Deep Learning in Gait Parameter Prediction for OA and TKA Patients Wearing IMU Sensors

Commercially available activity monitors such as the fitbit charge and apple watch show poor validity in patients with gait aids after total knee arthroplasty

PURPOSE

The aim of this study is to determine the validity of consumer grade step counter devices during the early recovery period after knee replacement surgery.

METHODS

Twenty-three participants wore a Fitbit Charge or Apple Watch Series 4 smart watch and performed a walking test along a 50-metre hallway. There were 9 males and 14 females included in the study with an average age of 68.5 years and BMI of 32. Each patient wore both the Fitbit Charge and Apple Watch while completing the walking test and an observer counted the ground truth value using a thumb-push tally counter. This test was repeated pre-operatively with no gait aid, immediately post operatively with a walker, at 6 weeks follow up with a cane and at 6 months with no gait aid. Bland-Altman plots were performed for all walking tests to compare the agreement between measurement techniques.

RESULTS

Mean overall agreement of step count for pre-operative and at 6 months for subjects walking without gait aids was excellent for both the Apple Watch vs. actual and Fitbit vs. actual with bias values ranging from - 0.87 to 1.36 with limits of agreement (LOA) ranging between - 10.82 and 15.91. While using a walker both devices showed extremely little agreement with the actual step count with bias values between 22.5 and 24.37 with LOA between 11.7 and 33.3. At 6 weeks post-op while using a cane, both the Apple Watch and Fitbit devices had a range of bias values between - 2.8 and 5.73 with LOA between - 13.51 and 24.97.

CONCLUSIONS

These devices show poor validity in the early post operative setting, especially with the use of gait aids, and therefore results should be interpreted with caution.

Gait classification of knee osteoarthritis patients using shoe-embedded internal measurement units sensor

BACKGROUND

Knee osteoarthritis negatively affects the gait of patients, especially that of elderly people. However, the assessment of wearable sensors in knee osteoarthritis patients has been under-researched. During clinical assessments, patients may change their gait patterns under the placebo effect, whereas wearable sensors can be used in any environment.

METHODS

Sixty patients with knee osteoarthritis and 20 control subjects were included in the study. Wearing shoes with an IMU sensor embedded in the insoles, the participants were required to walk along a walkway. The sensor data were collected during the gait. To discriminate between healthy and knee osteoarthritis patients and to classify different subgroups of knee osteoarthritis patients (patients scheduled for surgery vs. patients not scheduled for surgery; bilateral knee osteoarthritis diagnosis vs. unilateral knee osteoarthritis diagnosis), we used a machine learning approach called the support vector machine. A total of 88 features were extracted and used for classification.

FINDINGS

The patients vs. healthy participants were classified with 71% accuracy, 85% sensitivity, and 56% specificity. The "patients scheduled for surgery" vs. "patients not scheduled for surgery" were classified with 83% accuracy, 83% sensitivity, and 81% specificity. The bilateral knee osteoarthritis diagnosis vs. unilateral knee osteoarthritis diagnosis was classified with 81% accuracy, 75% sensitivity, and 79% specificity.

INTERPRETATION

Gait analysis using wearable sensors and machine learning can discriminate between healthy and knee osteoarthritis patients and classify different subgroups with reasonable accuracy, sensitivity, and specificity. The proposed approach requires no complex gait factors and is not limited to controlled laboratory settings.

3D gait analysis in children using wearable sensors: feasibility of predicting joint kinematics and kinetics with personalized machine learning models and inertial measurement units

