New Considerations for Collecting Biomechanical Data Using Wearable Sensors: How Does Inclination Influence the Number of Runs Needed to Determine a Stable Running Gait Pattern?

Use of subject-specific models to detect fatigue-related changes in running biomechanics: a random forest approach

Running biomechanics are affected by fatiguing or prolonged runs. However, no evidence to date has conclusively linked this effect to running-related injury (RRI) development or performance implications. Previous investigations using subject-specific models in running have demonstrated higher accuracy than group-based models, however, this has been infrequently applied to fatigue. In this study, two experiments were conducted to determine whether subject-specific models outperformed group-based models to classify running biomechanics during non-fatigued and fatigued conditions. In the first experiment, 16 participants performed four treadmill runs at or around the maximal lactate steady state. In the second experiment, nine participants performed five prolonged runs using commercial wearable devices. For each experiment, two segments were extracted from each trial from early and late in the run. For each participant, a random forest model was applied with a leave-one-run-out cross-validation to classify between the early (non-fatigued) and late (fatigued) segments. Additionally, group-based classifiers with a leave-one-subject-out cross validation were constructed. For experiment 1, mean classification accuracies for the single-subject and group-based classifiers were 68.2 ± 8.2% and 57.0 ± 8.9%, respectively. For experiment 2, mean classification accuracies for the single-subject and group-based classifiers were 68.9 ± 17.1% and 61.5 ± 11.7%, respectively. Variable importance rankings were consistent within participants, but these rankings differed from each participant to those of the group. Although the classification accuracies were relatively low, these findings highlight the advantage of subject-specific classifiers to detect changes in running biomechanics with fatigue and indicate the potential of using big data and wearable technology approaches in future research to determine possible connections between biomechanics and RRI.

Minimum number of inertial measurement units needed to identify significant variations in walk patterns of overweight individuals walking on irregular surfaces

Gait data collection from overweight individuals walking on irregular surfaces is a challenging task that can be addressed using inertial measurement unit (IMU) sensors. However, it is unclear how many IMUs are needed, particularly when body attachment locations are not standardized. In this study, we analysed data collected from six body locations, including the torso, upper and lower limbs, to determine which locations exhibit significant variation across different real-world irregular surfaces. We then used deep learning method to verify whether the IMU data recorded from the identified body locations could classify walk patterns across the surfaces. Our results revealed two combinations of body locations, including the thigh and shank (i.e., the left and right shank, and the right thigh and right shank), from which IMU data should be collected to accurately classify walking patterns over real-world irregular surfaces (with classification accuracies of 97.24 and 95.87%, respectively). Our findings suggest that the identified numbers and locations of IMUs could potentially reduce the amount of data recorded and processed to develop a fall prevention system for overweight individuals.

Evaluating the difference in walk patterns among normal-weight and overweight/obese individuals in real-world surfaces using statistical analysis and deep learning methods with inertial measurement unit data

Unusual walk patterns may increase individuals' risks of falling. Anthropometric features of the human body, such as the body mass index (BMI), influences the walk patterns of individuals. In addition to the BMI, uneven walking surfaces may cause variations in the usual walk patterns of an individual that will potentially increase the individual's risk of falling. The objective of this study was to statistically evaluate the variations in the walk patterns of individuals belonging to two BMI groups across a wide range of walking surfaces and to investigate whether a deep learning method could classify the BMI-specific walk patterns with similar variations. Data collected by wearable inertial measurement unit (IMU) sensors attached to individuals with two different BMI were collected while walking on real-world surfaces. In addition to traditional statistical analysis tools, an advanced deep learning-based neural network was used to evaluate and classify the BMI-specific walk patterns. The walk patterns of overweight/obese individuals showed a greater correlation with the corresponding walking surfaces than the normal-weight population. The results were supported by the deep learning method, which was able to classify the walk patterns of overweight/obese (94.8 ± 4.5%) individuals more accurately than those of normal-weight (59.4 ± 23.7%) individuals. The results suggest that application of the deep learning method is more suitable for recognizing the walk patterns of overweight/obese population than those of normal-weight individuals. The findings from the study will potentially inform healthcare applications, including artificial intelligence-based fall assessment systems for minimizing the risk of fall-related incidents among overweight and obese individuals.

Between-Day Reliability of Commonly Used IMU Features during a Fatiguing Run and the Effect of Speed

The purpose of this study was to determine if fatigue-related changes in biomechanics derived from an inertial measurement unit (IMU) placed at the center of mass (CoM) are reliable day-to-day. Sixteen runners performed two runs at maximal lactate steady state (MLSS) on a treadmill, one run 5% above MLSS speed, and one run 5% below MLSS speed while wearing a CoM-mounted IMU. Trials were performed to volitional exhaustion or a specified termination time. IMU features were derived from each axis and the resultant. Feature means were calculated for each subject during non-fatigued and fatigued states. Comparisons were performed between the two trials at MLSS and between all four trials. The only significant fatigue state × trial interaction was the 25th percentile of the results when comparing all trials. There were no main effects for trial for either comparison method. There were main effects for fatigue state for most features in both comparison methods. Reliability, measured by an intraclass coefficient (ICC), was good-to-excellent for most features. These results suggest that fatigue-related changes in biomechanics derived from a CoM-mounted IMU are reliable day-to-day when participants ran at or around MLSS and are not significantly affected by slight deviations in speed.

