A multi-task framework based on SDA–LSTM fusion network for gait phase recognition and gait cycle percentage progression prediction by IMU for forward walking

Recovering an independent and efficient gait is essential for social participation and quality of life after a stroke. However, muscle coactivation (MCo) is considered a major limiting factor. This study aimed to systematically review the methods used to assess lower limb MCo during overground walking in stroke survivors. Major scientific databases were searched, and 19 articles were ultimately included. Data extraction focused on gait cycle detection, surface electromyograph (sEMG) acquisition settings, signal processing, and MCo calculation, specifically the coactivation index (CoI) and coactivation duration (CoD). Results revealed variability in methodological choices across all stages required to assess MCo. Regarding acquisition and processing approaches, several commonalities were identified, such as the use of Surface ElectroMyoGraphy for the Non-Invasive Assessment of Muscles guidelines for electrode placement and the application of a single threshold for muscle activation detection. Conversely, significant variability was found in the selection of gait cycle phases and signal filtering techniques. Furthermore, this heterogeneity was particularly pronounced in the final MCo calculation stage. Specifically, CoI represents the most diverse approach, with seven distinct formulas, which prevents the proposal of a common method. Conversely, CoD measurement shared a unique calculation formula associated with more common methodological elements, allowing for the proposal of a standardized approach. This common method was also aligned with current sEMG recommendations. Overall, although standardized methods have been established to assess the MCo, neither CoI nor CoD has yet undergone sufficient reliability testing to support routine clinical use. This review underscores the need for methodological rigour to improve the comparability of MCo assessments in stroke rehabilitation research.

Gait analysis of children with neurodevelopmental disorders using gait profile score.

Prosthetic ankle-foot devices provide valuable assistance for individuals with a unilateral transtibial amputation (TTA) to effectively engage in daily living activities, although users often experience diminished walking performance such as increased metabolic cost, knee joint loading, and dynamic balance asymmetry due to the lack of torque control from commonly prescribed passive devices. Consequently, active powered prosthetic devices have been developed; however, it is unclear how to optimally tune them. The purpose of this study was to identify the optimal ankle torque profile of a powered ankle-foot prosthesis that improves walking performance for individuals with TTA. Specifically, we used a musculoskeletal simulation-based optimization framework to optimize a powered prosthesis torque profile while emulating group averaged kinematics and ground reaction forces (GRFs). We compared the metabolic cost, knee joint loading, sagittal plane dynamic balance symmetry, and torque profiles across the following simulated conditions: a passive prosthesis tracking individuals with TTA walking data, a powered prosthesis tracking able-bodied walking data, and a powered prosthesis that separately minimized metabolic cost, knee joint loading, and dynamic balance asymmetry. Distinct torque profiles emerged for each measure, but there was no clear trend in the positive prosthetic work performed, which suggests increased prosthetic work alone is insufficient to improve walking performance. Further analysis showed the prosthetic torque must be properly timed over the gait cycle to improve each measure. This study provides a framework for future work developing customized controllers for powered prostheses to improve various aspects of walking performance for individuals with TTA.

A decrease in Minimum Foot Clearance (MFC), which represents the minimum vertical distance of the foot from the ground surface during the swing phase of the gait cycle, is one of the primary contributors to tripping-related falls among people with stroke. Accurate prediction of upcoming MFC values, understanding their potential range, and evaluating whether future values will fall within the variability of actual MFC values are crucial for early fall risk assessment. This work proposes a new Transformer model for predicting multistep MFC values in stroke survivors collected from the affected lower-limb during walking on a treadmill utilizing a two-head self-attention mechanism. We introduce a data-driven conditional post-normalization projection approach to enhance the performance of multistep prediction for MFC values. In addition, we introduce a statistical moment-matching loss function during training to account for significant data variability, such as that observed in individuals with stroke. We compare the performance of our model with two other deep learning models. Our findings indicate that the Transformer model training is faster and achieves an average Mean Absolute Error (MAE) of approximately 0.0035, Maximum Absolute Error (MaxAE) of 0.0085 and a Root Mean Square Error (RMSE) of 0.0043, using the leave-one-out cross-validation (LOOCV) method. The highest prediction within the upper bound (PWUB) of around 89% is achieved with self-attention LSTM model. The main contribution of our work is to identify when an individual stroke survivor is at increased risk of tripping-related falls due to a lower MFC value in their affected lower limb.

