Search

Background: Low back pain (LBP) is a major global disability contributor with profound health and socio-economic implications. The predominant form is non-specific LBP (NSLBP), lacking treatable pathology. Active physical interventions tailored to individual needs and capabilities are crucial for its management. However, the intricate nature of NSLBP and complexity of clinical classification systems necessitating extensive clinical training, hinder customised treatment access. Recent advancements in machine learning and computer vision demonstrate promise in characterising NSLBP altered movement patters through wearable sensors and optical motion capture. This study aimed to develop and evaluate a machine learning model (i.e., 'BACK-to-MOVE') for NSLBP classification trained with expert clinical classification, spinal motion data from a standard video alongside patient-reported outcome measures (PROMs)., Methods: Synchronised video and three-dimensional (3D) motion data was collected during forward spinal flexion from 83 NSLBP patients. Two physiotherapists independently classified them as motor control impairment (MCI) or movement impairment (MI), with conflicts resolved by a third expert. The Convolutional Neural Networks (CNNs) architecture, HigherHRNet, was chosen for effective pose estimation from video data. The model was validated against 3D motion data (subset of 62) and trained on the freely available MS-COCO dataset for feature extraction. The Back-to-Move classifier underwent fine-tuning through feed-forward neural networks using labelled examples from the training dataset. Evaluation utilised 5-fold cross-validation to assess accuracy, specificity, sensitivity, and F1 measure., Results: Pose estimation's Mean Square Error of 0.35 degrees against 3D motion data demonstrated strong criterion validity. Back-to-Move proficiently differentiated MI and MCI classes, yielding 93.98% accuracy, 96.49% sensitivity (MI detection), 88.46% specificity (MCI detection), and an F1 measure of .957. Incorporating PROMs curtailed classifier performance (accuracy: 68.67%, sensitivity: 91.23%, specificity: 18.52%, F1: .800)., Conclusion: This study is the first to demonstrate automated clinical classification of NSLBP using computer vision and machine learning with standard video data, achieving accuracy comparable to expert consensus. Automated classification of NSLBP based on altered movement patters video-recorded during routine clinical examination could expedite personalised NSLBP rehabilitation management, circumventing existing healthcare constraints. This advancement holds significant promise for patients and healthcare services alike., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2024 Hartley et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Background: Exposure to prolonged periods in microgravity is associated with deconditioning of the musculoskeletal system due to chronic changes in mechanical stimulation. Given astronauts will operate on the Lunar surface for extended periods of time, it is critical to quantify both external (e.g., ground reaction forces) and internal (e.g., joint reaction forces) loads of relevant movements performed during Lunar missions. Such knowledge is key to predict musculoskeletal deconditioning and determine appropriate exercise countermeasures associated with extended exposure to hypogravity., Objectives: The aim of this paper is to define an experimental protocol and methodology suitable to estimate in high-fidelity hypogravity conditions the lower limb internal joint reaction forces. State-of-the-art movement kinetics, kinematics, muscle activation and muscle-tendon unit behaviour during locomotor and plyometric movements will be collected and used as inputs (Objective 1), with musculoskeletal modelling and an optimisation framework used to estimate lower limb internal joint loading (Objective 2)., Methods: Twenty-six healthy participants will be recruited for this cross-sectional study. Participants will walk, skip and run, at speeds ranging between 0.56-3.6 m/s, and perform plyometric movement trials at each gravity level (1, 0.7, 0.5, 0.38, 0.27 and 0.16g) in a randomized order. Through the collection of state-of-the-art kinetics, kinematics, muscle activation and muscle-tendon behaviour, a musculoskeletal modelling framework will be used to estimate lower limb joint reaction forces via tracking simulations., Conclusion: The results of this study will provide first estimations of internal musculoskeletal loads associated with human movement performed in a range of hypogravity levels. Thus, our unique data will be a key step towards modelling the musculoskeletal deconditioning associated with long term habitation on the Lunar surface, and thereby aiding the design of Lunar exercise countermeasures and mitigation strategies., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2022 Herssens et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

