Biomechanics

Background: This technical note proposes a deep learning-based, few-shot automatic key point tracking technique tailored to sports biomechanics research. Methods: The present method facilitates the arbitrary definition of key points on athletes’ bodies or sports equipment. Initially, a limited number of video frames are manually digitized to mark the points of interest. These annotated frames are subsequently used to train a deep learning model that leverages a pre-trained VGG16 network as its backbone and incorporates an additional convolutional head. Feature maps extracted from three intermediate layers of VGG16 are processed by the head network to generate a probability map, highlighting the most likely locations of the key points. Transfer learning is implemented by freezing the backbone weights and training only the head network. By restricting the training data generation to regions surrounding the manually annotated points and training specifically for each video, this approach minimizes training time while maintaining high precision. Conclusions: This technique substantially reduces the time and effort required compared to frame-by-frame manual digitization in various sports settings, and enables customized training tailored to specific analytical needs and video environments. Full article

Background: Gait assessment is a complex task involving locomotion and balance control across all body segments, requiring a global analysis in the event of motor disorders. Among these are spinal disorders, where an understanding of spinal kinematics during walking is important to improve treatment decisions and outcomes. The technique of stereophotogrammetric motion analysis is currently the gold standard in this context. A new integrated protocol for whole-body kinematic gait analysis is proposed in this study, which takes into account the movements of the spine. Methods: A new protocol with 30 passive markers was developed to analyze gait. Of these markers, 22 implemented the Davis protocol for gait measurement, while the other 8 were placed onto the spine to record spinal movements. The protocol’s accuracy was assessed through comparisons of the constructive angles of a manikin replicating the human body and the angles measured with the optoelectronic system. An assessment of intra- and inter-operator repeatability and protocol usability was carried out by recruiting and applying the protocol in a population composed of ten subjects (mean age 17.36—SD 10.12) without any history of spine pathology. Results: The protocol was validated successfully. The validation accuracy was more than satisfactory: the measured RMSE was 1.2 ± 1° for the data collected with the optoelectronic system with respect to the manikin. The intra-operator repeatability was also good in the sagittal and frontal planes (average ICC > 0.867), and the inter-operator repeatability was moderate or good in all planes (average ICC > 0.77). The usability score obtained using the System Usability Scale was satisfactory (mean 74.75, SD 5.88). Conclusions: This study proposes a new protocol to assess total body kinematics, including the spine in its three main segments, during gait. The successful validation of this protocol in terms of reliability and usability allows for its subsequent clinical application. Full article

Background/Objectives: Groin and hip injuries are common in sport, and muscle weakness has been identified as an intrinsic risk factor. So, analyzing the strength of the hip musculature becomes important. To date, there are no hip adductor isometric strength tests on force platforms. This study aims to analyze the intra-test reliability of a hip adductor strength test using force platforms. Methods: The study sample comprised 13 male professional soccer players with an average age of 22.3 ± 3 years, body mass of 75.8 ± 5.4 kg, and height of 1.8 ± 0.1 m. Assessments were conducted on a uniaxial force platform. The variables analyzed are peak force (PF), rate of force development (RFD), and impulse. Intra-test reliability was evaluated using the coefficient of variation (CV), intraclass correlation coefficient (ICC), and Bland–Altman plots. Results: Acceptable levels of reliability were identified solely for the variable of peak force, with CV values of D = 5.7% for the dominant profile and ND = 5.4% for the non-dominant profile. Furthermore, moderate and good relative reliability were observed in peak force for the dominant (ICC = 0.706) and non-dominant (ICC = 0.819) profiles, respectively. However, the remaining time-related variables, RFD and impulse, did not achieve acceptable levels of absolute reliability (CV > 10%) and displayed poor to moderate relative reliability. Conclusions: In summary, PF during the hip adductor isometric strength test demonstrated acceptable absolute and commendable relative reliability. Conversely, the time-related variables, specifically RFD and impulse, yielded unsatisfactory absolute and relative reliability levels. Full article