Geriatric hip fractures, particularly unstable intertrochanteric fractures, pose a significant global health challenge with high morbidity and mortality. The optimal surgical management remains debated, with both Closed Reduction and Internal Fixation (CRIF) and Bipolar Hemiarthroplasty (BHA) being widely accepted treatment options. However, there is a conspicuous paucity of objective, quantitative data comparing the longitudinal gait recovery trajectories following these two distinct surgical approaches. This study aimed to leverage validated wearable inertial sensors to systematically characterize and compare the spatiotemporal gait recovery patterns of elderly patients undergoing CRIF versus BHA for unstable intertrochanteric fractures. This prospective cohort study enrolled 48 patients aged > 70 years with unstable intertrochanteric fractures (AO/OTA 31-A2/A3). Patients were allocated to Group A (CRIF with PFNA, n = 21) or Group B (BHA, n = 27) based on surgeon preference and patient factors. Baseline frailty was assessed using the modified Frailty Index (mFI) and it was comparable between 2 groups. A comprehensive battery of spatiotemporal gait parameters was objectively quantified using validated Gait Up Physilog® 5 wearable inertial sensors at 2 weeks, 1 month, 3 months, and 6 months postoperatively. At 3 months postoperatively, the BHA group demonstrated markedly superior gait parameters compared to the CRIF group, including significantly higher cadence (82.9 ± 27.3 vs. 51.7 ± 22.3 steps/min; p = 0.025) and shorter gait cycle time (1.6 ± 0.5 s vs. 2.7 ± 1.0 s; p = 0.028), reflecting a more confident and fluid gait pattern in the BHA cohort. However, this functional advantage attenuated by 6 months, as the CRIF group exhibited a compensatory "catch-up" phenomenon, achieving comparable gait parameters (p > 0.05) once fracture consolidation was established. No postoperative complications were observed in either group during the study period. While both BHA and CRIF are effective treatments for unstable intertrochanteric fractures, BHA offers a clinically meaningful advantage in early-to-mid-term gait recovery. The immediate intrinsic stability provided by arthroplasty facilitates earlier return to confident ambulation, which is of paramount importance for frail geriatric patients with limited physiological reserve. For select high-risk patients, BHA may be preferred to mitigate risks of prolonged immobility and optimize functional recovery. Level III, Therapeutic Study.

Plantar pressure measurement is frequently employed to assess the gait characteristics of stroke patients. This study developed a gait assessment system for stroke patients utilizing a self-developed five-region smart insole, with the goal of objectively evaluating their gait characteristics. A total of 75 stroke patients (14 at Brunnstrom stage III, 18 at stage IV, 27 at stage V, and 16 at stage VI) and 24 healthy subjects (designated "Health") were recruited. A smart insole was used to collect the subjects' gait parameters and plot plantar pressure curves. The results showed the following: (1) There were statistically significant differences among the five groups in the durations of the gait cycle, the double support phase, the bilateral single support phases, and the swing phase, as well as in the ratio of peak pressure to body weight for the bilateral toe bone regions, medial metatarsal regions, lateral metatarsal regions, and heel regions. No statistically significant differences were found in the bilateral arch peak pressure-to-body weight ratio among the five groups. (2) Each of the five subject groups exhibited a unique plantar pressure curve profile. The smart insole used in this study can provide objective gait assessment for stroke patients and effectively differentiate the gait characteristics of patients at different Brunnstrom stages.

Most resurfacing arthroplasty is performed using the posterior approach, while total hip arthroplasty is increasingly performed via the anterior approach with objective benefit noted in the sagittal plane kinematics. We surmised that the muscle releases needed for resurfacing arthroplasty using the posterior approach would reduce the strength of push-off as seen in gait studies of total hip arthroplasty. Method 19 healthy controls were compared with 17 posterior and 17 anterior patients for hip resurfacing, recruited from a prospective gait study of lower limb arthroplasty. PROMs and gait characteristics were captured more than 12months following surgery. Oxford Hip Score, MET score and gait were analysed using ground reaction forces and motion capture. Statistical parametric mapping was used across the entire gait cycle. Results The median and modal OHS was 48/48 for both groups. The mean MET was 13.1 for posterior and 12.6 for anterior. Top walking speed of 7.5km/hr for the control and posterior were similar to the 7km/hr of anterior. Ground reaction forces were symmetric at all speeds. Maximum sagittal plane motion of 42° for posterior and 45° for anterior were indistinguishable from the control group throughout the range of motion. Coronal plane range was 15° for both groups – identical to the healthy controls. Discussion The gait characteristics of these two groups were indistinguishable even at higher speeds from the healthy control group. The speed reached and the different gait characteristics did not differ, despite the substantial difference in soft tissue releases needed. There may be a benefit to using the anterior approach for resurfacing arthroplasty, but ground reaction forces and motion capture in a gait lab at 12 months was unable to detect any lasting impact of either approach. Surgeons should continue to use their preferred approach.

