Background. Spastic muscles of patients with cerebral palsy (CP) are considered structurally as shortened muscles, that produce high force in short muscle lengths. Yet, previous intraoperative studies in which muscles’ forces are measured directly as a function of joint angle showed consistently that spastic knee flexor muscles produce a low percentage of their maximum force in flexed knee positions. They also showed effects of epimuscular myofascial force transmission (EMFT): simultaneous activation of different muscles elevated target muscle's force. However, quantification of spastic muscle's force - muscle-tendon unit length (l. MTU. ) data during gait is lacking. Aim. Combining intraoperative experiments with participants’ musculoskeletal models developed based on their gait analyses, we aimed to test the following hypotheses: activated spastic semitendinosus (ST) muscle (1) operates at short l. MTU. 's during gait, forces are (2) low at short l. MTU. 's and (3) increase by co-activating other muscles. Methods. Ten limbs of seven children with CP (GMFCS-II) were tested. Pre-surgery, gait analyses were conducted. Intraoperatively, isometric spastic ST distal forces were measured in ten hip-knee joint angle combinations, in two conditions: (i) activation of the ST individually and (ii) simultaneously with the gracilis, biceps femoris, and rectus femoris muscles endorsing EMFT. In OpenSim, gait_2392 model was used for each limb to (a) calculate l. MTU. per each hip and knee angle combination and the gait relevant l. MTU. range, and (b) analyze gait relevant spastic muscle force - l. MTU. data. Two-way ANOVA was used to compare the patients’ l. MTU. to those of the seven age-matched typically developing (TD) children. l. MTU. values were normalized for the participants’ thigh length. (a) was used to test hypothesis (1) and (b) to test hypotheses (2) and (3): in condition (i), the percent of peak force exerted at the shortest l. MTU. calculated per limb was used as a metric for (2). In condition (ii), mean percent change in muscle force calculated within gait-relevant l. MTU. range was used as a metric for (3). Results. Modeling showed that l. MTU. of spastic ST during gait is shorter on average by 14.1% compared to TD. The ST active force at the shortest gait-relevant l. MTU. was 68.6 (20.6)% (39.9–99.2%) of the peak force. Simultaneous activation of other muscles caused substantial increases in force (minimally by 11.1%, up to several folds, with an exception for one limb). Therefore, only the first and third hypotheses were confirmed. Conclusion. The modeling showed in concert with the clinical considerations that spastic ST may be a shortened muscle that produces high force in short muscle lengths. However, this contrasts intraoperative data, which shows only low forces in flexed knee positions. Note that, the model does not distinguish the muscle-belly and tendon lengths. Therefore, it cannot isolate shorter muscle length and how this compares to the data of TD children remains unknown. Yet, the effects of co-activation of other muscles shown intraoperatively to cause an increase of the spastic ST's force are observed also in muscle force - l. MTU. data characterizing gait. Therefore, if indeed spastic ST produces high forces in short muscle-belly lengths alone, elevated forces due to co-activation of other muscles may be considered as a contributor to the patients’ pathological gait. Otherwise, such EMFT effect may be the main determinant of the pathological condition

Altered mechanical loading is a widely suggested, but poorly understood potential cause of cartilage degeneration in osteoarthritis. In rodents, osteoarthritis is induced following destabilization of the medial meniscus (DMM). This study estimates knee kinematics and contact forces in rats with DMM to gain better insight into the specific mechanisms underlying disease development in this widely-used model. Unilateral knee surgery was performed in adult male Sprague-Dawley rats (n=5 with DMM, n=5 with sham surgery). Radio-opaque beads were implanted on their femur and tibia. 8 weeks following knee surgery, rat gait was recorded using the 3D²YMOX setup (Sanctorum et al. 2019, simultaneous acquisition of biplanar XRay videos and ground reaction forces). 10 trials (1 per rat) were calibrated and processed in XMALab (Knörlein et al. 2016). Hindlimb bony landmarks were labeled on the XRay videos using transfer learning (Deeplabcut, Mathis et al. 2019; Laurence-Chasen et al. 2020). A generic OpenSim musculoskeletal model of the rat hindlimb (Johnson et al. 2008) was adapted to include a 3-degree-of-freedom knee. Inverse kinematics, inverse dynamics, static optimization of muscle forces, and joint reaction analysis were performed. In rats with DMM, knee adduction was lower compared to sham surgery. Ground reaction forces were less variable with DMM, resulting in less variability in joint external moments. The mediolateral ground reaction force was lower, resulting in lower hip adduction moment, thus less force was produced by the rectus femoris. Rats with DMM tended to break rather than propel, resulting in lower hip flexion moment, thus less force was produced by the semimembranosus. These results are consistent with lower knee contact forces in the anteroposterior and axial directions. These preliminary data indicate no overloading of the knee joint in rats with DMM, compared with sham surgery. We are currently expanding our workflow to finite element analysis, to examine mechanical cues in the cartilage of these rats (Fig1G)