Introduction: Children's walking patterns evolve with age, exhibiting less repetitiveness at a young age and more variability than adults. Three-dimensional gait analysis (3DGA) is crucial for understanding and treating lower limb movement disorders in children, traditionally performed using Optical Motion Capture (OMC). Inertial Measurement Units (IMUs) offer a cost-effective alternative to OMC, although challenges like drift errors persist. Machine learning (ML) models can mitigate these issues in adults, prompting an investigation into their applicability to a heterogeneous pediatric population. This study aimed at 1) quantifying personalized and generalized ML models' performance for predicting gait time series in typically developed (TD) children using IMUs data, 2) Comparing random forest (RF) and convolutional neural networks (CNN) models' performance, 3) Finding the optimal number of IMUs required for accurate predictions. Methodology: Seventeen TD children, aged 6 to 15, participated in data collection involving OMC, force plates, and IMU sensors. Joint kinematics and kinetics (targets) were computed from OMC and force plates' data using OpenSim. Tsfresh, a Python package, extracted features from raw IMU data. Each target's ten most important features were input in the development of personalized and generalized RF and CNN models. This procedure was initially conducted with 7 IMUs placed on all lower limb segments and then performed using only two IMUs on the feet. Results: Findings suggested that the RF and CNN models demonstrated comparable performance. RF predicted joint kinematics with a 9.5% and 19.9% NRMSE for personalized and generalized models, respectively, and joint kinetics with an NRMSE of 10.7% for personalized and 15.2% for generalized models in TD children. Personalized models provided accurate estimations from IMU data in children, while generalized models lacked accuracy due to the limited dataset. Furthermore, reducing the number of IMUs from 7 to 2 did not affect the results, and the performance remained consistent. Discussion: This study proposed a promising personalized approach for gait time series prediction in children, involving an RF model and two IMUs on the feet.

Kinematic Analysis of Human Gait in Healthy Young Adults Using IMU Sensors: Exploring Relevant Machine Learning Features for Clinical Applications

BACKGROUND

Gait is the manner or style of walking, involving motor control and coordination to adapt to the surrounding environment. Knowing the kinesthetic markers of normal gait is essential for the diagnosis of certain pathologies or the generation of intelligent ortho-prostheses for the treatment or prevention of gait disorders. The aim of the present study was to identify the key features of normal human gait using inertial unit (IMU) recordings in a walking test.

METHODS

Gait analysis was conducted on 32 healthy participants (age range 19-29 years) at speeds of 2 km/h and 4 km/h using a treadmill. Dynamic data were obtained using a microcontroller (Arduino Nano 33 BLE Sense Rev2) with IMU sensors (BMI270). The collected data were processed and analyzed using a custom script (MATLAB 2022b), including the labeling of the four relevant gait phases and events (Stance, Toe-Off, Swing, and Heel Strike), computation of statistical features (64 features), and application of machine learning techniques for classification (8 classifiers).

RESULTS

Spider plot analysis revealed significant differences in the four events created by the most relevant statistical features. Among the different classifiers tested, the Support Vector Machine (SVM) model using a Cubic kernel achieved an accuracy rate of 92.4% when differentiating between gait events using the computed statistical features.

CONCLUSIONS

This study identifies the optimal features of acceleration and gyroscope data during normal gait. The findings suggest potential applications for injury prevention and performance optimization in individuals engaged in activities involving normal gait. The creation of spider plots is proposed to obtain a personalised fingerprint of each patient's gait fingerprint that could be used as a diagnostic tool. A deviation from a normal gait pattern can be used to identify human gait disorders. Moving forward, this information has potential for use in clinical applications in the diagnosis of gait-related disorders and developing novel orthoses and prosthetics to prevent falls and ankle sprains.

Validation of algorithms for calculating spatiotemporal gait parameters during continuous turning using lumbar and foot mounted inertial measurement units

Spatiotemporal gait parameters such as step time and walking speed can be used to quantify gait performance and determine physical function. Inertial measurement units (IMUs) allow for the measurement of spatiotemporal gait parameters in unconstrained environments but must be validated against a gold standard. While many IMU systems and algorithms have been validated during treadmill walking and overground walking in a straight line, fewer studies have validated algorithms during more complex walking conditions such as continuous turning in different directions. This study explored the concurrent validity in a population of healthy adults (range 26-52 years) of three different algorithms using lumbar and foot mounted IMUs to calculate spatiotemporal gait parameters: two methods utilizing an inverted pendulum model, and one method based on strapdown integration. IMU data was compared to a Vicon twelve-camera optoelectronic system, using data collected from 9 participants performing straight walking and continuous walking trials at different speeds, resulting in 162 walking trials in total. Intraclass correlation coefficients (ICCA,1) for absolute agreement were calculated between the algorithm outputs and Vicon output. Temporal parameters were comparable in all methods and ranged from moderate to excellent, except double support time which was poor. Strapdown integration performed better for estimating spatial parameters than pendulum models during straight walking, but worse during turning. Selecting the most appropriate model should take into consideration both speed and walking condition.

