Search

Neuromusculoskeletal (NMS) models enable non-invasive estimation of clinically important internal biomechanics. A critical part of NMS modelling is the estimation of musculotendon kinematics, which comprise musculotendon unit lengths, moment arms, and lines of action. Musculotendon kinematics, which are partially dependent on joint angles, define the non-linear mapping of muscle forces to joint moments and contact forces. Currently, real-time computation of musculotendon kinematics requires creation of a per-individual surrogate model. The computational speed and accuracy of these surrogates degrade with increasing number of coordinates. We developed a feed-forward neural network that completely encodes musculotendon kinematics of a target model across a wide anthropometric range, enabling accurate real-time estimates of musculotendon kinematics without need for a priori creation of a per-individual surrogate model. Compared to reference, the neural network had median normalized errors ~0.1% for musculotendon lengths, <0.4% for moment arms, and <0.10° for line of action orientations. The neural network was employed within an electromyogram-informed NMS model to calculate hip contact forces, demonstrating little difference (normalized root mean square error $1.23\pm 0.15$ %) compared to using reference musculotendon kinematics. Finally, execution time was <0.04 ms per frame and constant for increasing number of model coordinates. Our approach to musculoskeletal kinematics may facilitate deployment of complex real-time NMS modelling in computer vision or wearable sensors applications to realize biomechanics monitoring, rehabilitation, and disease management outside the research laboratory. [ABSTRACT FROM AUTHOR]

Remodeling of the Achilles tendon (AT) is partly driven by its mechanical environment. AT force can be estimated with neuromusculoskeletal (NMSK) modeling; however, the complex experimental setup required to perform the analyses confines use to the laboratory. We developed task-specific long short-term memory (LSTM) neural networks that employ markerless video data to predict the AT force during walking, running, countermovement jump, single-leg landing, and single-leg heel rise. The task-specific LSTM models were trained on pose estimation keypoints and corresponding AT force data from 16 subjects, calculated via an established NMSK modeling pipeline, and cross-validated using a leave-one-subject-out approach. As proof-of-concept, new motion data of one participant was collected with two smartphones and used to predict AT forces. The task-specific LSTM models predicted the time-series AT force using synthesized pose estimation data with root mean square error (RMSE) $\le 526$ N, normalized RMSE (nRMSE) $\le 0.21$ , R $^{{2}} \ge 0.81$. Walking task resulted the most accurate with RMSE $= 189\pm 62$ N; nRMSE $= 0.11\pm 0.03$ , R $^{{2}}= 0.92\pm 0.04$. AT force predicted with smartphones video data was physiologically plausible, agreeing in timing and magnitude with established force profiles. This study demonstrated the feasibility of using low-cost solutions to deploy complex biomechanical analyses outside the laboratory. [ABSTRACT FROM AUTHOR]

The Achilles tendon (AT) is sensitive to mechanical loading, with appropriate strain improving tissue mechanical and material properties. Estimating free AT strain is currently possible through personalized neuromusculoskeletal (NMSK) modeling; however, this approach is time-consuming and requires extensive laboratory data. To enable in-field assessments, we developed an artificial intelligence (AI) workflow to predict free AT strain during running from motion capture data. Ten keypoints commonly used in pose estimation algorithms (e.g., OpenPose) were synthesized from motion capture data and noise was added to represent real-world data obtained using video cameras. Two AI workflows were compared: (1) a Long Short-Term Memory (LSTM) neural network that predicted free AT strain directly (called LSTM only workflow); and (2) an LSTM neural network that predicted AT force which was subsequently converted to free AT strain using a personalized force-strain curve (called LSTM+ workflow). AI models were trained and evaluated using estimates of free AT strain obtained from a validated NMSK model with personalized AT force-strain curve. The effect of using different input features (position, velocity, and acceleration of keypoints, height and mass) on free AT strain predictions was also assessed. The LSTM+ workflow significantly improved the predictions of free AT strain compared to the LSTM only workflow (p < 0.001). The best free AT strain predictions were obtained using positions and velocities of keypoints as well as the height and mass of the participants as input, with average time-series root mean square error (RMSE) of 1.72±0.95% strain and r2 of 0.92±0.10, and peak strain RMSE of 2.20% and r2 of 0.54. In conclusion, we showed feasibility of predicting accurate free AT strain during running using low fidelity pose estimation data. [ABSTRACT FROM AUTHOR]