Find research datasets worth reusing

opencc-by-4.0Feb 2020View details →

Data for: Geometrical parameters for musculoskeletal modeling of hand

<p>Dataset to be linked with not yet published manuscript "Geometrical parameters for<br> musculoskeletal modeling of hand"</p> <p>In musculoskeletal modelling, parameters identification, such as the exact position and trajectories of muscle attachments, is a crucial issue. The main goal of this study was to calculate the position, attachment dimensions and cross section areas of twenty-five extrinsic and intrinsic hand muscle complexes. We integrated measurements taken from cadaveric preparations, magnetic resonance imaging and mathematical theory. Sixteen cadaveric preparations were dissected to draw up the anatomical maps including the position of muscle attachments, dimensions, shapes, cross section areas and variations. The magnetic resonance imaging of cadaveric upper extremity was performed to reconstruct the geometry of all bones and hand muscles. Using these outcomes, the muscle attachments and cross section areas were extracted and verified using the obtained morphological and morphometric analysis. The exact trajectories of muscle lines of action were computed using the modified weighted k-means method and Hungary algorithm. This work introduces a new approach to acquiring musculoskeletal modelling data in general and contributes extensive dataset to the hand musculature modelling in particular.</p> <p><strong>Acknowledgments</strong></p> <p>This work was supported by the project n. 182 “Obstetrics 2.0 - Virtual models for the prevention of injuries during childbirth” realised within the frame of the Program INTERREG V-A: Cross- border cooperation between the Czech Republic and the Federal State of Germany Bavaria, Aim European Cross-border cooperation 2014-2020. The realisation is supported by financial means of the European Regional Development Fund (85 % of the costs) and the state budget of the Czech Republic (5 %). KI is part-funded by project No. CZ.02.1.01/0.0/0.0/16_019/0000787 “Fighting INfectious Diseases“, awarded by the Ministry of Education, Youth and Sports of the Czech Republic, financed from The European Regional Development Fund.</p>

FIGURE 2 in Achieving kinematic identity across shape diversity in musculoskeletal modeling

FIGURE 2. The top images show an overlay of the reduced-asymmetry australopithecine pelvis (beige) with the ADL australopithecine pelvis (dark green). The bottom color-coded distance map pelvis show the distance between the reduced-asymmetry australopithecine pelvis with the ADL australopithecine pelvis.

FIGURE 4 in Achieving kinematic identity across shape diversity in musculoskeletal modeling

FIGURE 4. AnyBody australopithecine musculoskeletal model without (left) and with (right) muscle model visualization.

FIGURE 1 in Achieving kinematic identity across shape diversity in musculoskeletal modeling

FIGURE 1. The flowchart shows the major steps required to build the ADL australopithecine model. In the blue boxes, the ADL human model is driven with the Schreiber and Moissenet (2019) human locomotion data. From these ADL human simulations, the dimension of the pelvis and femur can be extracted as well as model motion profiles used at later stages of the process (Figure 5). The gray boxes show the major steps in transforming (TPS-based morphing) the ADL human pelvis to match the australopithecine morphology (A.L. 288-1 reduced-asymmetry pelvis; Australopithecus afarensis), thus creating the ADL australopithecine pelvis. The green boxes show the steps necessary to create the ADL australopithecine (hybrid) femur from the ADL human femur.

FIGURE 5 in Achieving kinematic identity across shape diversity in musculoskeletal modeling

FIGURE 5. This flowchart shows the major steps required to generate the C3D motion file to drive the walking simulations with an australopithecine hip. Blue, light blue, and blue/grey and blue/green dashed boxes are the same boxes from Figure 1. The original ADL human model (blue box) is morphed based on the australopithecine pelvis (blue/grey dashed box) and femur (blue/green dashed box) to create the ADL australopithecine model (orange box). The results from the human walking simulation (blue box) are combined with the L5-sacral offset translation (light blue box) to generate new "experimental marker data" that are combined with the original ground reaction force data from Schreiber and Moissenet (2019) (purple box). The ADL australopithecine model and new motion data are then used to drive the simulations of walking with an australopithecine hip.