Postoperative Activity and Knee Function of Patients after Total Knee Arthroplasty: A Sensor-Based Monitoring Study

Inertial measurement units (IMUs) are increasingly being used to assess knee function. The aim of the study was to record patients' activity levels and to detect new parameters for knee function in the early postoperative phase after TKA. Twenty patients (n = 20) were prospectively enrolled. Two sensors were attached to the affected leg. The data were recorded from the first day after TKA until discharge. Algorithms were developed for detecting steps, range of motion, horizontal, sitting and standing postures, as well as physical therapy. The mean number of steps increased from day 1 to discharge from 117.4 (SD ± 110.5) to 858.7 (SD ± 320.1), respectively. Patients' percentage of immobilization during daytime (6 a.m. to 8 p.m.) was 91.2% on day one and still 69.9% on the last day. Patients received daily continuous passive motion therapy (CPM) for a mean of 36.4 min (SD ± 8.2). The mean angular velocity at day 1 was 12.2 degrees per second (SD ± 4.4) and increased to 28.7 (SD ± 16.4) at discharge. This study shows that IMUs monitor patients' activity postoperatively well, and a wide range of interindividual motion patterns was observed. These sensors may allow the adjustment of physical exercise programs according to the patient's individual needs.

Enhancing Wearable Gait Monitoring Systems: Identifying Optimal Kinematic Inputs in Typical Adolescents

Machine learning-based gait systems facilitate the real-time control of gait assistive technologies in neurological conditions. Improving such systems needs the identification of kinematic signals from inertial measurement unit wearables (IMUs) that are robust across different walking conditions without extensive data processing. We quantify changes in two kinematic signals, acceleration and angular velocity, from IMUs worn on the frontal plane of bilateral shanks and thighs in 30 adolescents (8-18 years) on a treadmills and outdoor overground walking at three different speeds (self-selected, slow, and fast). Primary curve-based analyses included similarity analyses such as cosine, Euclidean distance, Poincare analysis, and a newly defined bilateral symmetry dissimilarity test (BSDT). Analysis indicated that superior-inferior shank acceleration (SI shank Acc) and medial-lateral shank angular velocity (ML shank AV) demonstrated no differences to the control signal in BSDT, indicating the least variability across the different walking conditions. Both SI shank Acc and ML shank AV were also robust in Poincare analysis. Secondary parameter-based similarity analyses with conventional spatiotemporal gait parameters were also performed. This normative dataset of walking reports raw signal kinematics that demonstrate the least to most variability in switching between treadmill and outdoor walking to help guide future machine learning models to assist gait in pediatric neurological conditions.

BioMAT: An Open-Source Biomechanics Multi-Activity Transformer for Joint Kinematic Predictions Using Wearable Sensors

Through wearable sensors and deep learning techniques, biomechanical analysis can reach beyond the lab for clinical and sporting applications. Transformers, a class of recent deep learning models, have become widely used in state-of-the-art artificial intelligence research due to their superior performance in various natural language processing and computer vision tasks. The performance of transformer models has not yet been investigated in biomechanics applications. In this study, we introduce a Biomechanical Multi-activity Transformer-based model, BioMAT, for the estimation of joint kinematics from streaming signals of multiple inertia measurement units (IMUs) using a publicly available dataset. This dataset includes IMU signals and the corresponding sagittal plane kinematics of the hip, knee, and ankle joints during multiple activities of daily living. We evaluated the model's performance and generalizability and compared it against a convolutional neural network long short-term model, a bidirectional long short-term model, and multi-linear regression across different ambulation tasks including level ground walking (LW), ramp ascent (RA), ramp descent (RD), stair ascent (SA), and stair descent (SD). To investigate the effect of different activity datasets on prediction accuracy, we compared the performance of a universal model trained on all activities against task-specific models trained on individual tasks. When the models were tested on three unseen subjects' data, BioMAT outperformed the benchmark models with an average root mean square error (RMSE) of 5.5 ± 0.5°, and normalized RMSE of 6.8 ± 0.3° across all three joints and all activities. A unified BioMAT model demonstrated superior performance compared to individual task-specific models across four of five activities. The RMSE values from the universal model for LW, RA, RD, SA, and SD activities were 5.0 ± 1.5°, 6.2 ± 1.1°, 5.8 ± 1.1°, 5.3 ± 1.6°, and 5.2 ± 0.7° while these values for task-specific models were, 5.3 ± 2.1°, 6.7 ± 2.0°, 6.9 ± 2.2°, 4.9 ± 1.4°, and 5.6 ± 1.3°, respectively. Overall, BioMAT accurately estimated joint kinematics relative to previous machine learning algorithms across different activities directly from the sequence of IMUs signals instead of time-normalized gait cycle data.

