Muscle synergies may improve optimization prediction of knee contact forces during walking.

The ability to predict patient-specific joint contact and muscle forces accurately could improve the treatment of walking-related disorders. Muscle synergy analysis, which decomposes a large number of muscle electromyographic (EMG) signals into a small number of synergy control signals, could reduce the dimensionality and thus redundancy of the muscle and contact force prediction process. This study investigated whether use of subject-specific synergy controls can improve optimization prediction of knee contact forces during walking. To generate the predictions, we performed mixed dynamic muscle force optimizations (i.e., inverse skeletal dynamics with forward muscle activation and contraction dynamics) using data collected from a subject implanted with a force-measuring knee replacement. Twelve optimization problems (three cases with four subcases each) that minimized the sum of squares of muscle excitations were formulated to investigate how synergy controls affect knee contact force predictions. The three cases were: (1) Calibrate+Match where muscle model parameter values were calibrated and experimental knee contact forces were simultaneously matched, (2) Precalibrate+Predict where experimental knee contact forces were predicted using precalibrated muscle model parameters values from the first case, and (3) Calibrate+Predict where muscle model parameter values were calibrated and experimental knee contact forces were simultaneously predicted, all while matching inverse dynamic loads at the hip, knee, and ankle. The four subcases used either 44 independent controls or five synergy controls with and without EMG shape tracking. For the Calibrate+Match case, all four subcases closely reproduced the measured medial and lateral knee contact forces (R2 ≥ 0.94, root-mean-square (RMS) error < 66 N), indicating sufficient model fidelity for contact force prediction. For the Precalibrate+Predict and Calibrate+Predict cases, synergy controls yielded better contact force predictions (0.61 < R2 < 0.90, 83 N < RMS error < 161 N) than did independent controls (-0.15 < R2 < 0.79, 124 N < RMS error < 343 N) for corresponding subcases. For independent controls, contact force predictions improved when precalibrated model parameter values or EMG shape tracking was used. For synergy controls, contact force predictions were relatively insensitive to how model parameter values were calibrated, while EMG shape tracking made lateral (but not medial) contact force predictions worse. For the subject and optimization cost function analyzed in this study, use of subject-specific synergy controls improved the accuracy of knee contact force predictions, especially for lateral contact force when EMG shape tracking was omitted, and reduced prediction sensitivity to uncertainties in muscle model parameter values.

Predictive Simulations of Neuromuscular Coordination and Joint-Contact Loading in Human Gait.

We implemented direct collocation on a full-body neuromusculoskeletal model to calculate muscle forces, ground reaction forces and knee contact loading simultaneously for one cycle of human gait. A data-tracking collocation problem was solved for walking at the normal speed to establish the practicality of incorporating a 3D model of articular contact and a model of foot-ground interaction explicitly in a dynamic optimization simulation. The data-tracking solution then was used as an initial guess to solve predictive collocation problems, where novel patterns of movement were generated for walking at slow and fast speeds, independent of experimental data. The data-tracking solutions accurately reproduced joint motion, ground forces and knee contact loads measured for two total knee arthroplasty patients walking at their preferred speeds. RMS errors in joint kinematics were < 2.0° for rotations and < 0.3 cm for translations while errors in the model-computed ground-reaction and knee-contact forces were < 0.07 BW and < 0.4 BW, respectively. The predictive solutions were also consistent with joint kinematics, ground forces, knee contact loads and muscle activation patterns measured for slow and fast walking. The results demonstrate the feasibility of performing computationally-efficient, predictive, dynamic optimization simulations of movement using full-body, muscle-actuated models with realistic representations of joint function.

Correction to: Predictive Simulations of Neuromuscular Coordination and Joint-Contact Loading in Human Gait.

In the "Materials and Methods" section, the link provided at the bottom of the second paragraph should be https://simtk.org/home/dcwithjtcontact/ .

Dynamic simulation of knee-joint loading during gait using force-feedback control and surrogate contact modelling.

The aim of this study was to perform multi-body, muscle-driven, forward-dynamics simulations of human gait using a 6-degree-of-freedom (6-DOF) model of the knee in tandem with a surrogate model of articular contact and force control. A forward-dynamics simulation incorporating position, velocity and contact force-feedback control (FFC) was used to track full-body motion capture data recorded for multiple trials of level walking and stair descent performed by two individuals with instrumented knee implants. Tibiofemoral contact force errors for FFC were compared against those obtained from a standard computed muscle control algorithm (CMC) with a 6-DOF knee contact model (CMC6); CMC with a 1-DOF translating hinge-knee model (CMC1); and static optimization with a 1-DOF translating hinge-knee model (SO). Tibiofemoral joint loads predicted by FFC and CMC6 were comparable for level walking, however FFC produced more accurate results for stair descent. SO yielded reasonable predictions of joint contact loading for level walking but significant differences between model and experiment were observed for stair descent. CMC1 produced the least accurate predictions of tibiofemoral contact loads for both tasks. Our findings suggest that reliable estimates of knee-joint loading may be obtained by incorporating position, velocity and force-feedback control with a multi-DOF model of joint contact in a forward-dynamics simulation of gait.

Contribution of tibiofemoral joint contact to net loads at the knee in gait.