Is This the Real Life, or Is This Just Laboratory? A Scoping Review of IMU-Based Running Gait Analysis

Inertial measurement units (IMUs) can be used to monitor running biomechanics in real-world settings, but IMUs are often used within a laboratory. The purpose of this scoping review was to describe how IMUs are used to record running biomechanics in both laboratory and real-world conditions. We included peer-reviewed journal articles that used IMUs to assess gait quality during running. We extracted data on running conditions (indoor/outdoor, surface, speed, and distance), device type and location, metrics, participants, and purpose and study design. A total of 231 studies were included. Most (72%) studies were conducted indoors; and in 67% of all studies, the analyzed distance was only one step or stride or <200 m. The most common device type and location combination was a triaxial accelerometer on the shank (18% of device and location combinations). The most common analyzed metric was vertical/axial magnitude, which was reported in 64% of all studies. Most studies (56%) included recreational runners. For the past 20 years, studies using IMUs to record running biomechanics have mainly been conducted indoors, on a treadmill, at prescribed speeds, and over small distances. We suggest that future studies should move out of the lab to less controlled and more real-world environments.

Continuous Analysis of Marathon Running Using Inertial Sensors: Hitting Two Walls?

Marathon running involves complex mechanisms that cannot be measured with objective metrics or laboratory equipment. The emergence of wearable sensors introduced new opportunities, allowing the continuous recording of relevant parameters. The present study aimed to assess the evolution of stride-by-stride spatio-temporal parameters, stiffness, and foot strike angle during a marathon and determine possible abrupt changes in running patterns. Twelve recreational runners were equipped with a Global Navigation Satellite System watch, and two inertial measurement units clamped on each foot during a marathon race. Data were split into eight 5-km sections and only level parts were analyzed. We observed gradual increases in contact time and duty factor as well as decreases in flight time, swing time, stride length, speed, maximal vertical force and stiffness during the race. Surprisingly, the average foot strike angle decreased during the race, but each participant maintained a rearfoot strike until the end. Two abrupt changes were also detected around km 25 and km 35. These two breaks are possibly due to the alteration of the stretch-shortening cycle combined with physiological limits. This study highlights new measurable phenomena that can only be analyzed through continuous monitoring of runners over a long period of time.

Real-time feedback by wearables in running: Current approaches, challenges and suggestions for improvements

Injuries and lack of motivation are common reasons for discontinuation of running. Real-time feedback from wearables can reduce discontinuation by reducing injury risk and improving performance and motivation. There are however several limitations and challenges with current real-time feedback approaches. We discuss these limitations and challenges and provide a framework to optimise real-time feedback for reducing injury risk and improving performance and motivation. We first discuss the reasons why individuals run and propose that feedback targeted to these reasons can improve motivation and compliance. Secondly, we review the association of running technique and running workload with injuries and performance and we elaborate how real-time feedback on running technique and workload can be applied to reduce injury risk and improve performance and motivation. We also review different feedback modalities and motor learning feedback strategies and their application to real-time feedback. Briefly, the most effective feedback modality and frequency differ between variables and individuals, but a combination of modalities and mixture of real-time and delayed feedback is most effective. Moreover, feedback promoting perceived competence, autonomy and an external focus can improve motivation, learning and performance. Although the focus is on wearables, the challenges and practical applications are also relevant for laboratory-based gait retraining.

Use of Machine Learning and Wearable Sensors to Predict Energetics and Kinematics of Cutting Maneuvers

Changes of directions and cutting maneuvers, including 180-degree turns, are common locomotor actions in team sports, implying high mechanical load. While the mechanics and neurophysiology of turns have been extensively studied in laboratory conditions, modern inertial measurement units allow us to monitor athletes directly on the field. In this study, we applied four supervised machine learning techniques (linear regression, support vector regression/machine, boosted decision trees and artificial neural networks) to predict turn direction, speed (before/after turn) and the related positive/negative mechanical work. Reference values were computed using an optical motion capture system. We collected data from 13 elite female soccer players performing a shuttle run test, wearing a six-axes inertial sensor at the pelvis level. A set of 18 features (predictors) were obtained from accelerometers, gyroscopes and barometer readings. Turn direction classification returned good results (accuracy > 98.4%) with all methods. Support vector regression and neural networks obtained the best performance in the estimation of positive/negative mechanical work (coefficient of determination R² = 0.42-0.43, mean absolute error = 1.14-1.41 J) and running speed before/after the turns (R² = 0.66-0.69, mean absolute error = 0.15-018 m/s). Although models can be extended to different angles, we showed that meaningful information on turn kinematics and energetics can be obtained from inertial units with a data-driven approach.