A comparison of machine learning models' accuracy in predicting lower-limb joints' kinematics, kinetics, and muscle forces from wearable sensors

A combination of wearable sensors' data and Machine Learning (ML) techniques has been used in many studies to predict specific joint angles and moments. The aim of this study was to compare the performance of four different non-linear regression ML models to estimate lower-limb joints' kinematics, kinetics, and muscle forces using Inertial Measurement Units (IMUs) and electromyographys' (EMGs) data. Seventeen healthy volunteers (9F, 28 ± 5 years) were asked to walk over-ground for a minimum of 16 trials. For each trial, marker trajectories and three force-plates data were recorded to calculate pelvis, hip, knee, and ankle kinematics and kinetics, and muscle forces (the targets), as well as 7 IMUs and 16 EMGs. The features from sensors' data were extracted using the Tsfresh python package and fed into 4 ML models; Convolutional Neural Networks (CNN), Random Forest (RF), Support Vector Machine, and Multivariate Adaptive Regression Spline for targets' prediction. The RF and CNN models outperformed the other ML models by providing lower prediction errors in all intended targets with a lower computational cost. This study suggested that a combination of wearable sensors' data with an RF or a CNN model is a promising tool to overcome the limitations of traditional optical motion capture for 3D gait analysis.

A Novel Classification of Coronal Plane Knee Joint Instability Using Nine-Axis Inertial Measurement Units in Patients with Medial Knee Osteoarthritis

The purpose of this study was to propose a novel classification of varus thrust based on gait analysis with inertial motion sensor units (IMUs) in patients with medial knee osteoarthritis (MKOA). We investigated thigh and shank acceleration using a nine-axis IMU in 69 knees with MKOA and 24 (control) knees. We classified varus thrust into four phenotypes according to the relative medial-lateral acceleration vector patterns of the thigh and shank segments: pattern A (thigh medial, shank medial), pattern B (medial, lateral), pattern C (lateral, medial), and pattern D (lateral, lateral). Quantitative varus thrust was calculated using an extended Kalman filter-based algorithm. We compared the differences between our proposed IMU classification and the Kellgren-Lawrence (KL) grades for quantitative varus thrust and visible varus thrust. Most of the varus thrust was not visually perceptible in early-stage OA. In advanced MKOA, increased proportions of patterns C and D with lateral thigh acceleration were observed. Quantitative varus thrust was significantly increased stepwise from patterns A to D. This novel IMU classification has better clinical utility due to its ability to detect subtle kinematic changes that cannot be captured with conventional motion analysis even in the early stage of MKOA.

Concerns surrounding application of artificial intelligence in hip and knee arthroplasty

