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.

Introduction

Many finite element (FE) studies have been performed in the past to assess the biomechanical performance of TKA and THA components. The boundary conditions have often been simplified to a few peak loads. With the availability of personalized musculoskeletal (MS) models we becomes possible to estimate dynamic muscle and prosthetic forces in a patient specific manner. By combining this knowledge with FE models, truly patient specific failure analyses can be performed.

In this study we applied this combined technique to the femoral part of a cementless THR and calculated the cyclic micro-motions of the stem relative to the bone in order to assess the potential for bone ingrowth.

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

Predictions of lower limb growth are based upon historical data, collected from patients who had coexistent poliomyelitis. By utilising standardised longitudinal prospective European data, our objective was to generate superior estimates for the age and rate at which lower limb skeletal maturity is reached; thus improving the timing of epiphysiodesis, for the management of leg length discrepancy. The Avon Longitudinal Study of Parents and Children of the 90s (ALSPAC) is a longitudinal cohort study of children recruited antenatally 2. Using a previously validated Multiplier Method, a sequence of leg length multipliers were calculated for each child. 15,458 individuals were recruited to the ALSPAC study; and of those whose growth was measured, 52% were boys and 48% girls, each with an average of eight recording episodes. 25,828 leg length multiplier (LLM) values were calculated with final recordings taken at a mean age of 15.5 years. From this data, the age at which girls reach skeletal maturity (LLM=1) is 11 months later than previously calculated and for boys nearly 9 months earlier. With nearly 4000 more children recruited in this cohort than preceding studies, this study brings increased power to future leg length calculations

Aims

Disorders of bone integrity carry a high global disease burden, frequently requiring intervention, but there is a paucity of methods capable of noninvasive real-time assessment. Here we show that miniaturized handheld near-infrared spectroscopy (NIRS) scans, operated via a smartphone, can assess structural human bone properties in under three seconds.