Anterior cruciate ligament reconstruction (ACLR) using a semitendinosus (ST) autograft, with or without gracilis (GR), results in donor muscle atrophy and varied tendon regeneration. The effects of harvesting these muscles on muscle moment arm and torque generating capacity have not been well described. This study aimed to determine between-limb differences (ACLR vs uninjured contralateral) in muscle moment arm and torque generating capacity across a full range of hip and knee motions. A secondary analysis of magnetic resonance imaging was undertaken from 8 individuals with unilateral history of ST-GR ACLR with complete ST tendon regeneration. All hamstring muscles and ST tendons were manually segmented. Muscle length (cm), peak cross-sectional area (CSA) (cm. 2. ), and volume (cm. 3. ) were measured in ACLR and uninjured contralateral limbs. OpenSim was used to simulate and evaluate the mechanical consequences of changes in normalised moment arm (m) and torque generating capacity (N.m) between ACLR and uninjured contralateral limbs. Compared to uninjured contralateral limbs, regenerated ST tendon re-insertion varied proximal (+) (mean = 0.66cm, maximum = 3.44cm, minimum = −2.17cm, range = 5.61cm) and posterior (+) (mean = 0.38cm maximum = 0.71cm, minimum = 0.02cm, range = 0.69cm) locations relative to native anatomical positions. Compared to uninjured contralateral limbs, change in ST tendon insertion point in ACLR limbs resulted in 2.5% loss in peak moment arm and a 3.4% loss in peak torque generating capacity. Accounting for changes to both max isometric force and ST moment arm, the ST had a 14.8% loss in peak torque generating capacity. There are significant deficits in ST muscle morphology and insertion points following ST-GR ACLR. The ST atrophy and insertion point migration following ACLR may affect force transmission and distribution within the hamstrings and contribute to persistent deficits in knee flexor and internal rotator strength

This study aimed to quantify self-reported outcomes and walking gait biomechanics in patients following primary and revision THA. The specific goals of this study were to investigate: (i) if primary and revision THA patients have comparable preoperative outcomes; and (2) if revision THA patients have worse postoperative outcomes than primary THA patients. Forty-three patients undergoing primary THA for osteoarthritis and 23 patients undergoing revision THA were recruited and followed longitudinally for their first 12 postoperative months. Reasons for revision were loosening (73%), dislocation (9%), and infection (18%). Patients completed the Hip dysfunction and Osteoarthritis Outcome Score (HOOS), and underwent gait analysis preoperatively, and at 3 and 12 months postoperatively. A 10 camera motion analysis system (V5 Vantage, Vicon, UK) recorded marker trajectories (100 Hz) during walking at self- selected speeds. A generic lower-body musculoskeletal model (Gait2392) was scaled using principal component analysis [1] and the inverse kinematics tool in Opensim 3.3 was used to compute joint angles for the lower limbs in the sagittal plane. Independent samples t-test were used to compare patient reported outcomes between the primary and revision groups at each timepoint. Statistical parametric mapping was used to compare gait patterns between the two groups at each timepoint. Preoperatively, patients undergoing primary THA reported significantly worse pain (p<0.001), symptoms (p<0.001), function (p<0.001), and quality of life (p=0.004). No differences were observed at 3 and 12 months postoperatively between patients who had received a primary or revision THA. The only observed difference in gait pattern was that patients with a revision THA had reduced hip extension at 3 months, but no differences were observed preoperatively and 12 months. Despite the suggestions in the literature that revision THA is bound to have worse outcomes compared to primary THA, we found no differences in in patient-reported outcomes and gait patterns at 12 months postoperatively. This suggests that it may be possible, in some circumstances, for patients following revision THA to achieve similar outcomes to their peers undergoing primary THA

Introduction. Gait laboratory measurement of whole-body kinematics and ground reaction forces during a wide range of activities is frequently performed in joint replacement patient diagnosis, monitoring, and rehabilitation programs. These data are commonly processed in musculoskeletal modeling platforms such as OpenSim and Anybody to estimate muscle and joint reaction forces during activity. However, the processing required to obtain musculoskeletal estimates can be time consuming, requires significant expertise, and thus seriously limits the patient populations studied. Accordingly, the purpose of this study was to evaluate the potential of deep learning methods for estimating muscle and joint reaction forces over time given kinematic data, height, weight, and ground reaction forces for total knee replacement (TKR) patients performing activities of daily living (ADLs). Methods. 70 TKR patients were fitted with 32 reflective markers used to define anatomical landmarks for 3D motion capture. Patients were instructed to perform a range of tasks including gait, step-down and sit-to-stand. Gait was performed at a self-selected pace, step down from an 8” step height, and sit-to-stand using a chair height of 17”. Tasks were performed over a force platform while force data was collected at 2000 Hz and a 14 camera motion capture system collected at 100 Hz. The resulting data was processed in OpenSim to estimate joint reaction and muscle forces in the hip and knee using static optimization. The full set of data consisted of 135 instances from 70 patients with 63 sit-to-stands, 15 right-sided step downs, 14 left-sided step downs, and 43 gait sequences. Two classes of neural networks (NNs), a recurrent neural network (RNN) and temporal convolutional neural network (TCN), were trained to predict activity classification from joint angle, ground reaction force, and anthropometrics. The NNs were trained to predict muscle and joint reaction forces over time from the same input metrics. The 135 instances were split into 100 instances for training, 15 for validation, and 20 for testing. Results. The RNN and TCN yielded classification accuracies of 90% and 100% on the test set. Correlation coefficients between ground truth and predictions from the test set ranged from 0.81–0.95 for the RNN, depending on the activity. Predictions from both NNs were qualitatively assessed. Both NNs were able to effectively learn relationships between the input and output variables. Discussion. The objective of the study was to develop and evaluate deep learning methods for predicting patient mechanics from standard gait lab data. The resulting models classified activities with excellent performance, and showed promise for predicting exact values for loading metrics for a range of different activities. These results indicate potential for real-time prediction of musculoskeletal metrics with application in patient diagnostics and rehabilitation. For any figures or tables, please contact authors directly

