Ground reaction forces are often used by sport scientists and clinicians to analyze the mechanical risk-factors of running related injuries or athletic performance during a running analysis. An interesting ground reaction force-derived variable to track is the maximal vertical instantaneous loading rate (VILR). This impact characteristic is traditionally derived from a fixed force platform, but wearable inertial sensors nowadays might approximate its magnitude while running outside the lab. The time-discrete axial peak tibial acceleration (APTA) has been proposed as a good surrogate that can be measured using wearable accelerometers in the field. This paper explores the hypothesis that applying machine learning to time continuous data (generated from bilateral tri-axial shin mounted accelerometers) would result in a more accurate estimation of the VILR. Therefore, the purpose of this study was to evaluate the performance of accelerometer-based predictions of the VILR with various machine learning models trained on data of 93 rearfoot runners. A subject-dependent gradient boosted regression trees (XGB) model provided the most accurate estimates (mean absolute error: 5.39 ± 2.04 BW•s −1 , mean absolute percentage error: 6.08%). A similar subject-independent model had a mean absolute error of 12.41 ± 7.90 BW•s −1 (mean absolute percentage error: 11.09%). All of our models had a stronger correlation with the VILR than the APTA (p < 0.01), indicating that multiple 3D acceleration features in a learning setting showed the highest accuracy in predicting the lab-based impact loading compared to APTA.

Methods to reduce impact in distance runners have been proposed based on real-time auditory feedback of tibial acceleration. These methods were developed using treadmill running. In this study, we extend these methods to a more natural environment with a proof-of-concept. We selected ten runners with high tibial shock. They used a music-based biofeedback system with headphones in a running session on an athletic track. The feedback consisted of music superimposed with noise coupled to tibial shock. The music was automatically synchronized to the running cadence. The level of noise could be reduced by reducing the momentary level of tibial shock, thereby providing a more pleasant listening experience. The running speed was controlled between the condition without biofeedback and the condition of biofeedback. The results show that tibial shock decreased by 27% or 2.96 g without guided instructions on gait modification in the biofeedback condition. The reduction in tibial shock did not result in a clear increase in the running cadence. The results indicate that a wearable biofeedback system aids in shock reduction during over-ground running. This paves the way to evaluate and retrain runners in over-ground running programs that target running with less impact through instantaneous auditory feedback on tibial shock.

Background Running retraining with the use of biofeedback on an impact measure has been executed or evaluated in the biomechanics laboratory. Here, the execution and evaluation of feedback‐driven retraining are taken out of the laboratory. Purpose To determine whether biofeedback can reduce the peak tibial acceleration with or without affecting the running cadence in a 3‐week retraining protocol. Study Design Quasi‐randomized controlled trial. Methods Twenty runners with high peak tibial acceleration were allocated to either the retraining (n = 10, 32.1 ± 7.8 years, 10.9 ± 2.8 g) or control (n = 10, 39.1 ± 10.4 years, 13.0 ± 3.9 g) groups. They performed six running sessions in an athletic training environment. A body‐worn system collected axial tibial acceleration and provided real‐time feedback. The retraining group received music‐based biofeedback in a faded feedback scheme. Pink noise was superimposed on tempo‐synchronized music when the peak tibial acceleration was ≥70% of the runner's baseline. The control group received tempo‐synchronized music, which acted as a placebo for blinding purposes. Speed feedback was provided to obtain a stable running speed of ~2.9 m·s−1. Peak tibial acceleration and running cadence were evaluated. Results A significant group‐by‐feedback interaction effect was detected for peak tibial acceleration. The experimental group had a decrease in peak tibial acceleration by 25.5% (mean: 10.9 ± 2.8 g versus 8.1 ± 3.9 g, p = 0.008, d = 1.08, mean difference = 2.77 [0.94, 4.61]) without changing the running cadence. The control group had no statistically significant change in peak tibial acceleration nor in running cadence. Conclusion The retraining protocol was effective at reducing the peak tibial acceleration in high‐impact runners by reacting to music‐based biofeedback that was provided in real time per wearable technology in a training environment. This reduction magnitude may have meaningful influences on injury risk.

ObjectivesRecreational runners show a large interindividual variation in spatiotemporal characteristics. This research focused on slow runners and intended: (1) to document the variance in duty factor (DF) between runners in a real-life running setting and (2) examine whether the interindividual variation in DF and stride frequency (SF) relates to differences in external loading parameters between runners.MethodsSpatiotemporal characteristics of 23 slow runners (ie, <2.6 m/s) were determined during a 5.2 km running event. To relate the interindividual variation in DF and SF to differences in external forces between runners (maximal vertical ground reaction force (FzMax), peak braking force (PBF) and vertical instantaneous loading rate (VILR)), 14 of them were invited to the lab. They ran at 1.9 m/s on a treadmill while ground reaction forces were recorded. A multiple linear regression analysis was conducted to investigate the effect of DF and SF on external force measures.ResultsDF between slow runners varied from 42.50% to 56.49% in a recreational running event. DF was found to be a significant predictor of FzMax (R²=0.755) and PBF (R²=0.430). SF only improved the model for PBF, but to a smaller extent than DF (R² change=0.191). For VILR, neither DF nor SF were significant predictors.ConclusionExternal forces are lower in recreational runners that run with higher DFs and slightly lower SFs. These findings may be important for injury prevention purposes, especially directed to recreational runners that are more prone to overuse injuries.

Background: Gait event detection of the initial contact and toe off is essential for running gait analysis, allowing the derivation of parameters such as stance time. Heuristic-based methods exist to estimate these key gait events from tibial accelerometry. However, these methods are tailored to very specific acceleration profiles, which may offer complications when dealing with larger data sets and inherent biological variability. Research question: Can a structured machine learning approach achieve a more accurate prediction of running gait event timings from tibial accelerometry, compared to the previously utilised heuristic approaches? Methods: Force-based event detection acted as the criterion measure in order to assess the accuracy, repeatability and sensitivity of the predicted gait events. 3D tibial acceleration and ground reaction force data from 93 rearfoot runners were captured. A heuristic method and two structured machine learning methods were employed to derive initial contact, toe off and stance time from tibial acceleration signals. Results: Both a structured perceptron model (median absolute error of stance time estimation: 10.00 ± 8.73 ms) and a structured recurrent neural network model (median absolute error of stance time estimation: 6.50 ± 5.74 ms) significantly outperformed the existing heuristic approach (median absolute error of stance time estimation: 11.25 ± 9.52 ms). Thus, results indicate that a structured recurrent neural network machine learning model offers the most accurate and consistent estimation of the gait events and its derived stance time during level overground running. Significance: The machine learning methods seem less affected by intra-and inter-subject variation within the data, allowing for accurate and efficient automated data output during rearfoot overground running. Furthermore offering possibilities for real-time monitoring and biofeedback during prolonged measurements, even outside the laboratory.