Musculoskeletal model-guided, customizable selection of shoulder and elbow muscles for a C5 SCI neuroprosthesis

Improving Musculoskeletal Model Scaling Using an Anatomical Atlas: The Importance of Gender and Anthropometric Similarity to Quantify Joint Reaction Forces

OBJECTIVE

The accuracy of a musculoskeletal model relies heavily on the implementation of the underlying anatomical dataset. Linear scaling of a generic model, despite being time and cost efficient, produces substantial errors as it does not account for gender differences and inter-individual anatomical variations. The hypothesis of this study is that linear scaling to a musculoskeletal model with gender and anthropometric similarity to the individual subject produces similar results to the ones that can be obtained from a subject-specific model.

METHODS

A lower limb musculoskeletal anatomical atlas was developed consisting of ten datasets derived from magnetic resonance imaging of healthy subjects and an additional generic dataset from the literature. Predicted muscle activation and joint reaction force were compared with electromyography and literature data. Regressions based on gender and anthropometry were used to identify the use of atlas.

RESULTS

Primary predictors of differences for the joint reaction force predictions were mass difference for the ankle (p < 0.001) and length difference for the knee and hip (p ≤ 0.017). Gender difference accounted for an additional 3% of the variance (p ≤ 0.039). Joint reaction force differences at the ankle, knee, and hip were reduced by between 50% and 67% (p = 0.005) when using a musculoskeletal model with the same gender and similar anthropometry in comparison with a generic model.

CONCLUSION

Linear scaling with gender and anthropometric similarity can improve joint reaction force predictions in comparison with a scaled generic model.

SIGNIFICANCE

The presented scaling approach and atlas can improve the fidelity and utility of musculoskeletal models for subject-specific applications.

A Muscle-Specific Rehabilitation Training Method Based on Muscle Activation and the Optimal Load Orientation Concept

Training based on muscle-oriented repetitive movements has been shown to be beneficial for the improvement of movement abilities in human limbs in relation to fitness, athletic training, and rehabilitation training. In this paper, a muscle-specific rehabilitation training method based on the optimal load orientation concept (OLOC) was proposed for patients whose motor neurons are injured, but whose muscles and tendons are intact, to implement high-efficiency resistance training for the shoulder muscles, which is one of the most complex joints in the human body. A three-dimensional musculoskeletal model of the human shoulder was used to predict muscle forces experienced during shoulder movements, in which muscles that contributed to shoulder motion were divided into 31 muscle bundles, and the Hill model was used to characterize the force-length properties of the muscle. According to the musculoskeletal model, muscle activation was calculated to represent the muscle force. Thus, training based on OLOC was proposed by maximizing the activation of a specific muscle under each posture of the training process. The analysis indicated that the muscle-specific rehabilitation training method based on the OLOC significantly improved the training efficiency for specific muscles. The method could also be used for trajectory planning, load magnitude planning, and evaluation of training effects.

Adaptive neuron-to-EMG decoder training for FES neuroprostheses

OBJECTIVE

We have previously demonstrated a brain-machine interface neuroprosthetic system that provided continuous control of functional electrical stimulation (FES) and restoration of grasp in a primate model of spinal cord injury (SCI). Predicting intended EMG directly from cortical recordings provides a flexible high-dimensional control signal for FES. However, no peripheral signal such as force or EMG is available for training EMG decoders in paralyzed individuals.

APPROACH

Here we present a method for training an EMG decoder in the absence of muscle activity recordings; the decoder relies on mapping behaviorally relevant cortical activity to the inferred EMG activity underlying an intended action. Monkeys were trained at a 2D isometric wrist force task to control a computer cursor by applying force in the flexion, extension, ulnar, and radial directions and execute a center-out task. We used a generic muscle force-to-endpoint force model based on muscle pulling directions to relate each target force to an optimal EMG pattern that attained the target force while minimizing overall muscle activity. We trained EMG decoders during the target hold periods using a gradient descent algorithm that compared EMG predictions to optimal EMG patterns.

MAIN RESULTS

We tested this method both offline and online. We quantified both the accuracy of offline force predictions and the ability of a monkey to use these real-time force predictions for closed-loop cursor control. We compared both offline and online results to those obtained with several other direct force decoders, including an optimal decoder computed from concurrently measured neural and force signals.

