Search

Background: Running is a popular physical activity that benefits health; however, running surface characteristics may influence loading impact and injury risk. Machine learning algorithms could automatically identify running surface from wearable motion sensors to quantify running exposures, and perhaps loading and injury risk for a runner., Research Question: (1) How accurately can machine learning algorithms identify surface type from three-dimensional accelerometer sensors? (2) Does the sensor count (single or two-sensor setup) affect model accuracy?, Methods: Twenty-nine healthy adults (23.3 ± 3.6 years, 1.8 ± 0.1 m, and 63.6 ± 8.5 kg) participated in this study. Participants ran on three different surfaces (concrete, synthetic, woodchip) while fit with two three-dimensional accelerometers (lower-back and right tibia). Summary features (n = 208) were extracted from the accelerometer signals. Feature-based Gradient Boosting (GB) and signal-based deep learning Convolutional Neural Network (CNN) models were developed. Models were trained on 90% of the data and tested on the remaining 10%. The process was repeated five times, with data randomly shuffled between train-test splits, to quantify model performance variability., Results: All models and configurations achieved greater than 90% average accuracy. The highest performing models were the two-sensor GB and tibia-sensor CNN (average accuracy of 97.0 ± 0.7 and 96.1 ± 2.6%, respectively)., Significance: Machine learning algorithms trained on running data from a single- or dual-sensor accelerometer setup can accurately distinguish between surfaces types. Automatic identification of surfaces encountered during running activities could help runners and coaches better monitor training load, improve performance, and reduce injury rates., (Copyright © 2019 Elsevier B.V. All rights reserved.)

Maintaining stability under dynamic conditions is an inherent challenge to bipedal running. This challenge may impose an energetic cost (Ec) thus hampering endurance running performance, yet the underlying mechanisms are not clear. Wireless triaxial trunk accelerometry is a simple tool that could be used to unobtrusively evaluate these mechanisms. Here, we test a cost of instability hypothesis by examining the contribution of trunk accelerometry-based measures (triaxial root mean square, step and stride regularity, and sample entropy) to interindividual variance in Ec (J/m) during treadmill running. Accelerometry and indirect calorimetry data were collected concurrently from 30 recreational runners (16 men; 14 women) running at their highest steady-state running speed (80.65 ± 5.99% V̇o 2max ). After reducing dimensionality with factor analysis, the effect of dynamic stability features on Ec was evaluated using hierarchical multiple regression analysis. Three accelerometry-based measures could explain an additional 10.4% of interindividual variance in Ec after controlling for body mass, attributed to anteroposterior stride regularity (5.2%), anteroposterior root mean square ratio (3.2%), and mediolateral sample entropy (2.0%). Our results lend support to a cost of instability hypothesis, with trunk acceleration waveform signals that are 1) more consistent between strides anteroposterioly, 2) larger in amplitude variability anteroposterioly, and 3) more complex mediolaterally and are energetically advantageous to endurance running performance. This study shows that wearable trunk accelerometry is a useful tool for understanding the Ec of running and that running stability is important for economy in recreational runners. NEW & NOTEWORTHY This study evaluates and more directly lends support to a cost of instability hypothesis between runners. Moreover, this hypothesis was tested using a minimalist setup including a single triaxial trunk mounted accelerometer, with potential transferability to biomechanical and performance analyses in typical outdoor settings.

Medial tibial stress syndrome (MTSS) is a common overuse running injury with pathomechanics likely to be exaggerated by fatigue. Wearable accelerometry provides a novel alternative to assess biomechanical parameters continuously while running in more ecologically valid settings. The purpose of this study was to determine the influence of outdoor running fatigue and MTSS on both dynamic loading and dynamic stability derived from trunk and tibial accelerometery. Runners with (n=14) and without (n=16) history of MTSS performed an outdoor fatigue run of 3200m. Accelerometer-based measures averaged per lap included dynamic loading of the trunk and tibia (i.e. axial peak positive acceleration, signal power magnitude, and shock attenuation) as well as dynamic trunk stability (i.e. tri-axial root mean square ratio, step and stride regularity, and sample entropy). Regression coefficients from generalised estimating equations were used to evaluate group by fatigue interactions. No evidence could be found for dynamic loading being higher with fatigue in runners with MTSS history (all measures p>0.05). One significant group by running fatigue interaction effect was detected for dynamic stability. Specifically, in MTSS only, decreases mediolateral sample entropy i.e. loss of complexity was associated with running fatigue (p<0.01). The current results indicate that entire acceleration waveform signals reflecting mediolateral trunk control is related to MTSS history, a compensation that went undetected in the non-fatigued running state. We suggest that a practical outdoor running fatigue protocol that concurrently captures trunk accelerometry-based movement complexity warrants further prospective investigation as an in-situ screening tool for MTSS individuals., (Copyright © 2017 Elsevier B.V. All rights reserved.)

Despite frequently declared benefits of using wireless accelerometers to assess running gait in real-world settings, available research is limited. The purpose of this study was to investigate outdoor surface effects on dynamic stability and dynamic loading during running using tri-axial trunk accelerometry. Twenty eight runners (11 highly-trained, 17 recreational) performed outdoor running on three outdoor training surfaces (concrete road, synthetic track and woodchip trail) at self-selected comfortable running speeds. Dynamic postural stability (tri-axial acceleration root mean square (RMS) ratio, step and stride regularity, sample entropy), dynamic loading (impact and breaking peak amplitudes and median frequencies), as well as spatio-temporal running gait measures (step frequency, stance time) were derived from trunk accelerations sampled at 1024Hz. Results from generalized estimating equations (GEE) analysis showed that compared to concrete road, woodchip trail had several significant effects on dynamic stability (higher AP ratio of acceleration RMS, lower ML inter-step and inter-stride regularity), on dynamic loading (downward shift in vertical and AP median frequency), and reduced step frequency (p<0.05). Surface effects were unaffected when both running level and running speed were added as potential confounders. Results suggest that woodchip trails disrupt aspects of dynamic stability and loading that are detectable using a single trunk accelerometer. These results provide further insight into how runners adapt their locomotor biomechanics on outdoor surfaces in situ., (Copyright © 2016 Elsevier B.V. All rights reserved.)