Search

Human walking gait is a personal story written by the body, a tool for understanding biological identity in healthcare and security. Gait analysis methods traditionally diverged between these domains but are now merging their complementary strengths to unlock new possibilities. Using large ground reaction force (GRF) datasets for gait recognition is a way to uncover subtle variations that define individual gait patterns. Previously, this was done by developing and evaluating machine learning models on the same individuals or the same dataset, potentially biasing findings towards population samples or walking conditions. This study introduces a new method for analysing gait variation across individuals, groups and datasets to explore how demographics and walking conditions shape individual gait patterns. Machine learning models were implemented using numerous configurations of four large walking GRF datasets from different countries (740 individuals, 7400 samples) and analysed using explainable artificial intelligence tools. Recognition accuracy ranged from 52 to 100%, with factors like footwear, walking speed and body mass playing interactive roles in defining gait. Models developed with individuals walking in personal footwear at multiple speeds effectively recognized novel individuals across populations and conditions (89-99% accuracy). Integrating force platform hardware and gait recognition software could be invaluable for reading the complex stories of human walking.

The studded outsole of a soccer boot provides additional traction to players to minimise the risk of slipping while performing high-speed manoeuvres. As excessive traction can lead to foot fixation and injury risk, there has been significant research surrounding the influence of stud configuration on the level of traction generated. This previous research, however, has predominately focused on the stud patterns, foot morphology and lower limb loading patterns of male players. As the popularity of women's soccer increases, the aim of this investigation was to examine the differences in translational traction of female-specific soccer boots and male soccer boots currently available. A custom-built apparatus was used to determine the translational traction on both natural and artificial grass for four different movement directions. It was hypothesised that the female-specific boot in each pair would produce lower levels of translational traction as they are designed to be safer for female players who are more at risk of lower limb fixation injuries compared to males. An independent samples T-test showed that while there were some loading conditions where female boots produced lower translational traction compared to male boots, across all loading scenarios there was no significant difference between male and female boots ( p  = 0.818), thus the null hypothesis was rejected.

Background: Soccer boots are produced with different stud patterns and configurations to provide players with extra traction on specific surface types to minimize slipping and improve player performance. Excessive traction, however, can lead to foot fixation injuries, particularly anterior cruciate ligament tears., Purpose/hypothesis: The purpose of this study was to explore the translational traction properties of 5 different outsole configurations moving in 4 different directions across both natural grass and artificial grass (AG) playing surfaces. It was hypothesized that longer studs or studs with an asymmetric shape would yield a higher traction coefficient compared with the recommended stud configuration for the given playing surface., Study Design: Descriptive laboratory study., Methods: A custom-built testing apparatus recorded the translational traction of 5 different soccer boots moving in an anterior, posterior, medial, or lateral direction on both natural grass and AG playing surfaces. A 3-way analysis of variance was performed to determine the effect of outsole configuration on the traction, and a post hoc Tukey analysis was performed to compare different outsole configurations with a control., Results: For the natural grass playing surface, the longer and asymmetric studs yielded a significantly higher ( P < .05) traction coefficient on 75% of loading scenarios, while on AG, they yielded a significantly higher traction on 50% of loading scenarios., Conclusion: Some soccer boots yielded higher traction values compared with the recommended configuration., Clinical Relevance: The results highlight the importance of boot selection on different playing surfaces. Higher traction values could increase the injury risk for players due to excessive traction and foot fixation., Competing Interests: One or more of the authors has declared the following potential conflict of interest or source of funding: Research funding was received from The University of Adelaide. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto. Ethical approval was not sought for the present study., (© The Author(s) 2024.)

