November 2017

Assessing runners’ gait using wearable sensors

Wearable sensors allow for the collection of running biomechanics data outside the laboratory in natural training environments, enabling clinicians to collect a large volume of information in a relatively short time to help identify and manage individuals who may be at risk for running-related injuries.

By Rachel Koldenhoven, MEd, ATC; and Alex DeJong, MEd, ATC

The majority of traditional running gait assessments represented in the medical literature have been captured in a laboratory setting. The laboratory equipment and techniques used to analyze gait are capable of producing precise measurements; however, these assessments are generally time consuming, expensive, and clinically impractical. In addition, participants in these studies are typically under direct supervision in a controlled environment and wear some type of markers or sensors on the body while they run, factors that could potentially alter the way someone runs.

The results of these lab assessments should not be discounted, as they still provide useful information that can be translated to clinical practice. However, exploring other measurement tools that can be used during typical training conditions may provide unique insight on running mechanics in more natural environments.

Wearable sensors

Wearables sensors (sometimes simply called “wearables”) such as pedometers, accelerometers, global positioning systems (GPS), and inertial measurement units (IMUs) have become increasingly popular in recent years. They are typically small external devices that are used to assess different aspects of gait depending on the type of sensor(s) that is chosen. Each wearable sensor used to measure running mechanics includes one or more internal measurement devices.

Wearables can collect a range of outcome measures that can be used for monitoring training, implementing interventions, and enhancing athletic performance in runners.

Pedometers are commonly used to measure step count, distance travelled, and energy expenditure. Accelerometers are slightly more advanced, and are capable of recording data in up to three planes of motion. These data can be used to estimate frequency, intensity, and duration of physical activity. Gyroscopes feature a wheel or disc that spins around an axis and can be used to measure angular velocities. Similarly, IMUs use a 3D local coordinate system, typically with respect to a global fixed coordinate system such as a GPS. IMUs are capable of measuring acceleration, angular rates, and magnetic field vectors. Lastly, GPS uses radio signals from satellites to reach GPS devices and uses geometry to estimate location in three dimensions. GPS is typically used to assess position, navigation, and timing of movement.

Advantages of wearable sensors

Wearable sensors are advantageous as they allow for data collection with lightweight technology outside of the laboratory setting. It has been shown that individuals—whether adults or children—have different gait patterns when walking outside the laboratory and not under direct supervision or examination, compared with walking in a laboratory environment.1 Additionally, running outdoors introduces alterations in terrain and environment that cannot be replicated with treadmill running.2 Changes in surface type3 and the introduction of sloped terrain4,5 have been found to influence lower extremity running mechanics; therefore, collecting information in these environments is warranted. Gait speed is also subject to instantaneous changes during overground running that cannot be easily replicated with treadmill running.6

Thus, data obtained from wearable sensors may be more generalizable than laboratory data, as the wearable-based information is collected during natural running conditions and allows for collections and analysis of a larger number of steps.

There is potential to use wearable sensors as a mode of feedback as well as for overall assessment of running.7 For example, runners may use a wrist watch and a foot pod simultaneously to receive information such as step rate that can be manipulated during the run.7

 

Disadvantages of wearable sensors

Data collection outside of the laboratory setting improves the external validity and generalizability of the information. The internal validity suffers, however, as there is less overall control of the collection environment. Internal validity refers to the “soundness” of the study and a lack of confounding variables. In other words, how strong are the research methods? Conversely, external validity relates to how well the study findings can be applied to the real world. Lastly, laboratory equipment serves as the gold standard for assessing most gait-related variables; therefore, information from sensors may be less valid than traditional laboratory outputs.8-10

The wearable sensors that are currently available do not yet provide information on muscle activity or other measures that can be assessed in lab environments. Being able to obtain this information outside the lab would be beneficial as it would provide a more comprehensive view of running movement patterns. The extent of muscle activation and the duration of muscle activity during running could also provide useful information. This information could be helpful for understanding how injured runners use their muscles differently than an uninjured runner. Understanding these variations could help guide clinical targeted strength and gait retraining protocols.

The decision to use wearable sensors or laboratory equipment to analyze running biomechanics should be made on a case-by-case basis, making it important to determine which resources are available and what the specific goals are for evaluating running gait.

Measuring running gait with wearables

Wearable technology can collect a variety of outcomes measures that can be used for monitoring training, implementing intervention techniques, and enhancing performance. Table 1 provides a list of types of wearable sensors and what they can measure.

The most common outcome measurements include number of steps, distance, speed, duration of activity, and energy expenditure during activity. However, more sophisticated forms of wearable sensors use accelerometers and gyroscopes to additionally analyze kinematic and kinetic data. Several examples of kinematic data that can be collected at the ankle include pronation excursion, maximum pronation velocity, and sagittal plane stance excursion. Kinetic data that can be obtained from wearable sensors include contact time, impact forces, and braking forces.