Background/Objectives: Currently, there is no strong evidence to support interventions for medial tibial stress syndrome (MTSS), a common running injury associated with tibial loading. Vertical ground reaction force (vGRF) and axial tibial acceleration (TA) are the most common methods of estimating tibial loads, yet clinical recommendations for technique modification to reduce these metrics are not well documented. This study investigated whether changes to speed, cadence, stride length, and foot-strike pattern influence vGRF and TA. Additionally, machine-learning models were evaluated for their ability to estimate vGRF metrics. Methods: Sixteen runners completed seven 1 min trials consisting of preferred technique, ±10% speed, ±10% cadence, forefoot, and rearfoot strike. Results: A 10% speed reduction decreased peak tibial acceleration (PTA), vertical average loading rate (VALR), vertical instantaneous loading rate (VILR), and vertical impulse by 13%, 10.9%, 9.3%, and 3.2%, respectively. A 10% cadence increase significantly reduced PTA (11.5%), VALR (15.6%), VILR (13.5%), and impulse (3.5%). Forefoot striking produced significantly lower PTA (26.6%), VALR (68.3%), and VILR (68.9%). Habitual forefoot strikers had lower VALR (58.1%) and VILR (47.6%) compared to rearfoot strikers. Machine-learning models predicted all four vGRF metrics with mean average errors of 9.5%, 10%, 10.9%, and 3.4%, respectively. Conclusions: This study demonstrates that small-scale modifications to running technique effectively reduce tibial load estimates. Machine-learning models offer an accessible, affordable tool for gait retraining by predicting vGRF metrics without reliance on IMU data. The findings support practical strategies for reducing MTSS risk. Full article

Background/Objectives: The manual tuning of exoskeleton control parameters is tedious and often ineffective for adapting to individual users. Human-in-the-loop (HIL) optimization offers an automated approach, but existing methods typically rely on metabolic cost, which requires prolonged data collection times of at least 60 s. Surface electromyography (EMG) signals, as an alternative, enable faster optimization with reduced data acquisition times. Methods: This study develops a rapid EMG-based HIL Bayesian optimization framework to tune hip exoskeleton controllers for assisting free leg swinging. Eight participants are asked to perform leg swinging at two frequencies with assistance from a hip exoskeleton. EMG signals from four sensors, representing muscle activity during forward and backward swings, are dynamically processed into cost functions. Bayesian optimization with Gaussian processes tunes four controller parameters using an expected improvement acquisition function. Optimization outcomes are validated against no device, zero torque, and general control baselines. Results: Optimization converges within an average of 142 s with a standard deviation of 24 s across all participants. The controller yields muscle activity reductions of 16.1% (p < 0.001) compared to no device, 21.7% (p < 0.001) versus zero torque, and 15.1% (p < 0.001) versus general control. EMG-based tuning is faster than metabolic-cost-based methods and perceived as less effortful, with Borg scale reductions of up to 39.5%. Conclusions: EMG-based HIL optimization significantly enhances controller tuning speed and effectiveness, demonstrating its potential for scalable and user-specific exoskeleton applications. Full article

Background/Objectives: Kinematic measurements obtained from functional tests have been used to identify associated and risk factors for the development of lower limb dysfunction, allowing targeted interventions to reduce potential risks and guide rehabilitation. It is necessary to identify variables and tests with adequate reliability and with the capability to discriminate individuals with and without lower limb functional deficits. This study aimed to determine which single-legged test (single-leg squat and single-leg landing) and variables (angle at deepest instant and range of motion) present the best reliability and capability to discriminate individuals with and without lower limb functional deficits. Methods: The frontal plane kinematics of 86 adults, divided into 2 groups (43 with lower limb functional deficits and 43 without), as classified by the Lower Extremity Functional Scale, were assessed during single-leg squat and single-leg landing tasks. The differences between groups in trunk, pelvis, hip, and knee ranges of motion and angles were tested using the independent T test or Mann–Whitney U test, and the test–retest, inter-rater, and intra-rater absolute (standard error of measurement and minimal detectable difference) and relative (intraclass correlation coefficient) reliability were calculated. Results: Trunk (r = 0.47), hip (r = 0.40), and knee (r = 0.35) angles at the deepest instant, as well as range of motion of the trunk (r = 0.33), pelvis (r = 0.47), and knee (r = 0.32) during the single-leg landing discriminated between groups (p < 0.05). For the single-leg squat, no variable discriminated the groups. Test–retest, inter-rater, and intra-rater reliability ranged from poor to excellent, with minimal detectable differences remaining below 19°. Conclusions: The single-leg landing and pelvis range of motion were the most effective tests and variables for discriminating individuals with and without lower limb functional deficits. Most variables demonstrated moderate test–retest and excellent inter-rater and intra-rater reliability. Full article