Understanding of human neck stiffness and range of motion (ROM) with minimal neck muscle activation ("passive") is important for clinical and bioengineering applications. The aim of this study was to develop, implement, and evaluate the reliability of methods for assessing passive-lying stiffness and ROM, in six head-neck rotation directions. Six participants completed two assessment sessions. To perform passive-lying tests, the participant's head and torso were strapped to a bending (flexion, extension, lateral bending) or a rotation (axial rotation) apparatus, and clinical bed, respectively. The head and neck were manually rotated by the researcher to the participant's maximum ROM, to assess passive-lying stiffness. Participant-initiated ("active") head ROM was also assessed in the apparatus, and seated. Various measures of apparatus functionality were assessed. ROM was similar for all assessment configurations in each motion direction except flexion. In each direction, passive stiffness generally increased throughout neck rotation. Within-session reliability for stiffness (ICC > 0.656) and ROM (ICC > 0.872) was acceptable, but between-session reliability was low for some motion directions, probably due to intrinsic participant factors, participant-apparatus interaction, and the relatively low participant number. Moment-angle corridors from both assessment sessions were similar, suggesting that with greater sample size, these methods may be suitable for estimating population-level corridors., (© 2024. The Author(s).)

Well characterised mechanical response of the normal head-neck complex during passive motion is important to inform and verify physical surrogate and computational models of the human neck, and to inform normal baseline for clinical assessments. For 10 male and 10 female participants aged 20 to 29, the range of motion (ROM) of the neck about three anatomical axes was evaluated in active-seated, passive-lying and active-lying configurations, and the neck stiffness was evaluated in passive-lying. Electromyographic signals from the agonist muscles, normalised to maximum voluntary contractions, were used to provide feedback during passive motions. The effect of sex and configuration on ROM, and the effect of sex on linear estimates of stiffness in three regions of the moment-angle curve, were assessed with linear mixed models and generalised linear models. There were no differences in male and female ROM across all motion directions and configurations. Flexion and axial rotation ROM were configuration dependent. The passive-lying moment-angle relationship was typically non-linear, with higher stiffness (slope) closer to end of ROM. When normalising the passive moment-angle curve to active lying ROM, passive stiffness was sex dependent only for lateral bending region 1 and 2. Aggregate moment-angle corridors were similar for males and females in flexion and extension, but exhibited a higher degree of variation in applied moment for males in lateral bending and axial rotation. These data provide the passive response of the neck to low rate bending and axial rotation angular displacement, which may be useful for computational and surrogate modelling of the human neck., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 The Authors. Published by Elsevier Ltd.. All rights reserved.)

Walking gait data acquired with force platforms may be used for person re-identification (re-ID) in various authentication, surveillance, and forensics applications. Current force platform-based re-ID systems classify a fixed set of identities (IDs), which presents a problem when IDs are added or removed from the database. We formulated force platform-based re-ID as a deep metric learning (DML) task, whereby a deep neural network learns a feature representation that can be compared between inputs using a distance metric. The force platform dataset used in this study is one of the largest and the most comprehensive of its kind, containing 193 IDs with significant variations in clothing, footwear, walking speed, and time between trials. Several DML model architectures were evaluated in a challenging setting where none of the IDs were seen during training (i.e., zero-shot re-ID) and there was only one prior sample per ID to compare with each query sample. The best architecture was 85% accurate in this setting, though an analysis of changes in walking speed and footwear between measurement instances revealed that accuracy was 28% higher on same-speed, same-footwear comparisons, compared to cross-speed, cross-footwear comparisons. These results demonstrate the potential of DML algorithms for zero-shot re-ID using force platform data, and highlight challenging cases.