Synthesized information on running-related risk factors has shown altered kinematics and kinetics are associated with the development of running-related injuries.11 Therefore, concomitant collection of these variables could provide clinicians with a more global view of runners’ movement profiles that could help to identify gait-related risk factors for injury and ultimately guide more targeted clinical interventions.

Several laboratory studies have associated tibial stress fractures in runners with increased loading rate of vertical forces12-14 and excessive hip adduction.13 Information from laboratory studies has the potential to translate to measurements or interventions in the field setting. For example, one study used wearable sensors to implement in-field gait retraining for individuals demonstrating risk factors for tibial stress fractures.7 The intervention group received feedback about their steps per minute via a wristwatch during running and were instructed to increase their preferred step rate by 7.5%, while the control group received no instructions to change gait.7 Three-dimensional motion capture in a laboratory was used at baseline, postintervention, and one month postintervention, but all training sessions were unsupervised and done in the field.

The intervention group could alter their running gait by increasing their step rate (9% above their baseline rate), which was associated with decreases in loading rate measures (18%-19%), peak hip adduction (3º), knee joint work per stance (27%), and per kilometer of running (21%). In addition, the gait training group was able to maintain these changes one month postintervention.7 Meanwhile, runners in the control group exhibited no changes in gait parameters during the study.

Our research

A wearable foot pod that uses an accelerometer and gyroscope to measure ankle kinetics, kinematics, and spatiotemporal variables has been tested at our institution; we presented some of the preliminary findings at the National Athletic Trainers’ Association 2017 annual meeting in Houston, TX.10,15,16 We performed a concurrent validation study in a laboratory setting to compare the wearable sensor to traditional 3D motion capture.10 This study was performed to ensure justified use of the wearable sensor for research in the field.

Excellent ICC (intraclass correlation) estimate agreements with small mean differences were found for maximum pronation velocity, cycle time, and contact time.10 This suggests the wearable device we studied produces acceptable information for those specific variables. Pronation excursion did not have the same level of consistency between the two measurement systems; therefore, pronation excursion assessments performed using the wearable device we studied should be interpreted with caution.10 We would not recommend direct comparisons between the variables from the wearable device and the 3D motion capture system.

Following the validation study, we determined it may be beneficial to use the wearable device to capture running at slow and fast running speeds, on various surface types, and in different populations.15,16 To compare changes associated with different self-selected speeds (slow, fast) and different surfaces (track, grass), runners completed a 1600-m run in each condition. The sensors were able to detect increased stride length, decreased cycle time, and decreased contact time during the fast runs compared with the slow runs, regardless of surface type.

Impact gs were described as the vertical component of force and braking gs were described as the horizontal component of force. The pronation excursion variable was described as the amount of pronation range of motion from initial contact to maximum pronation. As speed increased, pronation excursion, maximum pronation velocity, impact gs, and braking gs also increased. These findings coincide with traditional laboratory study findings that found increased running speed was related to increased sagittal plane ankle motion and increased peak impact forces.17,18

When comparing the surface types (track vs grass field), the sensors were able to detect decreases in cycle time and contact time on the track surface compared with the grass surface.15 If the track is considered a firm surface and the grass a softer surface, then our findings are consistent with those of a similar study in which contact time was about 12 ms shorter for firm surface conditions than for softer surfaces.2,3

With regard to kinematic and kinetic measures, the sensors detected higher pronation excursion, maximum pronation velocity, impact gs, and braking gs on the track than on grass.15 Increased forces on the body have been identified when running on harder surfaces as well, which support the results of the increased impact and braking gs found in our study.2

When considering all of our study results, the sensors identified the expected changes in spatiotemporal, kinematic, and kinetic running mechanics when participants ran at different speeds and on different surfaces.15 This wearable sensor appears to be well suited for clinical use as well as for answering future sports medicine research questions.

Assessing ankle instability with wearables

The wearable sensor described has been used to measure differences in spatiotemporal, kinetic, and kinematic measures between individuals with chronic ankle instability (CAI) and healthy controls.16 Participants ran 1600 m at a self-selected slow pace and 1600 m at a self-selected fast pace around a track. The CAI group demonstrated decreased maximum pronation velocity (~90-110º/sec) and pronation excursion (~4º), and increased contact time (~40-50 ms) compared with the control group throughout the 1600-m run. During the first 400 m of the run, the CAI group had higher impact gs (~1g).

The authors speculated differences in running mechanics could be due to more constrained movement patterns used by the CAI group relative to controls, something previous laboratory studies of walking and jogging have indentified.19 Clinicians monitoring gait mechanics while treating individuals with CAI should be aware of the existing differences in running mechanics and potential increases in physiological stress to either the ankle joint or other lower extremity structures in these individuals.