Background/Objectives: Postural control describes our ability to maintain an upright position. This study explored the impact of prolonged ankle plantar-flexed standing on postural control variability and strategy in an older adult population. The ability to perceive balance change was also assessed via subjective balance-related variables. Methods: Twenty-four community-dwelling older adults were recruited via convenience sampling. Each participant completed a balance confidence and falls efficacy questionnaire at baseline. Five barefoot quiet standing trials on a force plate then followed (Timepoint 1). After this, the participants stood with their ankles in a plantar-flexed position for up to 7.5 min before completing another quiet standing trial on the force plate. Four further ankle plantar-flexed standing trials of 2 min were then completed, interspersed with quiet standing trials on a force plate (Timepoint 2). The balance confidence and falls efficacy questionnaires were then completed again. For measures of postural control variability (sway path length, root mean square [RMS], sway area) and strategy (fractal dimension), mean values for the five trials were calculated for Timepoints 1 and 2 separately. Results: The sway path length and RMS measures were significantly increased (p < 0.05) at Timepoint 2. However, the fractal dimension did not change. There was also no change in balance confidence or falls efficacy. Conclusions: The findings suggest that prolonged standing can impact measures of postural variability without a change in postural control strategy. Postural control change also occurred without a change in subjective balance measures, suggesting that the altered balance may not be practically significant or perceptible to the individual. Full article

Background/Objectives: This study aimed to examine how uncertainties in inertial properties and minimal sets of inertial parameters (MSIP) affect inverse-dynamics simulations of high-acceleration sport movements and to demonstrate that applying MSIP identified through the free vibration measurement method improves simulation accuracy. Methods: Monte Carlo simulations were performed for running, side-cutting, vertical jumping, arm swings, and leg swings by introducing uncertainties in inertial properties and MSIP. Results: These uncertainties significantly affect the joint torques and ground reaction forces and moments (GRFs&Ms), especially during large angular acceleration. The mass and longitudinal position of the center of gravity had strong effects. Subsequently, MSIP identified by our methods with free vibration measurement were applied to the same tasks, improving the accuracy of the predicted ground reaction forces compared with the standard regression-based estimates. The root mean square error decreased by up to 148 N. Conclusions: These results highlight that uncertainties in inertial properties and MSIP affected the calculated joint torques and GRFs&Ms, and combining experimentally identified MSIP with dynamics simulations enhances precision. These findings demonstrate that utilizing the MSIP from free vibration measurement in inverse dynamics simulations improves the accuracy of dynamic models in sports biomechanics, thereby providing a robust framework for precise biomechanical analyses. Full article

Adductor strains are prevalent injuries in professional soccer. The purpose of this study is to identify further evidence of characteristics associated with adductor injury. MLS and other worldwide leagues have differing styles of play warranting further investigation of injury mechanisms. A descriptive cohort study was conducted with a single professional team in the MLS. Injury data was collected between the 2016 to 2022 seasons to characterize adductor injury. Player position type, age, previous injury, and mechanism(s) of injury (MOI) were assessed to understand the injured population. Generalized estimating equations (GEEs) were utilized to assess the odds of future injury among the injured and non-injured populations. Adductor strains (n = 30) made up 15.5% of all soft-tissue, lower extremity injuries (n = 194) in a single MLS cohort. These injuries were the second most common defined soft-tissue, non-contact injury after hamstring strains (26.4%) and before quadricep strains (11.9%). Among the position types, 28% of defenders, 25% of goalkeepers, 21.4% of forwards, and 20.5% of midfielders experienced at least one adductor strain. The MOI most responsible for these injuries were overuse (30%), change of direction (26.7%), running (13.3%), and kicking (10%). Athletes with previous adductor injury had 167.2 times the odds of adductor injury in a future half-season compared to non-injured athletes. The findings from this study provide further descriptive evidence of player position types and mechanisms related to adductor strain. Insights into the nature of injury within an MLS team and support of previous evidence shows the prevalence of adductor injuries in elite soccer players. Full article