ASSESSING DISABILITY LEVEL AND FATIGUE IN MULTIPLE SCLEROSIS WITH SMART SOCK SENSOR TECHNOLOGY.

The One-Year Postoperative Trajectories of Knee Joint Angle Recovery During Walking After Total Knee Arthroplasty in Women: A Cross-Sectional Study.

Individuals with unilateral transtibial amputation (TTA) often demonstrate asymmetrical gait patterns, which are further exacerbated during load carriage as passive prostheses cannot modulate their mechanical stiffness to accommodate the increased demand. Load carriage also increases the loads on the intact limb that can lead to overuse injuries and is also associated with increased metabolic cost and decreased forward propulsion. While active and passive load-suspended backpacks have been studied in healthy populations, no studies have explored the use of active backpacks to improve the walking performance for individuals with TTA. Therefore, the purpose of this study was to identify active backpack loading patterns to improve the metabolic cost, forward propulsion, and knee joint loading of individuals with TTA using a musculoskeletal simulation-based optimization framework. Different loading patterns were simulated using a time-varying actuator force applied to an active backpack over the gait cycle. The magnitude and timing of the actuator force were optimized for each performance criterion that resulted in a unique actuation pattern for each biomechanical measure. Interestingly, similar improvements relative to a passive backpack were observed across all actuation patterns, regardless of the optimization criteria. With all active backpacks, the force impulse experienced by the body from the dynamic load decreased, which resulted in increased forward propulsion, decreased intact knee joint loading, and improved metabolic cost. This suggests that active backpacks have the potential to improve walking performance during load carriage for individuals with TTA.

Artificial intelligence predictions of knee kinematics, kinetics, and internal biomechanics during walking in people with knee osteoarthritis: A systematic review and meta-analysis.

Translation gain is a key Redirected Walking (RDW) technique in Virtual Reality (VR) that enables users to navigate virtual environments (VEs) larger than the available physical space. The technique was originally developed to scale users' walking distance in the VE and is typically applied continuously, regardless of the user's motion state. We introduce Gait-Synced Translation Gain (GSTG), a novel approach that adapts translation gain by synchronizing it with the user's gait cycle. GSTG leverages the single-limb support phase of walking-when users are less stable and thus less sensitive to external disturbances-to apply higher levels of gain. This approach allows greater manipulation while preserving natural walking sensations and avoiding additional cybersickness. A user study comparing GSTG with continuous translation gain demonstrates significant improvements in perceived naturalness and comfort. Our results highlight the potential of gait-synchronized gain to enhance immersion, offering new possibilities for more realistic and comfortable VR locomotion.

Kinetic Interlimb Asymmetries in Loading and Propulsion During Walking From 2 to 8 Months After Anterior Cruciate Ligament Reconstruction.

To dissect specific gait abnormalities associated with upper motor neuron (UMN) dysfunction in amyotrophic lateral sclerosis (ALS) by controlling for overall disease severity and to develop a multivariate classification model. We performed 3D gait analysis on 118 ALS patients and 1796 healthy controls (HC). ALS patients were categorized into those with ALS with UMN dysfunction((ALS-UMN), n = 70) and those without ALS without UMN signs ((ALS-Numn), n = 48) lower limb UMN signs based on neurological examination. Gait parameters were compared, and their association with UMN involvement was analyzed using partial correlation (controlling for ALSFRS-R score) and machine learning models (Random Forest and Least Absolute Shrinkage and Selection Operator (Lasso) regression). Compared with HC, ALS patients exhibited widespread gait deterioration (e.g., reduced speed, increased step width, p < 0.001). After controlling for ALSFRS-R, specific parameters, including reduced stride, increased step width, prolonged double support, and elevated gait cycle time asymmetry, remained independently associated with UMN severity (PENN score, p < 0.01). A multivariate model incorporating key features demonstrated fair discriminative ability for identifying ALS-UMN patients, with an area under the curve (AUC) of 0.690, a sensitivity of 0.816, and a specificity of 0.418. Quantitative gait analysis reveals a distinct spatiotemporal pattern linked to UMN dysfunction in ALS. A model based on gait features shows potential, particularly high sensitivity, for identifying patients with pyramidal signs, supporting the exploratory utility of objective gait metrics for motor phenotyping in ALS, pending external validation.