The accurate detection of foot-strike and toe-off is often critical in the assessment of running biomechanics. The gold standard method for step event detection requires force data which are not always available. Although kinematics-based algorithms can also be used, their accuracy and generalisability are limited, often requiring corrections for speed or foot-strike pattern. The purpose of this study was to develop FootNet, a novel kinematics and deep learning-based algorithm for the detection of step events in treadmill running. Five treadmill running datasets were gathered and processed to obtain segment and joint kinematics, and to identify the contact phase within each gait cycle using force data. The proposed algorithm is based on a long short-term memory recurrent neural network and takes the distal tibia anteroposterior velocity, ankle dorsiflexion/plantar flexion angle and the anteroposterior and vertical velocities of the foot centre of mass as input features to predict the contact phase within a given gait cycle. The chosen model architecture underwent 5-fold cross-validation and the final model was tested in a subset of participants from each dataset (30%). Non-parametric Bland-Altman analyses (bias and [95% limits of agreement]) and root mean squared error (RMSE) were used to compare FootNet against the force data step event detection method. The association between detection errors and running speed, foot-strike angle and incline were also investigated. FootNet outperformed previously published algorithms (foot-strike bias = 0 [-10, 7] ms, RMSE = 5 ms; toe-off bias = 0 [-10, 10] ms, RMSE = 6 ms; and contact time bias = 0 [-15, 15] ms, RMSE = 8 ms) and proved robust to different running speeds, foot-strike angles and inclines. We have made FootNet's source code publicly available for step event detection in treadmill running when force data are not available., Competing Interests: The authors have declared that no competing interests exist.

Head collisions in sport can result in catastrophic injuries to the cervical spine. Musculoskeletal modelling can help analyse the relationship between motion, external forces and internal loads that lead to injury. However, impact specific musculoskeletal models are lacking as current viscoelastic values used to describe cervical spine joint dynamics have been obtained from unrepresentative quasi-static or static experiments. The aim of this study was to develop and validate a cervical spine musculoskeletal model for use in axial impacts. Cervical spine specimens (C2-C6) were tested under measured sub-catastrophic loads and the resulting 3D motion of the vertebrae was measured. Specimen specific musculoskeletal models were then created and used to estimate the axial and shear viscoelastic (stiffness and damping) properties of the joints through an optimisation algorithm that minimised tracking errors between measured and simulated kinematics. A five-fold cross validation and a Monte Carlo sensitivity analysis were conducted to assess the performance of the newly estimated parameters. The impact-specific parameters were integrated in a population specific musculoskeletal model and used to assess cervical spine loads measured from Rugby union impacts compared to available models. Results of the optimisation showed a larger increase of axial joint stiffness compared to axial damping and shear viscoelastic parameters for all models. The sensitivity analysis revealed that lower values of axial stiffness and shear damping reduced the models performance considerably compared to other degrees of freedom. The impact-specific parameters integrated in the population specific model estimated more appropriate joint displacements for axial head impacts compared to available models and are therefore more suited for injury mechanism analysis., Competing Interests: The authors have declared that no competing interests exist.

Cervical spine trauma from sport or traffic collisions can have devastating consequences for individuals and a high societal cost. The precise mechanisms of such injuries are still unknown as investigation is hampered by the difficulty in experimentally replicating the conditions under which these injuries occur. We harness the benefits of computer simulation to report on the creation and validation of i) a generic musculoskeletal model (MASI) for the analyses of cervical spine loading in healthy subjects, and ii) a population-specific version of the model (Rugby Model), for investigating cervical spine injury mechanisms during rugby activities. The musculoskeletal models were created in OpenSim, and validated against in vivo data of a healthy subject and a rugby player performing neck and upper limb movements. The novel aspects of the Rugby Model comprise i) population-specific inertial properties and muscle parameters representing rugby forward players, and ii) a custom scapula-clavicular joint that allows the application of multiple external loads. We confirm the utility of the developed generic and population-specific models via verification steps and validation of kinematics, joint moments and neuromuscular activations during rugby scrummaging and neck functional movements, which achieve results comparable with in vivo and in vitro data. The Rugby Model was validated and used for the first time to provide insight into anatomical loading and cervical spine injury mechanisms related to rugby, whilst the MASI introduces a new computational tool to allow investigation of spinal injuries arising from other sporting activities, transport, and ergonomic applications. The models used in this study are freely available at simtk.org and allow to integrate in silico analyses with experimental approaches in injury prevention., Competing Interests: The authors have declared that no competing interests exist.