Prediction of propulsion kinematics and performance in wheelchair sports has the potential to improve capabilities of individual wheelchair prescription while minimizing testing requirements. While propulsion predictions have been developed for daily propulsion, these have not been extended for maximal effort in wheelchair sports. A two step-approach to predicting the effects of changing set-up in wheelchair rugby was developed, consisting of: (One) predicting propulsion kinematics during a 5 m sprint by adapting an existing linkage model; and (Two) applying partial least-squares regression to wheelchair set-up, propulsion kinematics, and performance. Eight elite wheelchair rugby players completed 5 m sprints in nine wheelchair set-ups while varying seat height, seat depth, seat angle, and tire pressure. Propulsion kinematics (contact and release angles) and performance (sprint time) were measured during each sprint and used for training and assessment for both models. Results were assessed through comparison of predicted and experimental propulsion kinematics (degree differences) for Step One and performance times (seconds differences) for Step Two. Kinematic measures, in particular contact angles, were identified with mean prediction errors less than 5 degrees for 43 of 48 predictions. Performance predictions were found to reflect on-court trends for some players, while others showed weaker prediction accuracy. More detailed modeling approaches that can account for individual athlete activity limitations would likely result in improved accuracy in propulsion and performance predictions across a range of wheelchair sports. Although this would come at an increased cost, developments would provide opportunities for more suitable set-ups earlier in an athlete's career, increasing performance and reducing injury risk., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 Haydon, Pinder, Grimshaw, Robertson and Holdback.)

Prescription of wheelchair rugby chairs is difficult due to the range of athlete impairment types and severities in the sport, difficulty in adjusting wheelchair settings, and assessing on-court performance. Currently, elite players rely on experiential knowledge (personal, coaches, and support staff) to select an appropriate set-up. Technological advancements, such as with inertial measurement units and processing algorithms, and representative testing approaches, has improved the potential for assessing set-ups at an individual level. An orthogonal design approach was implemented using an adjustable wheelchair to investigate the effect of seat height, seat depth, seat angle, and tyre pressure on performance, mobility, and propulsion kinematics. Six elite wheelchair rugby players completed testing in nine individually tailored wheelchair set-ups while monitoring both quantitative and qualitative measures of performance. From this testing, a recommended set-up was compared with the current set-up for each individual. A single case-study approach shows how the assessment method identifies parameter settings that can potentially improve performance. Three of six players reported a blind preference for the recommended set-up over the current set-up, whilst remaining players often displayed similar performance between their current and recommended set-ups. This approach can improve upon the current prescription process for rugby wheelchairs.

Due to the detrimental influence of unnecessary mass on performance, racing wheelchair instrumentation used in both competition assessment and research is currently limited. Attaining key kinetic parameters of propulsion can enhance technique and provide athletes with a competitive advantage. This research examined the plausibility of inertial measurement units (IMUs) to estimate propulsion forces, during a simulated wheelchair race start and training. Start propulsion data calculated from an IMU system was compared to reference force plate data; steady state motion data was compared with existing literature. Some agreement in kinetic parameters between IMU data was observed under steady state motion, with data from athletes following a linear force-velocity relationship. In this context, it is important to identify that this cannot be directly compared to the existing literature due to the different methods of force measurement and the lack of data for similar force measurements using IMUs. IMUs were ineffective when used with wheelchairs having spoked wheels. Performance was best for measurements in the direction of motion. Although exact agreement was not observed, the IMU can provide an effective tool in the in-field assessment of propulsion kinetics.

Purpose: To evaluate the effect of the braced arm-to-thigh technique (BATT) (versus self-selected techniques) on three-dimensional trunk kinematics and spinal loads for three common activities of daily living (ADLs) simulated in the laboratory: weeding (gardening), reaching for an object in a low cupboard, and car egress using the two-legs out technique., Methods: Ten young healthy males performed each task using a self-selected technique, and then using the BATT. The pulling action of weeding was simulated using a magnet placed on a steel plate. Cupboard and car egress tasks were simulated using custom apparatus representing the dimensions of a kitchen cabinet and a medium-sized Australian car, respectively. Three-dimensional trunk kinematics and L4/L5 spinal loads were estimated using the Lifting Full-Body OpenSim model and compared between techniques. Paired t-tests were used to compare peak values between methods (self-selected vs BATT)., Results: The BATT significantly reduced peak extension moments (13-51%), and both compression (27-45%) and shear forces (31-62%) at L4/L5, compared to self-selected techniques for all three tasks (p < 0.05). Lateral bending angles increased with the BATT for weeding and cupboard tasks, but these changes were expected as the BATT inherently introduces asymmetric trunk motion., Conclusion: The BATT substantially reduced L4/L5 extension moments, and L4/L5 compression and shear forces, compared to self-selected methods, for three ADLs, in a small cohort of ten young healthy males without prior history of back pain. These study findings can be used to inform safe procedures for these three ADLs, as the results are considered representative of a mature population.