The present study plays a crucial role in enhancing the safety and perceived quality of life for users of bone-anchored prostheses. It focuses on developing an innovative protective component using various metallic materials to identify and mitigate potential risks during use, thereby reducing the likelihood of sudden fracture and maintaining the system’s structural integrity. The protective element is manufactured from Ti6Al4V alloy, while the safety pin is made from ductile cast iron. This combination allows controlled fracture of the protective element without complete separation of the prosthesis, thereby reducing the risk of falls. To optimise the numerical analysis, a 3D model of the prosthesis and its protective component was created using SolidWorks software. Loading conditions were simulated to reflect two critical phases of the gait cycle: heel strike and toe-off. The analysis revealed that the highest stress occurred during the toe-off phase, reaching 248 MPa, with a safety factor of 1.6, demonstrating the design’s ability to prevent sudden failure. Tensile testing showed that ductile cast iron is a suitable material for the safety component. Although Ti6Al4V alloy surpasses it in tensile strength, ductile cast iron’s lower strength ensures a controlled and less catastrophic failure under excessive loading. Numerical results confirmed a high safety factor for the protective system, indicating improved reliability and mechanical load resistance. This study presents a novel approach aimed at improving the safety of bone-anchored prostheses by minimising injury risks due to mechanical overload, ultimately enhancing user comfort and confidence.

Myeloid neoplasms (MNs) are most frequently diagnosed among adults aged 60 years and older. Cancer and chemotherapy can cause gait disturbances and increase fall risk in older adults with MNs. Exercise may improve gait, but there is a lack of research among older adults with MNs undergoing active chemotherapy. We explored gait changes following a home-based mobile health exercise intervention during 2 cycles of outpatient chemotherapy (8-12 weeks). In a single-arm pilot study, we included adults aged 60 years and older with MNs undergoing chemotherapy. Geriatric Oncology-Exercise for Cancer Patients intervention integrates a progressive aerobic walking and resistance exercise program with a mobile app. We assessed gait by using a waist-worn G-Walk motion sensor during a 6-minute walk at the preintervention and postintervention time points. Spatiotemporal outcomes included cadence (steps per minute), velocity (meters per minute), normalized stride (stride length normalized over height), and swing duration (percentage of the gait cycle during which a foot is in the air when walking). Regularity outcomes that measure gait rhythm included variability of normalized stride and variability of swing duration. Variability for both outcomes was quantified as the SD across all gait cycles. We calculated Cohen d effect sizes (ESs) for change in gait outcomes and used the Spearman rank correlation to correlate changes in daily steps and resistance exercise duration with gait outcomes. We included 13 patients (mean age 71, SD 4.8 years); most were male (n=8, 61.5%), White individuals (n=12, 92.3%), and non-Hispanic individuals (n=13, 100%). Average daily steps were 3084 (SD 1765.5) at the preintervention time point and 3757 (SD 2623.6) at the postintervention time point. Patients performed resistance exercises for 25 minutes per day, 4 days per week at low intensity (mean rating of perceived exertion 3/10, SD 1.3). At the postintervention time point, we observed numerical changes in gait outcomes, including increased cadence (mean +4.6, SD 14.6 steps per minute; P=.24; ES=0.38) and decreased variability in normalized stride (mean -1.4%, SD 8.5%; P=.34; ES=-0.18) and swing duration (mean -0.1%, SD 1.1%; P=.54; ES=-0.15), although these improvements were not statistically significant. Increased daily steps significantly correlated with decreased swing duration variability (r=-0.72; P=.01). Resistance exercise duration significantly correlated with increased cadence (r=0.54; P=.06) and velocity (r=0.56; P=.05). In our exploratory analyses, better adherence to exercise correlated with improved gait outcomes. Our ongoing pilot randomized controlled trial (ClinicalTrials.gov identifier: NCT04981821) will further examine the effects of the Geriatric Oncology-Exercise for Cancer Patients intervention on gait outcomes in this population.