Joint assessment through manual physical examination is a fundamental skill that must be acquired by orthopaedic surgeons. These joint assessments allow surgeons to identify soft tissue injuries (e.g. ligament tears) which are critical in identifying appropriate treatment options. The difficulty in communicating the feeling of different joint conditions and the limited opportunities for practice can make these skills challenging to learn, resulting in reduced treatment effectiveness and increased costs. This research seeks to improve the training of joint assessment with the creation of a haptic joint simulator that can train surgeons with increased effectiveness. A first of its kind haptic simulator is presented, which incorporates: a newly defined kinetic knee simulation, a haptic device for user interaction, and a haptic control algorithm. The knee model has been specifically created for this application and allows six degree-of-freedom manipulation of the tibia while considering the effects of ten knee ligament bundles. The model has been mathematically formulated to allow for the high update rates necessary for smooth and stable haptic simulation. Two quantitative assessments were made of the model to confirm its clinical validity. The first was against the widely used OpenSim biomechanical simulation software. Simulations of the model's performance for both anterior-posterior draw tests and varus-valgus rotation tests showed less than 0.7%RMSE for force and 5.5%RMSE for moments. Crucially, the proposed model could generate updated forces in less than 1ms, compared to 188ms for OpenSim. The second validation of the model was against a cadaveric knee that was tested using a validated robotic testing platform. This comparison showed that the model could generate similar force- motion pathways to the cadaveric knee after the model's parameters were scaled to match. Having demonstrated that it is possible to create a computational knee model that has good conformance to gold-standard knee simulations and cadaveric recordings, while updating at less than 1ms, this research has overcome a major hurdle. The next stage of this research will be to incorporate the knee model into a full haptic simulator and perform skill acquisition trials. Given the effectiveness of past haptic training systems in aiding clinical skills acquisition, this research offers a promising way to improve surgeon training, and therefore also patient diagnosis and treatment

An understanding of forces that act on the shoulder joint is important for designing, testing, and evaluating shoulder arthroplasty products. Last year, we presented data describing upper arm motion during eight in-situ hours of occupational and recreational tasks. Using that data the associated humeral head joint forces were calculated with an upper extremity model in OpenSim. Five subjects from a nonrandom sampling of occupations wore the Inertial Measurment Units during a four hour period while at their work place performing their normal work duties and then during the four hour period of non-work activities immediately following. An unscented Kalman filter (UKF) was used to produce the 3D humeral – thoracic angles at 128 Hz from the IMU data. Because of the very large number of data points collected with the IMUs, ninety samples of twenty second duration were randomly selected from each four hour collection period. Using the sampled files, the time scales of the sampled files were scaled by a factor of five and then analyzed with the SUEM static optimization and joint reaction features. Not every sample file could be modeled resulting in an average number of sampled files of 66.7 per subject and condition (work/recreation). The humeral – thoracic angles were then used as input to the Stanford Upper Extremity Model (SUEM) in the OpenSim environment. The SUEM model allowed 2 rotation degree of freedom (rdof) for the forearm (flexion twist), 3 rdof at the humeral – scapular joint, and predicted scapular motion based on the humeral – thoracic elevation angle. All models were run for an assumed 80 kg body weight and included the bone mass of the scapula, clavical, humerus, radius, and ulna, but none of the soft tissue mass. Shoulder muscles were represented by 15 actuators: three heads for each of the deltoid, latissimus dorsi, and pectoralis, and 1 head each for the coracobrachialis, infraspinatous, subscapularis, supraspinatous, termes major, and teres minor. The 5. th. , 50. th. , and 95. th. percentile values of each force component acting on the humeral head from each sampled file for each subject and condition were calculated and the distribution of forces was plotted as a histogram. The overall mean and standard deviation for the 5. th. , 50. th. , and 95. th. percentile values were also determined. Of the A-P and S-I force components, anterior and inferior directed force components were larger than the posterior and superior directed force components. For the M-L force component, the forces were nearly exclusively directed in the medial direction. The 5. th. and 95. th. percentile forces during these subjects' ADL were generally lower than those described by Westerhoff 2009, suggesting that within the limitations of the modeling assumptions, loading experienced during in-situ ADL may be different than loading during laboratory simulation of representative motions

