The back squat (BS) is one of the most effective exercises for enhancing lower limb strength, but an unstable squat can increase shear forces in the lower back. Understanding how to assess the squat stability is useful for avoiding potential sports injury. During the BS, the trunk lean and center of pressure (COP) are relevant to squat safety, and these kinematics can be altered by elevating the heel. However, there is no relevant meta-analysis on the impact of different heel heights on squat stability. This study aims to bridge the gap in the literature by conducting a systematic review and network meta-analysis on how heel elevation affects squat stability. By quantifying the influence of different heel heights on key biomechanical parameters, such as the center of pressure deviation and ankle dorsiflexion, the study provides actionable insights for athletes, trainers, and clinicians. Fourteen articles were included, and the majority of these studies demonstrated that elevated heels (EHs) can reduce COP deviation and trunk lean. In addition, 25 mm heels may be the preferred option for squat stability in the AP direction when COP data and network meta-analysis are combined. However, in the ML direction, the capacity to maintain balance is rather questionable; when ankle peak dorsiflexion is combined, 8 mm heels have higher COP deviation values and 5 mm heels have lower COP deviation values. Regarding limitations, reliance on a single bias assessment tool (Cochrane Risk of Bias Tool) might not fully capture methodological variability across non-RCT studies. Future systematic reviews could consider using multiple bias assessment tools for robust assessment.

As one of many fundamental sports techniques, the landing maneuver is also frequently used in clinical injury screening and diagnosis. However, the landing patterns are different under different constraints, which will cause great difficulties for clinical experts in clinical diagnosis. Machine learning (ML) have been very successful in solving a variety of clinical diagnosis tasks, but they all have the disadvantage of being black boxes and rarely provide and explain useful information about the reasons for making a particular decision. The current work validates the feasibility of applying an explainable ML (XML) model constructed by Layer-wise Relevance Propagation (LRP) for landing pattern recognition in clinical biomechanics. This study collected 560 groups landing data. By incorporating these landing data into the XML model as input signals, the prediction results were interpreted based on the relevance score (RS) derived from LRP. The interpretation obtained from XML was evaluated comprehensively from the statistical perspective based on Statistical Parametric Mapping (SPM) and Effect Size. The RS has excellent statistical characteristics in the interpretation of landing patterns between classes, and also conforms to the clinical characteristics of landing pattern recognition. The current work highlights the applicability of XML methods that can not only satisfy the traditional decision problem between classes, but also largely solve the lack of transparency in landing pattern recognition. We provide a feasible framework for realizing interpretability of ML decision results in landing analysis, providing a methodological reference and solid foundation for future clinical diagnosis and biomechanical analysis.

Background: Finding the best subset of gait features among biomechanical variables is considered very important because of its ability to identify relevant sports and clinical gait pattern differences to be explored under specific study conditions. This study proposes a new method of metaheuristic optimization-based selection of optimal gait features, and then investigates how much contribution the selected gait features can achieve in gait pattern recognition. Methods: Firstly, 800 group gait datasets performed feature extraction to initially eliminate redundant variables. Then, the metaheuristic optimization algorithm model was performed to select the optimal gait feature, and four classification algorithm models were used to recognize the selected gait feature. Meanwhile, the accuracy results were compared with two widely used feature selection methods and previous studies to verify the validity of the new method. Finally, the final selected features were used to reconstruct the data waveform to interpret the biomechanical meaning of the gait feature. Results: The new method finalized 10 optimal gait features (6 ankle-related and 4-related knee features) based on the extracted 36 gait features (85 % variable explanation) by feature extraction. The accuracy in gait pattern recognition among the optimal gait features selected by the new method (99.81 % ± 0.53 %) was significantly higher than that of the feature-based sorting of effect size (94.69 % ± 2.68 %), the sequential forward selection (95.59 % ± 2.38 %), and the results of previous study. The interval between reconstructed waveform-high and reconstructed waveform-low curves based on the selected feature was larger during the whole stance phase. Significance: The selected gait feature based on the proposed new method (metaheuristic optimization-based selection) has a great contribution to gait pattern recognition. Sports and clinical gait pattern recognition can benefit from population-based metaheuristic optimization techniques. The metaheuristic optimization algorithms are expected to provide a practical and elegant solution for sports and clinical biomechanical feature selection with better economy and accuracy.

This study aimed to examine the effects of wearable resistance (WR) training on change of direction ability (CODA), muscle activation patterns, and knee joint stress in athletes. Fifteen healthy male football players participated in a pre- and post-training intervention designed to target the quadriceps, hamstrings, and calf muscles to improve neuromuscular control and joint stability. Surface electromyography (EMG) was used to assess muscle activation, and finite element analysis (FEA) was applied to evaluate stress distribution in the knee joint. Following the WR training program, there was a significant reduction in knee abduction angle during the stance phase (p = 0.001), indicating enhanced joint stability. Strength in the calf muscles increased significantly, while muscle activation levels in the quadriceps (p < 0.001) and hamstrings (p = 0.007) were also elevated. Enhanced co-activation between quadriceps and hamstrings was observed, and FEA demonstrated a significant decrease in the maximal von Mises stress in the anterior cruciate ligament (ACL) and meniscus. These findings suggest that WR training improves CODA and lower limb muscle coordination while reducing internal knee joint stress, potentially lowering the risk of ACL injuries and enhancing athletic performance.

Background: Patients undergoing anterior cruciate ligament reconstruction (ACLR) are at high risk of osteoarthritis or secondary injuries, with abnormal knee contact forces (KCFs) identified as a key factor in joint degeneration. Traditional KCF assessment relies on expensive lab systems while advances in computer vision and AI now enable low-cost alternatives. However, currently available methods oversimplify knee mechanics and neglect compensatory movements, highlighting the urgent need for intelligent, real-time monitoring tools for personalized rehabilitation. Therefore, the aim of this study was to develop and validate an integrated, non-invasive framework for accurate KCFs prediction in ACLR patients during daily activities. We hypothesized that combining enhanced musculoskeletal modeling with a deep learning architecture incorporating spatiotemporal attention would improve the prediction accuracy across multiple movement tasks. Methods: This study simultaneously recorded three daily movements of 29 post-ACLR patients using both Vicon and OpenCap. Motion trajectories captured by Vicon were imported into OpenSim for musculoskeletal modeling and KCFs calculation. Dataset comprising OpenCap-derived kinematics and OpenSim-computed KCFs was used to train 3 learning models for the prediction of KCFs in ACLR patients across different movements. Results: Among three models, CNN-BiGRU-Attention model demonstrated the best predictive performance across all three movement tasks (R2walking = 0.973 ± 0.003, R2running = 0.982 ± 0.004, R2descending stairs = 0.951 ± 0.007). CNN and self-attention mechanism collectively enhanced the model's ability to capture key features in ACLR patients' movement data, thereby improving KCF prediction accuracy. Furthermore, for the three daily activities, all models showed superior KCFs prediction performance in running and stair-descent tasks compared to walking. Conclusion: The developed framework successfully achieved high-precision prediction of KCFs. This technological breakthrough not only provides a real-time quantitative tool for rehabilitation monitoring in patients with ACLR, but also facilitates a paradigm shift from static laboratory analysis to dynamic real-time monitoring, with broad application prospects in sports medicine, rehabilitation engineering.