Hoof trimming and shoeing techniques are used to manage and prevent equine limb injuries. However, quantitative studies comparing the effects of different shoeing techniques on equine joint biomechanics over the full gait cycle are lacking. To measure and compare joint motion and net torques at the distal forelimb joints when horses walk overground unshod, with a standard flat shoe, and with a rocker shoe. In vivo study. Gait data were recorded from 12 sound horses during walking. Three shoeing conditions were tested: unshod, flat shoe, and rocker shoe. Data were recorded for each shoeing condition immediately after trimming (short hoof condition) and again after 6 weeks of hoof growth (long hoof condition). Three-dimensional motion capture and retro-reflective skin markers recorded left forelimb motion, while synchronised force plates measured the corresponding ground reaction force. Inverse dynamics was used to calculate the net torques developed about the distal forelimb joints. Statistical comparisons were performed with multilevel mixed effects generalised linear models. While there were limited effects of trimming and shoeing, the rocker shoe was associated with higher walking speed (by 9.3 ± 9.7%) and reduced stride duration (by 4.9 ± 6.9%) compared with the flat shoe for the short hoof condition (p < 0.001). Throughout the stride cycle, the fetlock joint was less extended (by 9.0 ± 13.7°) while the distal interphalangeal joint (DIPJ) was more extended (by 10.7 ± 16.6°) for both shoeing types compared to unshod regardless of hoof growth (p < 0.005). Higher peak torques were generated at the DIPJ for flat shoe compared to unshod (by 0.05 ± 0.27 Nm/kg) in the short hoof condition, and for flat shoe compared to rocker shoe (by 0.03 ± 0.14 Nm/kg) in the long hoof condition (p < 0.05 for both). The horses were tested at a low-speed walking gait. Forelimb joint biomechanics did not differ substantially across the three shoeing and two hoof-growth conditions. Future studies should test the robustness of these findings at the trot and canter.

Ear-worn wearables (aka: earbuds, hearables, or earables) are commonly used by runners for entertainment, and many modern devices also include inertial sensors for user interaction. We propose harnessing the technology embedded in earbuds to capture fundamental aspects of running mechanics and make them available to the wider community of users, outside a lab setting. While other wearables such as insolesor ankle-/sacrum-mounted inertial measurement units have already been presented, ear-worn devices may have a better potential for adoption and therefore offer an optimal compromise between validity of running gait analysis and usability. Thirty healthy participants (18 males, 12 females) ran on an instrumented treadmill (54,000 gait cycles) and floor-mounted force plates (2800 gait cycles) at a variety of speeds. Building on the information brought about by the vibrations transmitted to, and motion of, the head, we devised a gait event detection algorithm and a regression model to predict vertical ground reaction force waveforms. The validation of outcomes against quantities from force plates shows an average mean absolute percentage error of 4.8% on temporal metrics and 9.0% on scalar ground reaction force derived metrics. Additionally, the model tracks the full vertical ground reaction force curve well, achieving an normalized root mean square error of 11.1% on unseen participants. Overall, we show comparable accuracy from an ear-worn consumer device in temporal and kinetic gait parameter estimation to specialist devices, paving the way for accessible running gait monitoring.

Wearable sensors enable collection of ground reaction force (GRF) data in real-world settings, but translating these data into meaningful activity classifications remains challenging, particularly for subject-specific monitoring. We present a dataset of wearable GRF measurements from 14 subjects walking across 18 combinations of speed and incline, along with a machine learning pipeline for step-level classification of loading behaviors. Continuous GRF signals were segmented into individual gait cycles and transformed into features using TSFRESH, followed by feature selection and Random Forest classification. Subject-specific models achieved a mean Top-1 accuracy of 0.664 (SD = 0.053), exceeding chance performance (0.056), with Top-2 and Top-3 accuracies of 0.836 and 0.904, respectively. Accuracy remained similar for incline-only classification (0.688 ± 0.030) but increased for speed-only classification (0.903 ± 0.097). These results demonstrate that step-level GRF data can support accurate classification of locomotion-related loading conditions and enable the development of subject-specific models for monitoring individual activity. The dataset and pipeline provide a foundation for future work in wearable biomechanics and personalized analysis of musculoskeletal loading.