Introduction and Objective. Clinically, it is considered that spastic muscles of patients with cerebral palsy (CP) are shortened, and produce higher force in shorter muscle lengths. Yet, direct quantification of spastic muscles’ forces is rare. Remarkably, previous intraoperative tests in which muscle forces are measured directly as a function of joint angle showed for spastic gracilis (GRA) that its passive forces are low, and only a small percentage of its maximum active force is measured in flexed knee positions. However, the relationship of force characteristics of spastic GRA with its muscle-tendon unit length (l. MTU. ) is unknown. Combining intraoperative experiments with participants’ musculoskeletal models developed based on their gait analyses, we aimed to test if spastic GRA muscle (1) operates at short l. MTU. compared to that of typically developing (TD) children, and exerts higher (2) passive and (3) active forces at shorter lengths, within gait-relevant l. MTU. range. Materials and Methods. Ten limbs of seven children with CP (GMFCS-II) were tested. Pre-surgery, gait analyses were conducted. Intraoperatively, isometric spastic GRA distal forces were measured in ten hip-knee joint angle combinations, in two conditions: (i) passive state and (ii) maximal activation of the GRA exclusively. In OpenSim, gait_2392 model was used for each limb to calculate l. MTU. 's per each hip and knee angle combination and the gait-relevant l. MTU. range, and to analyze gait relevant spastic muscle force - l. MTU. data. l. MTU. values were normalized for the participants’ thigh lengths. Two-way ANOVA was used to compare the patients’ l. MTU. to those of the seven age-matched TD children to test the first hypothesis. In order to test the second and the third hypotheses, Spearman's rank correlation coefficient (ρ) was calculated to seek a correlation between the muscle's operational length (represented by mean l. MTU. within gait cycle) and muscular force characteristics (the percent force at shortest l. MTU. of peak force, either in passive or in active conditions) within gait-relevant l. MTU. range. Results. ANOVA showed that l. MTU. 's of spastic GRA are shorter (on average by 15.4%) compared to those of TD. At the shortest gait-relevant l. MTU. , the GRA passive force was 84.6 (13.7)% of the peak passive force; and the active force was 55.8 (33.9)% of the peak active force. Passive state forces show an increase at longer lengths, whereas active state force characteristics vary in a patient-specific way. Spearman's rank correlation indicated weak correlations between muscle's operational length and muscular force characteristics (ρ= −0.30 P= 0.40, and ρ= −0.27 P= 0.45, for passive and active states, respectively). Therefore, only the first hypothesis was confirmed. Conclusions. Novel muscle force - l. MTU. data for spastic GRA were obtained using intraoperative data and modelling combined. The modelling showed in concert with the clinical considerations that spastic GRA may be a shortened muscle. However, because the model does not distinguish the muscle-belly and tendon lengths, it cannot isolate shorter muscle belly length and how this compares to the data of TD children remains unknown. Moreover, the absence of a strong correlation between shorter operational muscle length and higher force production either in passive or in active conditions highlights the influence of other factors (e.g., muscle structural proteins, and muscle mechanical characteristics including intermuscular interactions etc.) on the pathology rather than ascribing it solely to the length of a spastic muscle itself

INTRODUCTION. Aseptic loosening is the most common failure mode for Total Elbow Arthroplasty (TEA) and is considered to be associated with accelerated polyethylene bearing wear [1, 2]. This study aimed to evaluate three commercially available implant designs under loads associated with daily living. The hypothesis was that more recent designs (Discovery and Nexel) provide greater articular contact areas resulting in lower polyethylene stresses compared to the Coonrad/Morrey (CM). METHODS. Motion tracking was performed on a healthy volunteer during elbow flexion at 0, 45, and 90° shoulder abduction because most daily activities occur with some shoulder abduction [3] resulting in varus stress about the elbow. This kinematic data was used in an OpenSim upper extremity musculoskeletal model [4] to estimate muscle and joint reaction loads with 5lb in hand, consistent with the common clinical restrictions following TEA. Computer aided assemblies of the smallest size implants for each system were imported to ANSYS for finite element analysis. Metallic components were treated as rigid and polyethylene components were modeled using a nonlinear elastoplastic constitutive model calibrated to material data. Articular contacts were frictional. Physiologic joint reaction forces and moments quantified in OpenSim were applied and the resulting peak articular contact area and peak bearing von Mises stresses were assessed. RESULTS. Simulated deformation patterns of CM bearings corresponded well to those reported in retrievals studies [1, 2] supporting the clinical relevance of the modeling approach. Peak stresses for CM and Nexel were consistently found in the central and side bearings respectively. The central bearing stresses remained 2–2.6 times lower in Nexel compared to CM. Peak stress for all three TEA systems increased with shoulder abduction (Fig.1, 2). Highest peak stresses (Fig.2) were obtained in CM and consistently exceeded the polyethylene yield limit; CM showed the lowest contact area (Fig.3). Nexel and Discovery experienced peak polyethylene stresses 26–34% and 17–39% lower than CM respectively (Fig.2). DISCUSSION. Our results support the hypothesis that newer TEA systems provide increased articular contact area and reduced bearing stresses during physiological loading. The cylindrical CM central bearing carries both the joint reaction force and moment leading to edge loading and high stresses (Fig.1). The design of the Nexel central bearing provides limited resistance to varus-valgus moment, thus transferring the moment to the side bearings and reducing central bearing stresses. The hemispherical Discovery bearing design was confirmed to offer a large articular contact area. However, non-concentricity of the contact spheres can lead to edge loading and high polyethylene stresses under off-axis forces. CM and Discovery utilize conventional polyethylene, whereas Nexel utilizes highly cross-linked Vitamin-E polyethylene. This study does not account for the increased wear resistance of Vitamin-E as compared to conventional polyethylene [5]. Long term clinical data are needed to demonstrate how these wear properties, as well as the geometric design which has been shown to impact stresses and contact patterns, translate to in vivo performance. For figures, please contact authors directly