Background/Objectives: Despite widespread use in clinical and athletic settings, validity of handgrip strength (HGS) as a surrogate for maximal strength remains debated, particularly regarding how testing posture influences its predictive value. Moreover, while HGS is frequently considered a marker of ‘total strength’, this term is often vaguely defined, lacking a clear, performance-based framework. Therefore, this study investigates HGS as a potential surrogate measure for one-repetition maximum (1RM) performances in key compound lifts via back squat (BS), bench press (BP), deadlift (DL), and total (TOT), while accounting for variations in testing posture. Methods: Two distinct testing conditions were used to account for postural influences: Experiment 1 implemented high-output standing HGS (HGS_STAND) in 22 recreationally trained males [Wilks Score: 318.51 ± 44.61 au] vs. Experiment 2, which included low-output seated HGS (HGS_SIT) in 22 competitive powerlifters [409.86 ± 46.76 au], with all testing immediately followed by 1RM assessment. Results: Correlational analyses identified the strongest association between HGS_STAND and 1RM DL (r = 0.693, BF₁₀ = 106.42), whereas HGS_SIT exhibited the strongest relationship with 1RM BP (r = 0.732, BF₁₀ = 291.32). Postural effects had a significant impact on HGS outcomes (p < 0.001, η² = 0.413), with HGS_STAND producing higher outputs than HGS_SIT despite lower absolute strength 1RM capabilities. Conclusions: These findings emphasize the role of biomechanical specificity and neuromuscular engagement in grip strength assessments, indicating that HGS can function as a practical surrogate for maximal strength, though its predictive value depends on posture. Strength practitioners, sport scientists, and clinicians should consider these confounding factors when implementing HGS-based monitoring strategies. Full article

Background/Objectives: Mechanical compression of the median nerve is believed to be responsible for idiopathic carpal tunnel syndrome (CTS) due to fibrosis of the subsynovial connective tissue (SSCT). Vascular consequences have also been observed in structures of the carpal tunnel, raising speculation regarding the role of factors such as ischemia and edema in CTS pathology. Methods: We performed a mega-analysis from our database of over 10 years of studies. Mixed-effects models were used to address the disconnect between mechanical and vascular influences on CTS; the effects of biomechanical factors and CTS status were evaluated on carpal tunnel tissue mechanics and blood flow. Altered blood flow was also induced during tissue motion to draw inferences regarding the cyclical relationship between tissue mechanics and fluid flow changes on CTS pathology. Results: Greater movement speed and flexed wrist postures were found to contribute to greater shear strain. Flexed wrist postures and greater fingertip force were found to increase median nerve blood flow. Greater CTS severity was associated with lower median nerve blood flow. Finally, brachial blood flow restriction as a surrogate for elevated carpal tunnel pressure was found to alter tissue motion and increase carpal tunnel tissue shear strain. Conclusions: Finger movement speed, force application, wrist posture, and altered fluid flow in the carpal tunnel contribute to changes in outcomes associated with the development of CTS. The mechanistic findings from this paper should be incorporated into future research to update the damage model for CTS pathology. Full article

Background/Objectives: This study compared step-by-step kinematic measurements from an infrared contact mat (IR-mat) and an inertial measurement unit (IMU) system during bounding and single leg jumping for speed, while also evaluating the validity of algorithms originally developed for sprinting and running when applied to horizontal jumps. The aim was to investigate differences in contact times between the systems. Methods: Nineteen female football players (15 ± 0.5 years, 61.0 ± 5.9 kg, 1.70 ± 0.06 m) performed attempts in both jumps over 20 m with maximum speed, of which the first eight steps were analysed. Results: Significant differences were found between the systems, with the IR-mat recording longer contact times than the IMU. The IR-mat began and ended its measurements slightly earlier and later, respectively, compared to the IMU system, likely due to the IMU’s algorithm, which was developed for sprinting with forefoot contact, while more midfoot and heel landing is used during jumps. Conclusions: Both systems provide reliable measurements; however, the IR mat consistently records slightly longer contact times for horizontal jumps. While the IMU is dependable, it exhibits a consistent bias compared to the IR mat. For bounding, the IR mat begins recording 0.018 s earlier at touch down and stops 0.021 s later. For single leg jumps, it starts 0.024 s earlier and ends 0.021 s later, resulting in contact times that are, on average, 0.039–0.045 s longer. These findings provide valuable insights for coaches and researchers in selecting appropriate measurement tools, highlighting the systematic differences between IR mats and IMUs in horizontal jump analysis. Full article

