by increasing running cadence. Previous work documented that increasing running cadence reduces deviations in mechanics tied to injury. The long-term effect of a cadence retraining intervention on running mechanics and energy expenditure is unknown. This study aimed to determine if increasing running cadence by 10% decreases running efficiency and changes kinematics and kinetics to make them less similar to those associated with injury. Additionally, this study aimed to determine if, after 6 weeks of cadence retraining, there would be carryover in kinematic and kinetic changes from an increased cadence state to a runner’s preferred running cadence without decreased running efficiency. We measured oxygen uptake, kinematic and kinetic data on six uninjured participants before and after a 6-week intervention. Increasing cadence did not result in decreased running efficiency but did result in decreases in stride length, hip adduction angle and hip abductor moment. Carryover was observed in runners’ post-intervention preferred running form as decreased hip adduction angle and vertical loading rate.

Methods

In this study, researchers screened for runners with a cadence between 150 and 170 steps per minute, which is on the low end of the range of cadences typically observed (six runners were included in the study, so a relatively small sample size). They then had these runners train for 6 weeks with a goal of running at least 50% of their miles at a cadence 10% higher than their baseline. To accomplish this they could use either a metronome app or songs with tempos corresponding to the target cadence. A variety of metabolic and biomechanical measurements were taken before and after the training period to assess the effects of the retraining protocol. They also looked at immediate effects of increasing cadence at the outset to see if changes would be similar to those previously reported in the literature (they were).

Results

After the six week training protocol, the runners preferred cadence had increased significantly from around 166 steps/min on average to around 170 steps/min. Thus, though they did not adopt a full 10% increase through the training, cadence did increase significantly. Along with the preferred cadence increase, the runners also exhibited reduced ankle dorsiflexion at contact (less pronounced heel strike), reduced peak hip adduction angle (the thigh did not angle medially/inward as much), and reduced vertical loading rate. Running efficiency was not significantly different.

Comments

Though the sample size was small (the authors admit this and refer to it as a pilot study in the title), it is encouraging to see that cadence training can lead runners to modify their form, and that changes can alter variables associated with injury (especially with the small cadence increase that they observed). Reduction in hip adduction angle could benefit those with patellofemoral pain or ITBS, and reduced vertical loading rate could benefit those with a history of tibial stress fracture. It’s also encouraging that these benefits are accrued without a hit to running efficiency.

I would like to see this study completed with a larger sample of runners, but it was encouraging the significant changes were found even with this small sample. I’d also like to see some analysis of which method of cadence training is most effective for a runner (music, metronome, real-time feedback via an app or watch).

All things considered, these results do suggest that increasing cadence is one option to consider for those experiencing ilitibial band syndrome, patellofemoral pain syndrome, or past tibial stress fracture. I also like that this is an intervention that runners can undertake on their own without the need to make repeated visits to a clinic. A useful next step would be an interventional study where runners experiencing one of these conditions undergo a cadence training program and symptoms are monitored before and after.

What I’m going to do is share data from a few runs that revealed some interesting patterns to give an idea of what you can learn from the device. I still find the data more a curiosity than something I would use to make changes to my form, but it is interesting to play with the information!

First, let’s take some data from a near-PR 5K I ran back in April – below you’ll see the run summary info and GPS track, as well as graphs for cadence, vertical oscillation, and ground contact time (GCT) all estimated by the HRM-Monitor:

For frame of reference, here’s how Garmin breaks down running dynamics across a range of runners ((note: I’m not sure exactly how Garmin determined these groupings):

Looks like my cadence and ground contact time during the race were on the high end (mostly purple – >95 percentile), whereas vertical oscillation was more middle of the road. This suggests to me that I have a short, quick stride with a moderate amount of up and down movement (assuming the HRM-monitor provides an accurate measure of vertical oscillation).

What’s kind of cool in the above graphs is that right after the 10:00 mark we went up, over and down an overpass, then returned over that same overpass just after 13:20. You can see how my cadence, vertical oscillation, and GCT changed going up and down the hill. You can also see that my cadence trended upward, and oscillation trended downward over the course of the race (presumably in part as I sped up at the end).

For an even more interesting view of data from the same race, I pulled the following graph off of the Connect Stats app, which pulls data from Garmin Connect and lets you look at it in different ways:

In the above graph you can see how my cadence changed from the beginning (blue dots) to the end of the race (orange/brown dots). You can also see on the right are two downward loops that probably represent the overpass mentioned above. What I find particularly interesting is that at the end of the race my cadence shot up way higher than it was at the same pace earlier in the race – wondering if this is a way of compensating for fatigue (it was a pretty flat course except for the one overpass so terrain doesn’t seem to explain this)?

You can see a similar pattern with the vertical oscillation data in the graph below – greater oscillation early in the race, less toward the end, even at similar paces run earlier on – perhaps I had to compensate for reduced air time by increasing my turnover?

