Fused Ultrasound And Electromyography-Driven Neuromuscular Model To Improve Plantarflexion Moment Prediction Across Walking Speeds

Abstract

Background: Improving the prediction ability of a human-machine interface (HMI) is critical to accomplish a bio-inspired or model-based control strategy for rehabilitation interventions, which are of increased interest to assist limb function post neurological injuries. A fundamental role of the HMI is to accurately predict human intent by mapping signals from a mechanical sensor or surface electromyography (sEMG) sensor. These sensors are limited to measuring the resulting limb force or movement or the neural signal evoking the force. As the intermediate mapping in the HMI also depends on muscle contractility, a motivation exists to include architectural features of the muscle as surrogates of dynamic muscle movement, thus further improving the HMI's prediction accuracy. Objective: The purpose of this study is to investigate a non-invasive sEMG and ultrasound (US) imaging-driven Hill-type neuromuscular model (HNM) for net ankle joint plantarflexion moment prediction. We hypothesize that the fusion of signals from sEMG and US imaging results in a more accurate net plantarflexion moment prediction than sole sEMG or US imaging. Methods: Ten young non-disabled participants walked on a treadmill at speeds of 0.50, 0.75, 1.00, 1.25, and 1.50 m/s. The proposed HNM consists of two muscle-tendon units. The muscle activation for each unit was calculated as a weighted summation of the normalized sEMG signal and normalized muscle thickness signal from US imaging. The HNM calibration was performed under both single-speed mode and inter-speed mode, and then the calibrated HNM was validated across all walking speeds. Results: On average, the normalized moment prediction root mean square error was reduced by 14.58 % (p = 0.012) and 36.79 % (p < 0.001) with the proposed HNM when compared to sEMG-driven and US imaging-driven HNMs, respectively. Also, the calibrated models with data from the inter-speed mode were more robust than those from single-speed modes for the moment prediction.Conclusions: The proposed sEMG-US imaging-driven HNM can significantly improve the net plantarflexion moment prediction accuracy across multiple walking speeds. The findings imply that the proposed HNM can be potentially used in bio-inspired control strategies for rehabilitative devices due to its superior prediction.

Hill-type models of skeletal muscle and neuromuscular actuators: a systematic review

“…The performance of a muscle Hill-type model can be evaluated by comparing its predictions against experimental quantities obtained in situ or from complete experimental datasets made available in the literature (Perreault et al, 2003; Wakeling, James M. et al, 2021). The gold standard quantity for model validation is the time-course of individual fibre or muscle forces, which can be obtained or derived from dissected muscles of anesthetized animals as in some of the eligible studies, from non-invasive ultrasound measurements of fascicle or tendon length in humans (Lichtwark et al, 2007; Dick et al, 2017; Monte et al, 2020; Zhang et al, 2021; Zhang et al, 2022), or from specific protocols that minimize muscle co-activation during joint torque measurements (Hatze, 1981). Since the latter methods are relatively complex, most models are validated by comparing experimental and predicted joint torques, also often used for model calibration for the same activity, joint angles, and neural EMG-like signals, as discussed in the Results Section and detailed in SM12, despite limitations in validation accuracy as simulated and experimental quantities accumulate errors independent from muscle contraction mechanisms.…”

A deep learning method to predict ankle joint moment during walking at different speeds with ultrasound imaging: A framework for assistive devices control

“…In comparison to existing literature, it is fundamental to note that the results should be compared only with analyses considering the continuous prediction/estimation of joint moment/torque or joint angle by using neuromuscular signals, and the error should be normalized to the corresponding peak value. By using machine learning methods, it is common to see in the literature estimation/prediction error of down to 7–15% (Shi et al, 2008; Dick et al, 2017; Huang et al, 2017; Jahanandish et al, 2019; Zhou et al, 2019; Ma et al, 2020; Jahanandish et al, 2021; Wang et al, 2021; Yu et al, 2021; Zhang et al, 2022a,2022b). However, most of these studies consider using either sEMG or US signals for the upper limb motion or moment estimation/prediction, while a few of them consider the lower limb applications but with relatively simple tasks.…”

“…The human ankle plantarflexor muscles play an essential role in the lower limbs’ activities of daily living. For example, they provide primary mechanical propulsion force in both forward and upward directions during walking, which is referred to as the “push-off” during the late stance phase of the gait cycle (Huang et al, 2015b; Zhang et al, 2022b). Weakened function or dysfunction of human ankle plantarflexor muscles due to neurological disorders and injuries such as stroke, multiple sclerosis, and spinal cord injury, can cause significant impairment of normal walking function, like decreased plantarflexion propulsion force, inefficient ankle mechanical power production, and reduced walking efficiency (Nuckols et al, 2020).…”

“…The focus is to verify the robustness of the proposed approach when incorporating new data from various walking scenarios for the personalized prediction of net plantarflexion moment. Unlike features extraction methods in our previous studies (Zhang et al, 2020a,b; Zhang et al, 2021; Zhang et al, 2022a,b), this deep CNN approach does not require the extraction of conventional explicit features for establishing the mapping between skeletal muscle’s US imaging signals and human joint mechanical functions. The implicit features would significantly increase the dimension of the neuromuscular features, which is beneficial for bridging the mapping relationship without losing a large quantity of necessary information.…”