FIGURE 6 in Achieving kinematic identity across shape diversity in musculoskeletal modeling

FIGURE 6. Motion of the pelvis and lower limb joints for one individual walking simulation with both human (red lines) and australopithecine (black circles) shaped hips. A. Pelvic rotation (transverse plane), tilt (sagittal plane) and drop (coronal plane). B. Hip flexion-extension, abduction-adduction, and internal-external rotation. C. Knee flexion-extension, ankle dorsi-plantar flexion, subtalar eversion-inversion.

opencc-by-4.0Aug 2024View details →

Acquiring musculoskeletal skills with curriculum-based reinforcement learning - model weights

<p><strong>Acquiring musculoskeletal skills with curriculum-based reinforcement learning, Neuron 2024</strong></p> <p>Here we provide the weights of the neural network policies used for the analysis presented in our article.</p> <p>The archives whose names start with a number (01 - 32) correspond to the 32 curriculum steps to train the Baoding Balls policy which ranked first at the MyoChallenge 2022. The code used for the training and which can be used to test the policies can be found at https://github.com/amathislab/myochallenge.</p> <p>The archives <em>hand_pose, hand_reach, pen </em>and <em>reorient</em> correspond to the other policies used in the article. They were developed in the paper <em>Latent exploration for reinforcement learning</em>, Chiappa et al., NeurIPS 2023. They can be loaded and tested with the code at https://github.com/amathislab/lattice.</p> <p>The archive <em>datasets</em> includes three subfolders: <em>rollouts, umap</em> and <em>csi</em>.</p> <ul> <li>The files in <em>rollouts </em>are the datasets of transitions resulting from the interaction between a policy and the environment. </li> <li>The files in <em>umap</em> are the pre-computed projections of specific subsets fo the datasets included in <em>rollouts</em> using UMAP.</li> <li>The files in <em>csi</em> report the performance of the policies described in our paper when applying Control Subspace Inactivation (CSI).</li> </ul> <p>These datasets are necessary to run the notebooks to reproduce the paper's figures and main results, with the code at https://github.com/amathislab/MyoChallengeAnalysis</p> <p>If you find these weights useful, please cite:</p> <div> <div>@article{chiappa2024acquiring,<br>title = {Acquiring musculoskeletal skills with curriculum-based reinforcement learning},<br>journal = {Neuron},<br>volume = {112},<br>number = {23},<br>pages = {3969-3983.e5},<br>year = {2024},<br>issn = {0896-6273},<br>doi = {https://doi.org/10.1016/j.neuron.2024.09.002},<br>url = {https://www.sciencedirect.com/science/article/pii/S0896627324006500},<br>author = {Alberto Silvio Chiappa and Pablo Tano and Nisheet Patel and Abigaïl Ingster and Alexandre Pouget and Alexander Mathis},<br>keywords = {motor control, motor learning, reinforcement learning, curriculum learning, motor skills, musculoskeletal control, muscle synergies},<br>}</div> <div> </div> <div>@article{chiappa2024latent,</div> <div>title={Latent exploration for reinforcement learning},</div> <div>author={Chiappa, Alberto Silvio and Marin Vargas, Alessandro and Huang, Ann and Mathis, Alexander},</div> <div>journal={Advances in Neural Information Processing Systems},</div> <div>volume={36},</div> <div>year={2024}</div> <div>}</div> </div>

opencc-zeroSep 2020View details →

How to build a dinosaur: musculoskeletal modelling and simulation of locomotor biomechanics in extinct animals