Purpose: This study explores the impact of take-off distance on hurdling and interval running kinematics in sprint hurdles, recognizing its potential to improve performance. While beginners often use shorter take-off distances, a deeper understanding could inform coaching strategies aimed at improving hurdle technique. Methods: Ten male elite and highly trained hurdlers ran 60 m hurdles under original, short, and long take-off distances (OTD, STD, and LTD, respectively). The sagittal plane kinematics of the fourth hurdle and interval running were obtained using two high-speed cameras at a rate of 120 frames per second. Intraindividual step parameters were compared between conditions. Results: Running speed and step frequency were significantly lower in the STD than in the OTD and LTD. Significant interactions were found for step length with a significantly longer recovery step length in the STD than in the LTD. Furthermore, the hurdling distance was significantly longer in the LTD than in the OTD. In addition, the touchdown distance was significantly shorter in the LTD and longer in the STD compared to the OTD. Therefore, an STD is associated with a shorter acceleration distance between hurdles, whereas an LTD is associated with a longer acceleration distance. Therefore, the take-off distance influenced the distance for acceleration between hurdles, and the recovery step was related to the take-off distance. Conclusions: STD has negative effects on hurdling and interval running, even among elite and highly trained hurdlers. Full article

Bone metastases occur when cancer cells from the primary tumor spread to the bones. The incidence of bone metastases is increasing due to the longer survival of patients with primary tumors, driven by advances in cancer treatments. In patients with multiple bone metastases, care is primarily palliative, aiming to improve their quality of life through pain relief. Bone metastases are strongly associated with pathological fractures, particularly in the femur. In these cases, minimally invasive treatments such as percutaneous cementoplasty and internal fixation with intramedullary nails are growing in popularity. Methods: This manuscript focuses on studying these two therapies by developing virtual models using ANSYS^® software. Thermal and thermomechanical analyses were conducted to evaluate the heat effect resulting from the polymerization of different types of bone cement and to assess the benefits of combining it with internal fixation using intramedullary nails made of different materials. Results: The results highlight the advantages of combining these two techniques compared to cementoplasty alone. Furthermore, the use of Gentamicin Bone Cement (CMW 3^®) with an intramedullary nail made of either material has been shown to provide a more significant functional improvement. Conclusions: The combination of cementoplasty with internal fixation is more effective than cementoplasty alone. The use of CMW 3^® cement with an intramedullary nail made of either material provides greater control over the growth of the metastatic lesion. The chosen injection angle results in an excessive volume of cement, causing a high degree of thermal necrosis. Full article

Background/Objectives: One potential risk factor that remains especially contentious in the anterior cruciate ligament (ACL) injury literature is the role of neuromuscular fatigue in ACL injury risk. Therefore, the purposes of this review are (i) to present the research and practical concepts of lower extremity neuromuscular fatigue; and (ii) to review the literature relating to neuromuscular fatigue as an ACL injury risk factor and mechanism. Methods: A structured review was performed in the Medline database using a search strategy that included terms such as “anterior cruciate ligament injury” and “knee injuries” combined with terms such as “injury” and “fatigue”. Articles were included if they included young healthy participants (18–35) and made a comparison between non-fatigued and fatigued states that were assessed with at least one lower extremity biomechanical variable associated with ACL injury risk. Results: Overall, there were 67 studies included, accounting for 1440 participants (627 male and 813 female) across a variety of sports and activities. Of these, 53 (79%) reported a post-fatigue change in the kinematics, kinetics, neuromuscular, and/or other (e.g., proprioceptive) outcomes that indicate that the participants would be at an increased risk of an ACL injury. The most common argument against fatigue as a risk factor is that ACL injuries do not tend to occur later in a game or season, when it is assumed that athletes would be most fatigued. Conclusions: The evidence presented in this review suggests that localized neuromuscular fatigue is a risk factor, among multiple factors, for ACL injuries, providing another modifiable risk factor that should be considered when developing ACL injury risk reduction interventions. Full article