Literature surrounding artificial intelligence (AI)-related applications for hip and knee arthroplasty has proliferated. However, meaningful advances that fundamentally transform the practice and delivery of joint arthroplasty are yet to be realized, despite the broad range of applications as we continue to search for meaningful and appropriate use of AI. AI literature in hip and knee arthroplasty between 2018 and 2021 regarding image-based analyses, value-based care, remote patient monitoring, and augmented reality was reviewed. Concerns surrounding meaningful use and appropriate methodological approaches of AI in joint arthroplasty research are summarized. Of the 233 AI-related orthopaedics articles published, 178 (76%) constituted original research, while the rest consisted of editorials or reviews. A total of 52% of original AI-related research concerns hip and knee arthroplasty (n = 92), and a narrative review is described. Three studies were externally validated. Pitfalls surrounding present-day research include conflating vernacular (“AI/machine learning”), repackaging limited registry data, prematurely releasing internally validated prediction models, appraising model architecture instead of inputted data, withholding code, and evaluating studies using antiquated regression-based guidelines. While AI has been applied to a variety of hip and knee arthroplasty applications with limited clinical impact, the future remains promising if the question is meaningful, the methodology is rigorous and transparent, the data are rich, and the model is externally validated. Simple checkpoints for meaningful AI adoption include ensuring applications focus on: administrative support over clinical evaluation and management; necessity of the advanced model; and the novelty of the question being answered.

Cite this article: Bone Joint J 2022;104-B(12):1292–1303.

The role of commercially available smartphone apps and wearable devices in monitoring patients after total knee arthroplasty: a systematic review

Purpose

Commercially available smartphone apps and wearable devices have proven valuable in a variety of clinical settings, yet their utility in measuring physical activity and monitoring patient status following total knee arthroplasty (TKA) remains unclear.

Methods

A systematic review was performed to assess the evidence supporting the use of smartphone apps and wearable devices to assist rehabilitation interventions following TKA. A search was conducted in the PubMed, Cochrane, Medline, and Web of Science databases in September 2021.

Results

One hundred and seventy-six studies were retrieved, of which 15 met inclusion criteria, including 6 randomized control trials. Four of these studies utilized smartphone apps, seven utilized wearable devices, and four utilized a combination of both. A total of 1607 TKA patients participated in the included studies. For primary outcomes, three reported on device accuracy, three on recovery prediction, two on functional recovery, two on physical activity promotion, two on patient compliance, two on pain control, and one on healthcare utilization.

Conclusion

Commercially available smartphone apps and wearable devices were shown to capably monitor physical activity and improve patient engagement following TKA, making them potentially viable adjuncts or replacements to traditional rehabilitation programs. Components of interventions such as step goals, app-based patient engagement platforms, and patient-specific benchmarks for recovery may improve effectiveness. However, future research should focus on the economics of implementation, long-term outcomes, and optimization of compliance and accuracy when using these devices.

A Deep Learning Approach for Gait Event Detection from a Single Shank-Worn IMU: Validation in Healthy and Neurological Cohorts

Many algorithms use 3D accelerometer and/or gyroscope data from inertial measurement unit (IMU) sensors to detect gait events (i.e., initial and final foot contact). However, these algorithms often require knowledge about sensor orientation and use empirically derived thresholds. As alignment cannot always be controlled for in ambulatory assessments, methods are needed that require little knowledge on sensor location and orientation, e.g., a convolutional neural network-based deep learning model. Therefore, 157 participants from healthy and neurologically diseased cohorts walked 5 m distances at slow, preferred, and fast walking speed, while data were collected from IMUs on the left and right ankle and shank. Gait events were detected and stride parameters were extracted using a deep learning model and an optoelectronic motion capture (OMC) system for reference. The deep learning model consisted of convolutional layers using dilated convolutions, followed by two independent fully connected layers to predict whether a time step corresponded to the event of initial contact (IC) or final contact (FC), respectively. Results showed a high detection rate for both initial and final contacts across sensor locations (recall ≥92%, precision ≥97%). Time agreement was excellent as witnessed from the median time error (0.005 s) and corresponding inter-quartile range (0.020 s). The extracted stride-specific parameters were in good agreement with parameters derived from the OMC system (maximum mean difference 0.003 s and corresponding maximum limits of agreement (−0.049 s, 0.051 s) for a 95% confidence level). Thus, the deep learning approach was considered a valid approach for detecting gait events and extracting stride-specific parameters with little knowledge on exact IMU location and orientation in conditions with and without walking pathologies due to neurological diseases.