Here’s one more graph from a different run (also in the Spring) that I found interesting:

This one was from an easy training run, and I was puzzled by the higher vertical oscillation at the beginning of the run (light blue dots). After a bit of thinking over a few runs I realized that my left knee had been bothering me at the time. When it does I often accentuate a forefoot strike and stiffen my leg on that side at the beginning of runs. However, once it warms up I settle into my more typical stride mechanics. My guess is that the early numbers represent modified form until I warmed up. Pretty cool, and suggests that you can get some interesting data from a device like this. I’m still not sure how accurate the numbers are, but I do think they are at least representative of general patterns, and there seem to be plausible explanations for some of the unusual patterns observed.

I’m curious now to try out the HRM-Run monitor for a comparison between a few different shoes – perhaps Hokas versus something more minimal. May just have to do some experimenting!

Have any of you played with the running dynamics data on the Forerunner 620. Any interesting patterns observed?

Every once in awhile I’ll check my stats out on Garmin Connect, and I happened to glance at the cadence graph for the workout – I always forget that the 620 does on-board cadence recording. Anyway, here’s what the graph looked like:

Here’s the corresponding pace graph:

I was struck by the regularity of the pattern – it’s almost like shifting gears. My walking cadence was around 125 steps/min, easy run cadence (8:00-8:30 min/mile) about 179-181 steps/min, and my fast running cadence (5:30-5:40 min/mile) was 200-203 steps/min. The pattern repeated almost perfectly with each cycle – my body settled into it’s preferred cadence for each speed/gait.

I’ve written before about the “myth of the 180 cadence,” which is basically the erroneous suggestion that everyone should aim to run at a cadence of 180, and I covered the topic extensively in my book. Though I do tend to run right at 180 at my easy pace, my cadence clearly changes with speed. Once I push the pace under 6:00/mile I’m up around 200 steps/min. 180 is not a magic number, and it’s not a number that every runner should aim for at all times.

Knowledge that cadence varies with speed influences my approach to dealing with runners as a coach and in the clinic. For example I’ve found that with my beginner 5K group cadences are often quite low – many are under 160 steps/min. But, I’m hesitant to change this since they are typically running 11:00/mile pace or slower when they first start out. Cadence will vary with pace, and this needs to be taken into account when assessing whether someone has a possible issue with overstriding (some beginning runners do, others do not).

In the clinic I often find that overstriding tends to be a bigger issue when runners increase speed. Since speed is a function of stride rate and stride length, some runners will increase speed by upping cadence, others will tend to try to reach out front with the leg and overstride (watch the final 0.10 of a local 5K and you will see a ton of this). Yet others can make use of good hip extension to increase stride length on the back side.

As an example of the latter, below are some data that I compiled from a video of the 5000m men’s final taken by physical therapist Jeff Moreno at the 2012 US Olympic Trials:

Based on the filming frame rate provided by Jeff, I was able to calculate cadence for each of the runners from the video (order is the order in which they pass the camera at this point of the race, not the final finishing order; details on the other measurements can be found here):

Aside from the fact that none of these elites were running with a cadence of 180, what I found interesting is that many have a cadence as many as 10 steps/minute or more lower than my fast running cadence at a pace that is well over a minute per mile faster (the winning time was 13:42, which translates to an average pace of 4:24 min/mile). I would likely attribute this to the elites having greater range of motion at the hip, with much greater hip extension than me. They can accomplish increased stride length through massive hip extension, whereas I, lacking that ability (I’m tight!), need to increase cadence to get to a higher speed.

(Note: If you are unclear what hip extension is, it’s basically the extent to which the thigh swings backward toward the end of stance through toe-off – the picture below of Ryan Hall demonstrates hip extension well.)

Supporting this, here’s another graph from a 4×400 workout I did last week in which I went near max effort on the run reps (which, sadly, is still 10-20 seconds per mile slower than the elites in the race above manged for a full 5k!):

At 4:40-4:50 min/mile my cadence shoots up to 210-220 steps/min. And you can also see the effect of fatigue coming in a bit in the purple dots above – on a few of the reps my cadence starts out high then trails downward as I slow down during the lap. I have some very interesting data from a 5K race I ran which I’ll share sometime as well – has some bearing on fatigue and gait change.

So what did I learn from all of this? For me, cadence is very consistent at a given pace on the track, and I seem to rely on cadence increase to accomplish faster speed than the elite 5000m runners do. Makes me wonder if I spent some time working on hip drive and hip mobility if that would benefit my speed…

In the video, Ian does a nice job explaining what limb stiffness is, how it is modified by varying surfaces/shoes, and what the implications for injury might be. It’s a topic that fascinates me, and one that I have pondered a lot as it relates to my own footwear preferences.

Gather the family, cook some popcorn, and enjoy! (sadly, my wife was not interested in watching this with me last night…sorry Ian!)