Background: Linear methods of analysis of variability are concerned with the magnitude of variability and often consider deviations from a central mean as errors. The utilization of nonlinear tools to examine variability allows for the exploration and measurement of the patterns of variability displayed by the system. This methodology explores the deterministic properties of biological signals, in this case, gait, or how previous iterations within the gait cycle influence subsequent and future iterations. The nonlinear analysis of gait variability of the joint angle time series has not been investigated in developing children. Methods: We collected 3 min of treadmill walking data for 28 children between the ages of 2 and 10 years old and analyzed their joint angle time series using nonlinear methods of analysis (sample entropy, largest Lyapunov exponent, and recurrence quantification analysis). Results: Our results indicate that the nonlinear variability of children’s gait increases as children age. Interestingly, this contrasts with the findings from our previous work that showed a decrease in linear variability as children age. The combination of a decrease in linear variability, or a refined and improved stability of gait, as well as an increase in nonlinear variability, or an increase in the sophistication and quality of movement patterns, suggest an overall maturation of the neuromuscular system. Conclusions: Our study indicate that there is a refining of gait with age and motor maturation. This refining speaks to the overall multifaceted organization of systems that defines the maturation of gait. Full article

Introduction: Management strategies for stage II tibialis posterior tendon dysfunction are centered on tendon transfers and osteotomies. One of the most commonly used tendon transfers is flexor digitorum longus (FDL) tendon to navicular, but its superiority over transfers to other locations or transfers of other tendons, along with the role of spring ligament and tibialis posterior tendons, have not been objectively evaluated. Aims: We aimed to quantify both the location and magnitude of secondary stresses that develop as a consequence of the initial pathology. Methods: In this study, we used a computational model to study flat foot development and evaluate the effects of various tendon transfers and failures of passive structural elements, as well as their effect on the biomechanics of the foot. Results: We found that both FDL and FHL transfers have biomechanical advantages and disadvantages. Neither of these transfers decrease the stress on the tibialis posterior tendon if the underlying pathologies such as spring ligament failure are not addressed. Conclusions: Of the tendon transfers evaluated, FDL transfer to the navicular had the most profound effect on reducing the stresses on the spring ligament. Full article

Background/Objectives: Computational models are increasingly used in orthopedic research, such as in the context of total knee arthroplasty (TKA). However, the models’ actual integration in clinical practice is far from routine. Major limitations include the amount of input data, effort, and time required for personalization and simulation. In this paper, we present and validate a patient-specific multi-body musculoskeletal TKA model based on sparse input data to address these limitations. Methods: The simulation model was individualized based on the patients’ bone and knee implant 3D geometries, predicted bony landmarks, and soft tissue attachments using annotated statistical shape models, a statistical squat motion pattern, and a statistically based load case. For the validation, we used publicly accessible in vivo knee contact forces during squatting from four patients of the Grand Challenge Competitions (GCCs). Results: The prediction accuracy was quantified using several error metrics, including the root mean square error (RSME). For GCC3 and GCC5, both the range and trend of the mean in vivo contact forces were well matched by the simulation (RMSE lateral: 8.2–26.1% of body weight (BW); RMSE medial: 15.9–42.7 %BW). In contrast, there were relevant deviations between the experiment and simulation in the trend of contact forces for patient GCC2, as well as in the range of medial contact forces for patient GCC6 (RMSE medial: 52.6 %BW). The model setup time was at the magnitude of 15 min per patient, and the simulation was completed in less than 4 min. Conclusions: When comparing our results with the literature, we found similar accuracy to state-of-the-art models in predicting knee contact forces. While remaining deviations between in vivo and simulation data still warrant investigation and evaluation for clinical significance, the model has already successfully addressed important limitations of these previous models, which represent significant barriers to clinical application. Full article

Background: Previous literature has demonstrated that footstep sounds can be related to the unique gait pattern of individuals. This paper investigates the potential of using footstep sounds as a diagnostic tool in gait analysis. Methods: Fifteen participants ran on a treadmill at 2.7 m/s (6.0 MPH) while simultaneously recording plantar pressure and acoustic signals. Participants repeated the same recordings after completing an exhaustive fatigue protocol, thereby creating a modified gait pattern. Results: The modified gait was evident in the center-of-force trajectory, contact pressures, and acoustic signatures. Analysis of the peak contact pressure and acoustic amplitude showed a modest, statistically significant correlation (r = 0.42, p = 0.02). A method to measure the gait stance time from features in the acoustic signature was tested. Conclusions: The results show that acoustic signals can be used to characterize gait changes, but additional work is needed to link acoustic signal features to gait events like toe lift. Full article