Conclusions

The advent of wearable sensors allows for running data collection outside the laboratory to examine running biomechanics in natural training environments. Sensors worn on the shoes can collect a range of information, from simple measures such as step count and distance to more sophisticated measures such as spatiotemporal and impact data. Wearable foot pods are a reliable means of collecting multiple gait parameters concomitantly and have started to be transitioned from the laboratory setting to clinical applications. This enables clinicians to collect a large volume of running data in a relatively short time to help identify individuals who may be at risk for running injuries and can ultimately influence clinical intervention for improved patient outcomes.

Rachel Koldenhoven, MEd, ATC, and Alex DeJong, MEd, ATC, are PhD students in the Department of Kinesiology – Sports Medicine at the University of Virginia in Charlottesville.

REFERENCES
  1. Najafi B, Khan T, Wrobel J. Wearable sensors and its advantages for gait analysis. 2011 Annu Int Conf IEEE Eng Med Biol Soc EMBC 2011:6507-6510.
  2. Tessutti V, Trombini-Souza F, Ribeiro AP, et al. In-shoe plantar pressure distribution during running on natural grass and asphalt in recreational runners. J Sci Med Sport 2010;13(1):151-155.
  3. Tessutti V, Ribeiro AP, Trombini-Souza F, Sacco IC. Attenuation of foot pressure during running on four different surfaces: asphalt, concrete, rubber, and natural grass. J Sports Sci 2012;30(14):1545-1550.
  4. McIntosh AS, Beatty KT, Dwan LN, Vickers DR. Gait dynamics on an inclined walkway. J Biomech 2006;39(13):2491-2502.
  5. Tulchin K, Orendurff M, Karol L. The effects of surface slope on multi-segment foot kinematics in healthy adults. Gait Posture 2010;32(4):446-450.
  6. Schache AG, Blanch PD, Dorn TW, et al. Effect of running speed on lower limb joint kinetics. Med Sci Sports Exerc 2011;43(7):1260-1271.
  7. Willy RW, Buchenic L, Rogacki K, et al. In-field gait retraining and mobile monitoring to address running biomechanics associated with tibial stress fracture. Scand J Med Sci Sports 2016;26(2):197-205.
  8. Chung PYM, Ng GYF. Comparison between an accelerometer and a three-dimensional motion analysis system for the detection of movement. Physiotherapy 2012;98(3):256-259.
  9. Ferguson T, Rowlands AV, Olds T, Maher C. The validity of consumer-level, activity monitors in healthy adults worn in free-living conditions: a cross-sectional study. Int J Behav Nutr Phys Act 2015;12:42.
  10. Koldenhoven RM. Validation of a wearable sensor for measuring running biomechanics. Presented at the National Athletic Trainers’ Association 68th Clinical Symposia and AT Expo, Houston, TX, June 2017.
  11. Napier C, Cochrane CK, Taunton JE, Hunt MA. Gait modifications to change lower extremity gait biomechanics in runners: a systematic review. Br J Sports Med 2015;49(21):1382-1388.
  12. Milner CE, Ferber R, Pollard CD, et al. Biomechanical factors associated with tibial stress fracture in female runners. Med Sci Sports Exerc 2006;38(2):323-328.
  13. Pohl MB, Mullineaux DR, Milner CE, et al. Biomechanical predictors of retrospective tibial stress fractures in runners. J Biomech 2008;41(6):1160-1165.
  14. Zadpoor AA, Nikooyan AA. The relationship between lower-extremity stress fractures and the ground reaction force: A systematic review. Clin Biomech 2011;26(1):23-28.
  15. Hollis C. Gait mechanics as measured by a wearable sensor during running at two speeds on different surfaces. Presented at National Athletic Trainers’ Association 68th Clinical Symposia and AT Expo, Houston, TX, June 2017.
  16. Colapietro M, Fraser JJ, Resch JE, Hertel J. Running mechanics during 1600-meter track runs in young adults with and without chronic ankle instability. Presented at National Athletic Trainers’ Association 68th Clinical Symposia and AT Expo, Houston, TX, June 2017.
  17. Nilsson J. Ground reaction forces at different speeds of human walking and running. Acta Physiol Scand 1989;136(2):217-227.
  18. Kyröläinen H, Belli A, Komi PV. Biomechanical factors affecting running economy. Med Sci Sports Exerc 2001;33(8):1330-1337.
  19. Herb CC, Hertel J. Shank-rearfoot joint coupling in young adults with chronic ankle instability: a cross-correlation analysis. J Sports Med Phys Fitness 2015;55(6):639-646.
(Visited 134 times, 1 visits today)

Leave a Reply

Your email address will not be published. Required fields are marked *

Spam Blocker * Time limit is exhausted. Please reload CAPTCHA.