Analysis of Basketball Technical Movements Based on Human-Computer Interaction with Deep Learning

With the continuous development of computer technology, analysis techniques based on various types of sports data sets are also evolving. One typical representative is image-based motion recognition technology, which enables video action recognition with a certain degree of feasibility. In basketball technical action videos, technical action has obvious characteristics. The athletes in the footage in sports videos are relatively fixed, and the scenes are relatively homogeneous, so technical action analysis of basketball technical action videos has certain advantages. However, there are many challenges in basketball technical action recognition, mainly including the fact that basketball techniques are numerous and complex. To address the above issues, this research proposes a 3D convolutional neural network framework that two different resolution image inputs are performed on the basketball technical action dataset. The experimental results show that the algorithmic process designed in this study is effective for action recognition on the basketball technical action dataset.

Predicting Knee Joint Kinematics from Wearable Sensor Data in People with Knee Osteoarthritis and Clinical Considerations for Future Machine Learning Models

Deep learning models developed to predict knee joint kinematics are usually trained on inertial measurement unit (IMU) data from healthy people and only for the activity of walking. Yet, people with knee osteoarthritis have difficulties with other activities and there are a lack of studies using IMU training data from this population. Our objective was to conduct a proof-of-concept study to determine the feasibility of using IMU training data from people with knee osteoarthritis performing multiple clinically important activities to predict knee joint sagittal plane kinematics using a deep learning approach. We trained a bidirectional long short-term memory model on IMU data from 17 participants with knee osteoarthritis to estimate knee joint flexion kinematics for phases of walking, transitioning to and from a chair, and negotiating stairs. We tested two models, a double-leg model (four IMUs) and a single-leg model (two IMUs). The single-leg model demonstrated less prediction error compared to the double-leg model. Across the different activity phases, RMSE (SD) ranged from 7.04° (2.6) to 11.78° (6.04), MAE (SD) from 5.99° (2.34) to 10.37° (5.44), and Pearson’s R from 0.85 to 0.99 using leave-one-subject-out cross-validation. This study demonstrates the feasibility of using IMU training data from people who have knee osteoarthritis for the prediction of kinematics for multiple clinically relevant activities.

Wearable Sensor-Based Step Length Estimation During Overground Locomotion Using a Deep Convolutional Neural Network

Step length is a critical gait parameter that allows a quantitative assessment of gait asymmetry. Gait asymmetry can lead to many potential health threats such as joint degeneration, difficult balance control, and gait inefficiency. Therefore, accurate step length estimation is essential to understand gait asymmetry and provide appropriate clinical interventions or gait training programs. The conventional method for step length measurement relies on using foot-mounted inertial measurement units (IMUs). However, this may not be suitable for real-world applications due to sensor signal drift and the potential obtrusiveness of using distal sensors. To overcome this challenge, we propose a deep convolutional neural network-based step length estimation using only proximal wearable sensors (hip goniometer, trunk IMU, and thigh IMU) capable of generalizing to various walking speeds. To evaluate this approach, we utilized treadmill data collected from sixteen able-bodied subjects at different walking speeds. We tested our optimized model on the overground walking data. Our CNN model estimated the step length with an average mean absolute error of 2.89 ± 0.89 cm across all subjects and walking speeds. Since wearable sensors and CNN models are easily deployable in real-time, our study findings can provide personalized real-time step length monitoring in wearable assistive devices and gait training programs.

Inertial measurement units and application for remote healthcare in hip and knee osteoarthritis: a narrative review (Preprint)

Background

Measuring and modifying movement-related joint loading is integral to the management of lower extremity osteoarthritis (OA). Although traditional approaches rely on measurements made within the laboratory or clinical environments, inertial sensors provide an opportunity to quantify these outcomes in patients’ natural environments, providing greater ecological validity and opportunities to develop large data sets of movement data for the development of OA interventions.

Objective

This narrative review aimed to discuss and summarize recent developments in the use of inertial sensors for assessing movement during daily activities in individuals with hip and knee OA and to identify how this may translate to improved remote health care for this population.

Methods

A literature search was performed in November 2018 and repeated in July 2019 and March 2021 using the PubMed and Embase databases for publications on inertial sensors in hip and knee OA published in English within the previous 5 years. The search terms encompassed both OA and wearable sensors. Duplicate studies, systematic reviews, conference abstracts, and study protocols were also excluded. One reviewer screened the search result titles by removing irrelevant studies, and 2 reviewers screened study abstracts to identify studies using inertial sensors as the main sensing technology and a primary outcome related to movement quality. In addition, after the March 2021 search, 2 reviewers rescreened all previously included studies to confirm their relevance to this review.