Inertial measurement units (IMUs) provide a practical solution for attaining key performance data for wheelchair sports. The effects of IMU placement position on the identification of propulsion characteristics are unknown. The aim of this study was to determine the variability in the reliability of cycle time measurements (time between hand contacts) across IMU locations on the chair frame (axle housings), and wheels (axle, push rim, outer rim), on both the left and right sides ( n  = 8). Contacts were defined by spikes in the resultant acceleration data, corresponding to impact between the hands and push rim, and verified against motion capture. Five elite wheelchair racing athletes propelled at racing speeds on a treadmill. Excellent inter-rater Intraclass Correlation Coefficient values indicated high reliability and repeatability for both motion capture and IMU signal analysis approaches ( R  = 0.997, p  < 0.001 and R  = 0.990, p  < 0.001, respectively). The best results were (as determined by the best between method agreement) were observed for IMUs located on the frame. Detection reliability was positively associated with signal-to-noise ratio of the acceleration data. The IMU assessment approach facilitates an automated processing capability, which is an improvement to the currently used video analysis.

Despite the common use of one-handed lifting techniques for activities of daily living, these techniques have received little attention in the biomechanics literature. The braced arm-to-thigh technique (BATT) is a one-handed lifting method in which the dominant hand picks up objects, while the free hand braces the trunk on the ipsilateral thigh. The aim of this study was to compare the BATT to two-handed or unsupported one-handed lifting techniques with loads of 2 and 10 kg, by evaluating trunk motion and spine loading at L4/L5. Twenty healthy participants (30-70 years old) matched in age and sex to 18 participants with low back pain were recruited to the study. A three-axis load cell secured to the distal anterior thigh measured the bracing forces applied by the hand. The OpenSim Lifting Full-Body model was used to estimate trunk kinematics and spinal loading at L4/L5. Linear mixed-effects models were developed to compare trunk angles and L4/L5 moments and forces between lifting techniques. Trunk flexion angles were significantly reduced for the BATT lift compared to one-handed and two-handed stoop lifts (9-20%). However, the BATT also increased asymmetric trunk kinematics and moments at L4/L5. The BATT produced significantly lower moments (28-38%), and compressive (25-32%) and antero-posterior shear (25-45%) forces at L4/L5, compared to unsupported lifting techniques. Bracing the hand on the thigh to support the trunk can substantially reduce low back loading during lifting tasks of 2 to 10 kg., Competing Interests: Declaration of Competing Interest No competing financial interests exist., (Copyright © 2019 Elsevier Ltd. All rights reserved.)

The anthropometries of elite wheelchair racing athletes differ from the generic, able-bodied anthropometries commonly used in computational biomechanical simulations. The impact of using able-bodied parameters on the accuracy of simulations involving wheelchair racing is currently unknown. In this study, athlete-specific mass segment inertial parameters of the head and neck, torso, upper arm, forearm, hand, thigh, shank, and feet for five elite wheelchair athletes were calculated using dual-energy X-ray absorptiometry (DXA) scans. These were compared against commonly used anthropometrics parameters of data presented in the literature. A computational biomechanical simulation of wheelchair propulsion using the upper extremity dynamic model in opensim assessed the sensitivity of athlete-specific mass parameters using Kruskal-Wallis analysis and Spearman correlations. Substantial between-athlete body mass distribution variances (thigh mass between 7.8% and 22.4% total body mass) and between-limb asymmetries (<62.4% segment mass; 3.1 kg) were observed. Compared to nonathletic able-bodied anthropometric data, wheelchair racing athletes demonstrated greater mass in the upper extremities (up to 3.8% total body mass) and less in the lower extremities (up to 9.8% total body mass). Computational simulations were sensitive to individual body mass distribution, with joint torques increasing by up to 31.5% when the scaling of segment masses (measured or generic) differed by up to 2.3% total body mass. These data suggest that nonathletic, able-bodied mass segment inertial parameters are inappropriate for analyzing elite wheelchair racing motion., (Copyright © 2019 by ASME.)