SIGNIFICANCE

This novel approach to training an adaptive EMG decoder could make a brain-control FES neuroprosthesis an effective tool to restore the hand function of paralyzed individuals. Clinical implementation would make use of individualized EMG-to-force models. Broad generalization could be achieved by including data from multiple grasping tasks in the training of the neuron-to-EMG decoder. Our approach would make it possible for persons with SCI to grasp objects with their own hands, using near-normal motor intent.

Modeling open-loop stability of a human arm driven by a functional electrical stimulation neuroprosthesis

Functional electrical stimulation (FES) can be used to restore movement control following paralysis. For complex multijoint systems, it is becoming increasingly apparent that closed-loop controllers are needed. Designing a closed-loop control system is easiest when the open-loop system is stable. In this study we developed a computational model to assess the open-loop stability of FES-control systems. We used the model to examine the open-loop stability of the human arm throughout its reachable workspace. For each simulated position of the hand we examined the stability of the arm, assuming that a minimal pattern of muscle activation was used to support the arm against gravity. Only muscles available to an existing FES user were considered. We found that with this reduced muscle set, the stability of the arm was severely compromised. We also demonstrated that muscle co-contraction can be an effective method to improve the stability for many postures.

Clinical applications of musculoskeletal modelling for the shoulder and upper limb

Musculoskeletal models have been developed to estimate internal loading on the human skeleton, which cannot directly be measured in vivo, from external measurements like kinematics and external forces. Such models of the shoulder and upper extremity have been used for a variety of purposes, ranging from understanding basic shoulder biomechanics to assisting in preoperative planning. In this review, we provide an overview of the most commonly used large-scale shoulder and upper extremity models and categorise the applications of these models according to the type of questions their users aimed to answer. We found that the most explored feature of a model is the possibility to predict the effect of a structural adaptation on functional outcome, for instance, to simulate a tendon transfer preoperatively. Recent studies have focused on minimising the mismatch in morphology between the model, often derived from cadaver studies, and the subject that is analysed. However, only a subset of the parameters that describe the model's geometry and, perhaps most importantly, the musculotendon properties can be obtained in vivo. Because most parameters are somehow interrelated, the others should be scaled to prevent inconsistencies in the model's structure, but it is not known exactly how. Although considerable effort is put into adding complexity to models, for example, by making them subject-specific, we have found little evidence of their superiority over current models. The current trend in development towards individualised, more complex models needs to be justified by demonstrating their ability to answer questions that cannot already be answered by existing models.

3D finite element models of shoulder muscles for computing lines of actions and moment arms

Accurate representation of musculoskeletal geometry is needed to characterise the function of shoulder muscles. Previous models of shoulder muscles have represented muscle geometry as a collection of line segments, making it difficult to account for the large attachment areas, muscle-muscle interactions and complex muscle fibre trajectories typical of shoulder muscles. To better represent shoulder muscle geometry, we developed 3D finite element models of the deltoid and rotator cuff muscles and used the models to examine muscle function. Muscle fibre paths within the muscles were approximated, and moment arms were calculated for two motions: thoracohumeral abduction and internal/external rotation. We found that muscle fibre moment arms varied substantially across each muscle. For example, supraspinatus is considered a weak external rotator, but the 3D model of supraspinatus showed that the anterior fibres provide substantial internal rotation while the posterior fibres act as external rotators. Including the effects of large attachment regions and 3D mechanical interactions of muscle fibres constrains muscle motion, generates more realistic muscle paths and allows deeper analysis of shoulder muscle function.

Electromyogram-based neural network control of transhumeral prostheses

Upper-limb amputation can cause a great deal of functional impairment for patients, particularly for those with amputation at or above the elbow. Our long-term objective is to improve functional outcomes for patients with amputation by integrating a fully implanted electromyographic (EMG) recording system with a wireless telemetry system that communicates with the patient's prosthesis. We believe that this should generate a scheme that will allow patients to robustly control multiple degrees of freedom simultaneously. The goal of this study is to evaluate the feasibility of predicting dynamic arm movements (both flexion/extension and pronation/supination) based on EMG signals from a set of muscles that would likely be intact in patients with transhumeral amputation. We recorded movement kinematics and EMG signals from seven muscles during a variety of movements with different complexities. Time-delayed artificial neural networks were then trained offline to predict the measured arm trajectories based on features extracted from the measured EMG signals. We evaluated the relative effectiveness of various muscle subsets. Predicted movement trajectories had average root-mean-square errors of approximately 15.7° and 24.9° and average R(2) values of approximately 0.81 and 0.46 for elbow flexion/extension and forearm pronation/supination, respectively.