Inverse dynamics analysis is commonly used to estimate the net loads at a joint during human motion. Most lower-limb models of movement represent the knee as a simple hinge joint when calculating muscle forces. This approach is limited because it neglects the contributions from tibiofemoral joint contact forces and may therefore lead to errors in estimated muscle forces. The aim of this study was to quantify the contributions of tibiofemoral joint contact loads to the net knee loads calculated from inverse dynamics for multiple subjects and multiple gait patterns. Tibiofemoral joint contact loads were measured in four subjects with instrumented implants as each subject walked at their preferred speed (normal gait) and performed prescribed gait modifications designed to treat medial knee osteoarthritis. Tibiofemoral contact loads contributed substantially to the net knee extension and knee adduction moments in normal gait with mean values of 16% and 54%, respectively. These findings suggest that knee-contact kinematics and loads should be included in lower-limb models of movement for more accurate determination of muscle forces. The results of this study may be used to guide the development of more realistic lower-limb models that account for the effects of tibiofemoral joint contact at the knee.

Simultaneous prediction of muscle and contact forces in the knee during gait.

Musculoskeletal models are currently the primary means for estimating in vivo muscle and contact forces in the knee during gait. These models typically couple a dynamic skeletal model with individual muscle models but rarely include articular contact models due to their high computational cost. This study evaluates a novel method for predicting muscle and contact forces simultaneously in the knee during gait. The method utilizes a 12 degree-of-freedom knee model (femur, tibia, and patella) combining muscle, articular contact, and dynamic skeletal models. Eight static optimization problems were formulated using two cost functions (one based on muscle activations and one based on contact forces) and four constraints sets (each composed of different combinations of inverse dynamic loads). The estimated muscle and contact forces were evaluated using in vivo tibial contact force data collected from a patient with a force-measuring knee implant. When the eight optimization problems were solved with added constraints to match the in vivo contact force measurements, root-mean-square errors in predicted contact forces were less than 10 N. Furthermore, muscle and patellar contact forces predicted by the two cost functions became more similar as more inverse dynamic loads were used as constraints. When the contact force constraints were removed, estimated medial contact forces were similar and lateral contact forces lower in magnitude compared to measured contact forces, with estimated muscle forces being sensitive and estimated patellar contact forces relatively insensitive to the choice of cost function and constraint set. These results suggest that optimization problem formulation coupled with knee model complexity can significantly affect predicted muscle and contact forces in the knee during gait. Further research using a complete lower limb model is needed to assess the importance of this finding to the muscle and contact force estimation process.

Decreased knee adduction moment does not guarantee decreased medial contact force during gait.

Excessive contact force is believed to contribute to the development of medial compartment knee osteoarthritis. The external knee adduction moment (KAM) has been identified as a surrogate measure for medial contact force during gait, with an abnormally large peak value being linked to increased pain and rate of disease progression. This study used in vivo gait data collected from a subject with a force-measuring knee implant to assess whether KAM decreases accurately predict corresponding decreases in medial contact force. Changes in both quantities generated via gait modification were analyzed statistically relative to the subject's normal gait. The two gait modifications were a "medial thrust" gait involving knee medialization during stance phase and a "walking pole" gait involving use of bilateral walking poles. Reductions in the first (largest) peak of the KAM (32-33%) did not correspond to reductions in the first peak of the medial contact force. In contrast, reductions in the second peak and angular impulse of the KAM (15-47%) corresponded to reductions in the second peak and impulse of the medial contact force (12-42%). Calculated reductions in both KAM peaks were highly sensitive to rotation of the shank reference frame about the superior-inferior axis of the shank. Both peaks of medial contact force were best predicted by a combination of peak values of the external KAM and peak absolute values of the external knee flexion moment (R(2) = 0.93). Future studies that evaluate the effectiveness of gait modifications for offloading the medial compartment of the knee should consider the combined effect of these two knee moments.

Evaluation of predicted knee-joint muscle forces during gait using an instrumented knee implant.

Musculoskeletal modeling and optimization theory are often used to determine muscle forces in vivo. However, convincing quantitative evaluation of these predictions has been limited to date. The present study evaluated model predictions of knee muscle forces during walking using in vivo measurements of joint contact loading acquired from an instrumented implant. Joint motion, ground reaction force, and tibial contact force data were recorded simultaneously from a single subject walking at slow, normal, and fast speeds. The body was modeled as an 8-segment, 21-degree-of-freedom articulated linkage, actuated by 58 muscles. Joint moments obtained from inverse dynamics were decomposed into leg-muscle forces by solving an optimization problem that minimized the sum of the squares of the muscle activations. The predicted knee muscle forces were input into a 3D knee implant contact model to calculate tibial contact forces. Calculated and measured tibial contact forces were in good agreement for all three walking speeds. The average RMS errors for the medial, lateral, and total contact forces over the entire gait cycle and across all trials were 140 +/- 40 N, 115 +/- 32 N, and 183 +/- 45 N, respectively. Muscle coordination predicted by the model was also consistent with EMG measurements reported for normal walking. The combined experimental and modeling approach used in this study provides a quantitative framework for evaluating model predictions of muscle forces in human movement.