Purpose and Background. Physical mechanisms underlying back pain impairment are poorly understood. Measuring movement features linked to back pain should help understand its causes and decide on best management. Previous kinematic studies have pointed to diverse features distinguishing back pain sufferers. However, the complexity of 3D kinematics means that it is difficult to choose, a priori, which variables or variable combinations are most important. This study set out to obtain a rich set of kinematic data from spinal regions and lower extremities during typical movement tasks, and analyse all of these variables simultaneously to obtain globally important distinguishing features. To this end, a novel distance metric between pairs of motion sequences was used to construct distance matrices. Analyses were carried out directly on these distance matrices. Methods and Results. 20 controls (age: 28 ± 7.6, 10 female) and 20 chronic LBP subjects (age: 41 ± 10.7, 4 female) were recruited. Kinematic data were obtained whilst subjects stood from sitting (‘STS’), picking up (‘Picking’) and lowering (‘Lowering’) a 5kg box, and walking (right (‘WalkRight’) and left sides (‘WalkLeft’)). For each task, permutation tests for group differences were carried out, based on the pseudo-F statistic calculated from the distance matrices. A similar approach was used to identify local differences at time points and joints. Group mean motion sequences were compared using a custom OpenSim model. Significant differences were obtained for STS (pseudo-F=2.8, p=0.017), WalkRight (pseudo-F=3.27, p=0.008) and WalkLeft (pseudo-F=3.39, p=0.005). Conclusion. Comparisons of movement tasks between groups revealed significant differences for STS and walking. Visualisation of group mean motion sequences, and local analyses assisted in the detailed understanding of these differences. This provides a visually intuitive means of studying complex motion differences between groups, without prior assumptions regarding which variables are important. No conflicts of interest. No funding. Original study funded by Arthritis Research UK MRC (Medical Research Council) Centre for Musculoskeletal Health and Work

Cam-type femoroacetabular impingement (FAI) is a common cause for athletic hip injury and early hip osteoarthritis. Although corrective cam FAI surgery can improve patient reported outcome measures (PROMs), it is not clear how surgery affects muscle forces and hip joint loading. Surgery for FAI may redistribute muscle forces and contact forces at the hip joint during routine activities. The purpose of this study was to examine the muscle contributions and hip contact forces during gait in patients prior and after two years of undergoing surgery for cam FAI. Kinematics and kinetics were recorded in 11 patients with symptomatic cam FAI as they completed a gait task. Muscle and hip contact forces during the stance phase were estimated using musculoskeletal modelling and static optimization in OpenSim. All patients reported improvements in PROMs. Post-operatively, patients showed reduced forces in the long head of the biceps femoris at ipsilateral foot-strike and in the rectus femoris at the contralateral foot-strike. The reduced muscle forces decreased sagittal hip moment but did not change hip contact forces. This was the first study to evaluate hip muscle and contact forces in FAI patients post-operatively. Although hip contact forces are not altered following surgery, muscle forces are decreased even after two years. These findings can provide guidance in optimizing recovery protocols after FAI surgery to improve hip flexor and extensor muscle forces

Head collisions in sport can result in catastrophic cervical spine injuries. Musculo-skeletal (MSK) modelling can help analyse the relationship between players' motion, external loading and internal stresses that lead to injury. However, the literature lacks sport specific MSK models. In automotive research the intervertebral disc behaviour has been represented as viscoelastic elements (“bushing”), whose stiffness and damping parameters are often estimated under quasi-static conditions and may lack validity in dynamic impacts. The aim of this study was to develop a validated cervical spine model for axial impacts for future use in the analysis of head-first rugby collisions. A drop test rig was used to replicate a sub-catastrophic axial head impact. A load of 80 N from 0.5 m was applied to the cranial aspect of a C2-C6 porcine spinal specimen mounted in the neutral position. The 3D motion of C3-C5 vertebras (4 kHz) and the cranial axial load (1 MHz) were measured via motion capture (Qualysis, Sweden) and a uniaxial load cell (RDP Electronics Ltd, UK). Specimen specific models were created in NMSBuilder and OpenSim after the vertebrae geometries were obtained from the segmentation of micro-CT images of the specimens. The compressive viscoelastic properties of four vertebral joints (C2-C3 through to C5-C6) were optimised via a Genetic Algorithm (MATLAB v2016b, The Mathworks Inc) to minimise tracking errors. The optimisation converged to a solution of 140–49000 kN/m and 2000–8000 Ns/m for stiffness and damping respectively (RMSE=5.1 mm). Simulated joint displacements ranged between 0.09 – 1.75 mm compared to experimental 0.1 – 0.8 mm. Optimal bushing parameters were higher than previously reported values measured through quasi-static testing. Higher stiffness and damping values could be explained by the higher-dynamics nature of the event analysed related to a different part of the non-linear intervertebral disc load-displacement curve