Results

From the search process, 43 studies were determined to be relevant and subsequently included in this review. Inertial sensors have been successfully implemented for assessing the presence and severity of OA (n=11), assessing disease progression risk and providing feedback for gait retraining (n=7), and remotely monitoring intervention outcomes and identifying potential responders and nonresponders to interventions (n=14). In addition, studies have validated the use of inertial sensors for these applications (n=8) and analyzed the optimal sensor placement combinations and data input analysis for measuring different metrics of interest (n=3). These studies show promise for remote health care monitoring and intervention delivery in hip and knee OA, but many studies have focused on walking rather than a range of activities of daily living and have been performed in small samples (<100 participants) and in a laboratory rather than in a real-world environment.

Conclusions

Inertial sensors show promise for remote monitoring, risk assessment, and intervention delivery in individuals with hip and knee OA. Future opportunities remain to validate these sensors in real-world settings across a range of activities of daily living and to optimize sensor placement and data analysis approaches.

The Use of Synthetic IMU Signals in the Training of Deep Learning Models Significantly Improves the Accuracy of Joint Kinematic Predictions

Gait analysis based on inertial sensors has become an effective method of quantifying movement mechanics, such as joint kinematics and kinetics. Machine learning techniques are used to reliably predict joint mechanics directly from streams of IMU signals for various activities. These data-driven models require comprehensive and representative training datasets to be generalizable across the movement variability seen in the population at large. Bottlenecks in model development frequently occur due to the lack of sufficient training data and the significant time and resources necessary to acquire these datasets. Reliable methods to generate synthetic biomechanical training data could streamline model development and potentially improve model performance. In this study, we developed a methodology to generate synthetic kinematics and the associated predicted IMU signals using open source musculoskeletal modeling software. These synthetic data were used to train neural networks to predict three degree-of-freedom joint rotations at the hip and knee during gait either in lieu of or along with previously measured experimental gait data. The accuracy of the models’ kinematic predictions was assessed using experimentally measured IMU signals and gait kinematics. Models trained using the synthetic data out-performed models using only the experimental data in five of the six rotational degrees of freedom at the hip and knee. On average, root mean square errors in joint angle predictions were improved by 38% at the hip (synthetic data RMSE: 2.3°, measured data RMSE: 4.5°) and 11% at the knee (synthetic data RMSE: 2.9°, measured data RMSE: 3.3°), when models trained solely on synthetic data were compared to measured data. When models were trained on both measured and synthetic data, root mean square errors were reduced by 54% at the hip (measured + synthetic data RMSE: 1.9°) and 45% at the knee (measured + synthetic data RMSE: 1.7°), compared to measured data alone. These findings enable future model development for different activities of clinical significance without the burden of generating large quantities of gait lab data for model training, streamlining model development, and ultimately improving model performance.

Deep learning approach for guiding three-dimensional computed tomography reconstruction of lower limbs for robotically-assisted total knee arthroplasty

BACKGROUND

Robotic-assisted total knee arthroplasty (TKA) was performed to promote the accuracy of bone resection and mechanical alignment. Among these TKA system procedures, 3D reconstruction of CT data of lower limbs consumes significant manpower. Artificial intelligence (AI) algorithms applying deep learning has been proved efficient in automated identification and visual processing.

METHODS

CT data of a total of 200 lower limbs scanning were used for AI-based 3D model construction and CT data of 20 lower limbs scanning were utilised for verification.

RESULTS

We showed that the performance of an AI-guided 3D reconstruction of CT data of lower limbs for robotic-assisted TKA was similar to that of the operator-based approach. The time of 3D lower limb model construction using AI was 4.7 min. AI-based 3D models can be used for surgical planning.

CONCLUSION

AI was used for the first time to guide the 3D reconstruction of CT data of lower limbs for facilitating robotic-assisted TKA. Incorporation of AI in 3D model reconstruction before TKA might reduce the workload of radiologists.