There is currently no validated full-body lifting model publicly available on the OpenSim modelling platform to estimate spinal loads during lifting. In this study, the existing full-body-lumbar-spine model was adapted and validated for lifting motions to produce the lifting full-body model. Back muscle activations predicted by the model closely matched the measured erector spinae activation patterns. Model estimates of intradiscal pressures and in vivo measurements were strongly correlated. The same spine loading trends were observed for model estimates and reported vertebral body implant measurements. These results demonstrate the suitability of this model to evaluate changes in lumbar loading during lifting.

Objectives: Use a task vehicle of sprint testing in wheelchair rugby (WCR) to explore the impact of small changes to test design using both group and individual analysis., Design: Exploratory, repeated measures, on-court study METHOD: 25 national or international level wheelchair rugby players completed 5×5m sprints under two conditions: (i) an acceleration from standstill in their own time, and (ii) an 'active' start, simulating a key aspect of performance. Video analysis and accelerometer data were used to measure key kinematic and performance variables with a focus on the first three strokes. Each player was grouped into a high-, mid-, or low-point group based on their sport-specific classification score. Group (paired sample t-tests) and individual (meaningful differences, performance coefficients, and Cohen's d effect sizes) analysis assessed differences between the two conditions., Results: The low-point classification group performed significantly slower in the active start (p<0.05). There were no differences in sprint time for the high- and mid-point groups. Mid-point players achieved greater peak accelerations for strokes two and three in the active start (p<0.05). Individual sprint performances varied substantially, ranging from 8% decrease to 14% increase in sprint time for the active start. Meaningful differences in peak accelerations were demonstrated for 23 out of the 25 players., Conclusions: Small amendments to test design can lead to significant differences in individual athlete performance. Traditional group analyses masked important individual responses to testing conditions. There is need to further consider representative test design, and individual analysis for monitoring physical and skill performance., (Copyright © 2018 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.)

Purpose: Maximal acceleration from standstill has been identified as a key performance indicator in wheelchair rugby; however, the impact of classification and kinematic variables on performance has received limited attention. This study aimed to investigate kinematic variables during maximal acceleration, with level of activity limitation accounted for using sport-classification scores., Methods: Based on their sporting classification scores, which reflect combined trunk, arm, and hand function, 25 elite wheelchair rugby players were analyzed in high-, mid-, and low-point groups before completing five 5-m sprints from a stationary position. Inertial measurement units and video analysis were used to monitor key kinematic variables., Results: Significant differences in kinematic variables were evident across the classification groups, particularly for the first stroke-contact angle (1-way ANOVA F 2,122  = 51.5, P < .05) and first stroke time (F 2,124  = 18.3, P < .05). High-point players used a first stroke-contact angle that was closer to top dead center of the wheel than either other group, while also using a shorter overall stroke time than low-point players. A linear mixed-effects model was used to investigate how kinematic variables influenced performance, with results suggesting that increased release angles (ie, farther around the wheel) and decreased stroke angles resulted in larger peak accelerations. Further investigation revealed that these results are likely influenced by strong relationships for the high-point group, as there was often no clear trend evident for midpoint and low-point groups., Conclusion: Findings show that various propulsion approaches exist across classification groups, with this information potentially informing individual wheelchair setups and training programs.