Manufacturers of reverse shoulder arthroplasty (RSA) implants have recently designed innovative implants to optimise performance in rotator cuff deficient shoulders. These advancements are not without tradeoff and can have negative biomechanical effects. The objective of this study was to develop an integrated FEA kinematic model to compare the muscle forces and joint reaction force (JRF) of 3 different RSA designs. A kinematic model of a normal shoulder joint was adapted from the Delft model and integrated with the OpenSim shoulder model. Static optimisations then allowed for calculation of the individual muscle forces, moment arms and JRF relative to net joint moments. Three dimensional computer models of humeral lateralised design (HLD), glenoid lateral design (GLD), and Grammont design (GD) RSA were integrated and parametric studies were performed. Overall there were decreases in deltoid and rotator cuff muscle forces for all 3 RSA designs. These decreases were greatest in the middle deltoid of the HLD model for abduction and flexion and in the rotator cuff muscles under both internal and external rotation. The joint reactive forces in abduction and flexion decreased similarly for all RSA designs compared to the normal shoulder model, with the greatest decrease seen in the HLD model. These findings demonstrate that the design characteristics implicit in these modified RSA prostheses result in kinematic differences most prominently seen in the deltoid muscle and overall joint reactive forces. Further research utilising this novel integrated model can help guide continued optimisation of RSA design and clinical outcomes

Introduction. Modification in joint loading, and specifically shear stress, is found to be an important mechanical factor in the development of osteoarthritis (OA). Cartilage shear stresses can be investigated using finite element (FE) modelling, where typically in vivo joint loading as measured by an instrumented hip prosthesis is used as boundary condition. However, subject-specific gait characteristics substantially affect joint loading. The goal of this study is to investigate the effect of subject-specific joint loading as calculated using a subject-specific musculoskeletal model and integrated motion capture data on acetabular shear stress. Methods. Three healthy control subjects walked at self-selected speed while measuring marker trajectories (Vicon, Oxford Metrics, UK) and force data (two AMTI force platforms; Watertown, MA). A subject-specific MRI-based musculoskeletal model consisting of 14 segments, 19 degrees of freedom and 88 musculotendon actuators, and including wrapping surfaces around the hip joint, was used. All analyses were performed in OpenSim 3.1. The model was scaled to the dimensions of each subject using the marker positions of a static pose. A kalman smoother procedure was used to calculate joint angles. Muscle forces were calculated using static optimization, minimizing the sum of squared muscle activations, and hip contact forces (HCF) were calculated and normalized to body weight (BW). To calculate shear stress, HCFs and joint angles calculated during the stance phase of gait were imposed to a hip finite element model (hip_n10rb) using FFEbio 2.5. In the model, femoral and acetabular cartilage were represented using the Mooney-Rivlin formulation (c1=6.817, bulk modulus=1358.86) and the pelvis and femur bones as rigid bodies. Peak HCF as well as maximal acetabular shear stress, magnitude and location, and the HCF at the time of maximal shear stress were compared between subjects. Results. Maximal shear stress was lower for S3 compared to S1 and S2 (9.14, 9.48 and 7.14 MPa for S1, S2 and S3 respectively). Nevertheless, HCF at the time instance of peak stress as well as peak HCF were highest for S3 (S1: 2.40/4.54 BW, S2: 2.97/4.78 BW and S3: 3.13/6.46 BW respectively). Maximal shear stress also occurred earlier in the stance phase for S3 compared to S1 and S2 (31, 26 and 11% of the stance phase for S1, S2 and S3 respectively). In addition, the location of the peak maximal shear stress was found to be more superior for S3. Discussion. Subject-specific loading patterns clearly influence the calculated maximal shear stress in the acetabular cartilage, affecting both the magnitude and the location of the stress. In addition, higher shear stresses are not coinciding with higher HCFs. This finding highlights the need of subject-specific rather than generic loading patterns when assessing cartilage shear stresses and associated risk in OA development in individual patients

Introduction. Unicompartmental Knee Replacement Arthroplasty (UKA) is a treatment option for early knee OA that appears under-utilised, partly because of a lack of clear guidance on how to best restore lasting knee function using such devices. Computational tools can help consider inherent uncertainty in patient anatomy, implant positioning and loading when predicting the performance of any implant. In the present research an approach for creating patient-specific finite element models (FEM) incorporating joint and muscle loads was developed to assess the response of the underlying bone to UKA implantation. Methods. As a basis for future uncertainty modelling of UKA performance, the geometriesof 173 lower limbs weregenerated from clinical CT scans. These were segmented (ScanIP, Simpleware Ltd, UK) to reconstruct the 3D surfaces of the femur, tibia, patella and fibula. The appropriate UKA prosthesis (DePuy, U.S.) size was automatically selected according to tibial plateau size and virtually positioned (Figure 1). Boolean operations and mesh generation were accomplished with ScanIP. A patient-specific musculoskeletal model was generated in open-source software OpenSim (Delp et al. 2007) based on the Gait2392 model. The model was scaled to a specific size and muscle insertion points were modified to corresponding points on lower limb of patient. Hip joint load, muscle forces and lower limb posture during gait cycle were calculated from the musculoskeletal model. The FE meshes of lower limb bones were transformed to the corresponding posture at each time point of a gait cycle and FE analyses were performed (Ansys, Inc. U.S) to evaluate the strain distribution on the tibial plateau in the implanted condition. Results. With the tibial component positioned above, along or below the joint line, the lower limb alignment was more varus, remained unaltered or more valgus respectively (Figure 2). With the tibial component positioned 3mm above the joint line, the peak strain in the underlying bone was 670 µstrain on medial (UKA) side and 6780 µstrain on the intact side. With the tibial component positioned 3mm below the joint line, the peak strain was 3010 µstrain on the medial side and 5330 µstrain on the intact side. Here, the strains on the medial side increased by 2640 µstrain whilst they were reduced by 1450 µstrain on the intact side compared to the unimplanted case. Conclusion. The present research has delivered a framework which can be exploited in future uncertainty modelling of UKA performance predictions. The patient-specific model incorporates loading, anatomical and material property variability, and can be applied to evaluate the performance of UKA prostheses for metrics such as stress/strain/micromotions in larger patient populations. For figures/tables, please contact authors directly.