<p>The intersection of paleontology and biomechanics can be reciprocally illuminating, helping to improve paleobiological knowledge of extinct species and furthering our understanding of the generality of biomechanical principles derived from study of extant species. However, working with data gleaned primarily from the fossil record has its challenges. Building on decades of prior research, we outline and critically discuss a complete workflow for biomechanical analysis of extinct species, using locomotor biomechanics in the Triassic theropod dinosaur <em>Coelophysis </em>as a case study. We progress from the digital capture of fossil bone morphology to creating rigged skeletal models, to reconstructing musculature and soft tissue volumes, to the development of computational musculoskeletal models, and finally to the execution of biomechanical simulations. Using a three-dimensional musculoskeletal model comprising 33 muscles, a static inverse simulation of the mid-stance of running shows that <em>Coelophysis </em>probably used more upright (extended) hindlimb postures, and was likely capable of withstanding a vertical ground reaction force of magnitude more than 2.5 times body weight. We identify muscle force-generating capacity as a key source of uncertainty in the simulations, highlighting the need for more refined methods of estimating intrinsic muscle parameters such as fibre length. Our approach emphasizes the explicit application of quantitative techniques and physics-based principles, which helps maximize results robustness and reproducibility. Although we focus on one specific taxon and question, many of the techniques and philosophies explored here have much generality to them, so they can be applied in biomechanical investigation of other extinct organisms.</p>

zenodo36/100

FIGURE 3 in Achieving kinematic identity across shape diversity in musculoskeletal modeling

FIGURE 3. Control landmarks for the AnyBody TPS-morphing on the ADL human femur and pelvis.

opencc-by-4.0Apr 2022View details →

zenodo36/100

A Differentiable Dynamic Model for Musculoskeletal Simulation and Exoskeleton Control

<p>An exoskeleton, a wearable device, was designed based on the user's physical and cognitive interactions. The control of the exoskeleton used biomedical signals reflecting user intention as input and its algorithm calculated an output to make the movement smooth. However, the process of transforming the input of biomedical signals, such as electromyography (EMG), into the output of adjusting the torque and angle of the exoskeleton is limited by a finite time lag and precision of trajectory prediction, which result in a mismatch between subject and exoskeleton. Here we propose an EMG-based single-joint exoskeleton system, merging a differentiable continuous system with a dynamic musculoskeletal model. The parameters of each muscle contraction were calculated and applied to the rigid exoskeleton system to predict the precise trajectory. The results revealed accurate torque and angle prediction for the knee exoskeleton and good performance of assistance during movement. Our method outperformed other models by rate of convergence and execution time. In conclusion, a differentiable continuous system merged with a dynamic musculoskeletal model supported effective and accurate performance of an exoskeleton controlled by EMG signals.</p>

opencc-zeroSep 2023View details →

Simulated cycling data set and musculoskeletal models

<p><span>This study used musculoskeletal modelling to explore the relationship between cycling conditions (power output and cadence) and muscle activation and metabolic power. We hypothesized that the cadence that minimized the simulated average active muscle volume would be higher than that which minimized the simulated metabolic power. We validated the simulation by comparing predicted muscle activation and fascicle velocities with experimental electromyography and ultrasound images. We found strong correlations for averaged muscle activations and moderate to good correlations for fascicle dynamics. These correlations tended to weaken when analyzed at the individual participant level. Our study revealed a curvilinear relationship between average active muscle volume and cadence, with the minimum active volume being aligned to the self-selected cadence. The simulated metabolic power was consistent with previous results and was minimized at lower cadences than that which minimized active muscle volume across power outputs. Whilst there are some limitations to the musculoskeletal modelling approach, the findings suggest that minimizing active muscle volume may be a more important factor than minimizing metabolic power for self-selected cycling cadence preferences. Further research is warranted to explore the potential of an active muscle volume-based objective function for control schemes across a wider range of cycling conditions.</span></p>

publicSep 2020View details →

How to build a dinosaur: musculoskeletal modelling and simulation of locomotor biomechanics in extinct animals

Open the record for dataset details and reuse information.