The Influence of Modeling Separate Neuromuscular Compartments on the Force and Moment Generating Capacities of Muscles of the Feline Hindlimb

Functional electrical stimulation (FES) has the capacity to regenerate motion for individuals with spinal cord injuries. However, it is not straightforward to determine the stimulation parameters to generate a coordinated movement. Musculoskeletal models can provide a noninvasive simulation environment to estimate muscle force and activation timing sequences for a variety of tasks. Therefore, the purpose of this study was to develop a musculoskeletal model of the feline hindlimb for simulations to determine stimulation parameters for intrafascicular multielectrode stimulation (a method of FES). Additionally, we aimed to explore the differences in modeling neuromuscular compartments compared with representing these muscles as a single line of action. When comparing the modeled neuromuscular compartments of biceps femoris, sartorius, and semimembranosus to representations of these muscles as a single line of action, we observed that modeling the neuromuscular compartments of these three muscles generated different force and moment generating capacities when compared with single muscle representations. Differences as large as 4 N m (∼400% in biceps femoris) were computed between the summed moments of the neuromuscular compartments and the single muscle representations. Therefore, modeling neuromuscular compartments may be necessary to represent physiologically reasonable force and moment generating capacities of the feline hindlimb.

Musculoskeletal model of trunk and hips for development of seated-posture-control neuroprosthesis

The paralysis resulting from spinal cord injury severely limits voluntary seated-posture control and increases predisposition to a number of health risks. We developed and verified a musculoskeletal model of the hips and lumbar spine using published data. We then used the model to select the optimal muscles for-and evaluate the likely functional recovery benefit of-an 8-channel seated-posture-control neuroprosthesis based on functional electrical stimulation (FES). We found that the model-predicted optimal muscle set included the erector spinae, oblique abdominals, gluteus maximus, and iliopsoas. We mapped muscle excitations to seated trunk posture so that the required excitations at any posture could be approximated using a static map. Using the optimal muscle set, the model predicted a maximum stimulated range of motion of 49 degrees flexion, 9 degrees extension, and 16 degrees lateral bend. In the nominal upright posture, the modeled user could hold almost 15 kg with arms at sides and elbows bent. We discuss in this article the practicality of using FES with the oblique abdominals. A seated-posture-control neuroprosthesis would increase the user's bimanual work space and include several secondary benefits.

Stimulation stability and selectivity of chronically implanted multicontact nerve cuff electrodes in the human upper extremity

Nine spiral nerve cuff electrodes were implanted in two human subjects for up to three years with no adverse functional effects. The objective of this study was to look at the long term nerve and muscle response to stimulation through nerve cuff electrodes. The nerve conduction velocity remained within the clinically accepted range for the entire testing period. The stimulation thresholds stabilized after approximately 20 weeks. The variability in the activation over time was not different from muscle-based electrodes used in implanted functional electrical stimulation systems. Three electrodes had multiple, independent contacts to evaluate selective recruitment of muscles. A single muscle could be selectively activated from each electrode using single-contact stimulation and the selectivity was increased with the use of field steering techniques. The selectivity after three years was consistent with selectivity measured during the implant surgery. Nerve cuff electrodes are effective for chronic muscle activation and multichannel functional electrical stimulation in humans.

Feasibility of EMG-based neural network controller for an upper extremity neuroprosthesis

The overarching goal of this project is to provide shoulder and elbow function to individuals with C5/C6 spinal cord injury (SCI) using functional electrical stimulation (FES), increasing the functional outcomes currently provided by a hand neuroprosthesis. The specific goal of this study was to design a controller based on an artificial neural network (ANN) that extracts information from the activity of muscles that remain under voluntary control sufficient to predict appropriate stimulation levels for several paralyzed muscles in the upper extremity. The ANN was trained with activation data obtained from simulations using a musculoskeletal model of the arm that was modified to reflect C5 SCI and FES capabilities. Several arm movements were recorded from able-bodied subjects and these kinematics served as the inputs to inverse dynamic simulations that predicted muscle activation patterns corresponding to the movements recorded. A system identification procedure was used to identify an optimal reduced set of voluntary input muscles from the larger set that are typically under voluntary control in C5 SCI. These voluntary activations were used as the inputs to the ANN and muscles that are typically paralyzed in C5 SCI were the outputs to be predicted. The neural network controller was able to predict the needed FES paralyzed muscle activations from "voluntary" activations with less than a 3.6% RMS prediction error.