Valgus knee unloader braces are often prescribed as treatment for knee osteoarthritis (OA). These braces are designed to redistribute the loading in the knee, thereby reducing medial contact forces. Patient response to bracing is variable; some patients experience improvements in joint loading, pain, and function, others see little to no effect. We hypothesised that patients who experienced beneficial response to the brace, measured by reductions in medial contact force, could be predicted based on static and dynamic measures. Participants completed a WOMAC questionnaire and walked overground with and without an OA Assist knee brace in a motion capture lab. Eighteen patients with medial compartment OA (8 female, 53.8±7.0 years, BMI 30.3±4.1, median Kellgren-Lawrence grade 4 (range 1–4)) were evaluated. The abduction moment applied by the brace was estimated by multiplying brace deflection by the pre-determined brace stiffness. A generic musculoskeletal model was scaled for each participant based on standing full length radiographs and anatomical markers. Inverse kinematics, inverse dynamics, residual reduction, and muscle analysis were completed in OpenSim 3.2. A static optimisation was then performed to estimate muscle forces and then tibiofemoral contact forces were calculated. Brace effectiveness was defined by the difference in the first peak of the medial contact force between braced and unbraced conditions. Principal component analysis was performed on the hip, knee, and ankle angles and moments from the unbraced walking condition to extract the principal component (PC) scores for these variables. A linear regression procedure was used to determine which variables related to brace effectiveness. Potential regressors included: hip-knee-ankle angle and medial joint space measured radiographically; KL grade; mass; WOMAC scores; unbraced walking speed; and the first two principal component scores for each of the unbraced hip, knee, and ankle joint angles and moments. KL grade, walking speed, and hip adduction moment PC1, which represented the magnitude of the first peak were all found to be correlated with change in medial contact force. The brace was more successful in reducing medial contact force in subjects with higher KL grades, faster self-selected walking speeds, and larger peak external hip adduction moments. The R2 value for the overall regression model was 0.78. The best predictor of brace effectiveness was the hip adduction moment, indicating the need to consider dynamic measures. Participants who had hip adduction moments and walking speeds similar to those of their healthy counterparts saw a greater reduction in medial contact force. Thus, those who responded to bracing had more severe OA as measured by the KL grade but had not experienced changes in their hip adduction moment due to OA. The results of this study suggest that there is potential for an objective criterion for valgus knee brace use to be established

Several studies have shown that gait kinematics[1–3] and hip contact forces (HCFs)[4, 5] of patients following total hip arthroplasty (THA) do not return to normal, although improvements in kinematics are found compared to the pre-surgery. However, the evolution of HCFs after surgery has not been investigated. The goal of this study is to evaluate HCFs during gait in OA patients before and at 2 evaluation moments post-THA. Fourteen unilateral hip OA patients before and 3- and 12-months post-THA surgery walked at self-selected speed, as well as 18 healthy control subjects. 3D marker trajectories were captured using Vicon (Oxford Metrics, UK) and force data was measured using two AMTI force platforms (Watertown, MA). A musculoskeletal model consisting of 14 segments, 19 degrees of freedom and 88 musculotendon actuators and including wrapping surfaces around the hip joint was used[6]. All analyses were performed in OpenSim 3.1[7]. The model was scaled to the dimensions of each subject using the marker positions of a static pose. A kalman smoother procedure was used to calculate joint angles[8]. Muscle forces were calculated using static optimization, minimizing the sum of squared muscle activations. HCFs were calculated and normalized to body weight (BW). First and second peak HCFs were determined and used for statistical analysis. To determine differences between HCFs of OA patients at the different evaluation moments, a Friedman test was used. In case of a significant difference, post-hoc rank-based multiple comparison tests with a Bonferonni adjustment was used. To compare controls and patients at each evaluation moment separate Man-Whitney U tests were used. Differences in HCFs between the affected and non-affected legs were expressed by a symmetry index (SI), i.e. the ratio between the HCFs of the affected leg over the non-affected leg, averaged over the stance phase of the gait cycle. At the first and second HCF peaks, no significant differences were found between pre-, 3- and 12-months post-surgery (first peak average HCF: 2.68, 2.72 and 2.78BW respectively; second peak average HCF: 3.21, 3.83 and 3.77BW respectively). Compared to controls, significant differences are found for all evaluation moments at the first and second HCF peaks (average HCF controls: 3.43 and 5.15BW respectively). The SI was below 1 pre- and 3-months post-surgery (0.88 and 0.85 respectively), indicating decreased loading of the affected compared to the non-affected leg. At 12-months post-surgery SI was close to 1 (0.98). As reported before[4, 5], first or second peak HCFs do not return to normal after THA. Although HCFs increase after THA compared to pre-surgery, significant differences with controls remain. Surprisingly, no significant differences are found between the different evaluation moments of the patients, indicating no clear improvements are found after THA. Further, average HCF peaks at 3- and 12-months post-surgery are similar, indicating no further improvements are found 3-months post-surgery. However, the SI was above or close to 1 at 12-months post-surgery, indicating hip loading evolved to a more symmetrical loading 12-months post-surgery

Children with cerebral palsy (CP) often present femoral bone deformities not accounted for in generic musculoskeletal models [1,2]. MRI-based models can be used to include subject-specific muscle paths [3,4], although this is a time-demanding process. Recently, non-rigid deformation techniques have been used to transform generic bone geometry, including muscle points, onto personalized bones [5]. However, it is still unknown to what extent such an approximation of subject-specific detail affects calculated hip contact forces (HCFs) during gait in CP children. Seven children diagnosed with diplegic CP walked independently at self-selected speed. 3D marker trajectories were captured using Vicon (Oxford Metrics, UK) and force data was measured using two AMTI force platforms (Watertown, MA). MR-images were acquired (Philips Ingenia 1.5T) of all subjects lying supine. Firstly, a generic model [6] was scaled using the marker positions of a static pose. Secondly, a MRI-model containing the subject-specific bone structures and muscle paths of all hip and upper leg muscles was created [3]. Thirdly, the generic femur and pelvis geometries and muscle points were transformed onto the image-based femur and pelvis using an advanced non-rigid deformation procedure (Materialise N.V.). For all models, further analyses were performed in OpenSim 3.1 [7]. A kalman smoother procedure was used to calculate joint angles [8]. Muscle forces were calculated using a static optimization minimizing the sum of squared muscle activities. Next, HCFs were calculated and normalized to body weight (BW). First and second peak HCFs were determined and used for a Kruskal-Wallis test to determine differences between models. In case of a significant difference, a post-hoc rank-based multiple comparison test with Bonferonni adjustment was used. Further, average absolute differences in muscle points between the models was calculated, as well as average differences in moment arm lengths (MALs), reflecting muscle function. Where the scaled generic muscle points differed on average 2.49cm from the MRI points, the non-rigidly deformed points differed 1.54cm from the MRI muscle points. Specifically, the tensor fascia latae differed most between the deformed and MRI models (11.7cm). When considering MALs, the gluteii muscles present an altered function for the generic and deformed models compared to the MRI model for all degrees of freedom of the hip at the time of both HCF peaks. The differences between models resulted in a significantly increased second peak HCF for the MRI models compared to the generic models (first peak average HCF: 3.88BW, 3.95BW and 4.90BW; second peak average HCF: 3.03BW, 4.89BW and 5.32BW for the generic, MRI and non-rigidly deformed models respectively). Although not significantly different, the deformed models calculated slightly increased HCFs compare to the MRI models. The generic models underestimated HCFs compared to the MRI models, while the non-rigidly deformed models slightly overestimated HCFs. However, differences between the deformed and MRI models in terms of muscle points and MALs remain, specifically for the gluteii muscles. Therefore, further user-guided modification of the model based on MR-images will be necessary

The Stanford Upper Extremity Model (SUEM) (Holzbauer, Murray, Delp 2005, Ann Biomed Eng) includes the major muscles of the upper limb and has recently been described in scientific literature for various biomechanical purposes including modeling the muscle behavior after shoulder arthroplasty (Hoenecke, Flores-Hernandez, D'Lima 2014, J Shoulder Elbow Surg; Walker, Struk, Banks 2013, ISTA Proceedings). The initial publication of the SUEM compared the muscle moment arm predictions of the SUEM against various moment arm studies and all with the scapula fixed. A more recent study (Ackland, Pak, and Pandy 2008, J Anat) is now available that can be used to compare SUEM moment arm predictions to cadaver data for similar muscle sub-regions, during abduction and flexion motions, and with simulated scapular motion. SUEM muscle moment arm component vectors were calculated using the OpenSim Analyze Tool for an idealized abduction and an idealized flexion motion from 10° to 90° that corresponded to the motions described in Ackland for the cadaver arms. The normalized, averaged muscle moment arm data for the cadavers was manually digitized from the published figures and then resampled into uniform angles matching the SUEM data. Standard deviations of the muscle moment arms from the cadaver study were calculated from source data provided by the study authors. Python code was then used to calculate the differences, percent differences, and root-mean-square (RMS) values between the data sets. Of the 14 muscle groups in the SUEM, the smallest difference in predicted and measured moment arm was for the supraspinatus during the abduction task, with an RMS of the percent difference of 11.4%. In contrast, the middle latissimus dorsi had an RMS percent difference over 400% during the flexion task. The table presents the RMS difference and the RMS of the percent difference for the muscles with the largest abduction and adduction moment arms (during abduction) and the largest flexion and extension moment arms (during flexion). The moment arm data for the SUEM model and the cadaver data (with 1 standard deviation band) during the motion of the same muscles are provided in Figure 1 for the Abduction motion task and in Figure 2 for the Flexion motion task. It is challenging to simulate the three dimensional, time variant geometries of shoulder muscles while maintaining model fidelity and optimizing computational cost. Dividing muscles in to sub regions and using wrapping line segment approximations appears a reasonable strategy though more work could improve model accuracy especially during complex three dimensional motions