Recurrent neural network to predict hyperelastic constitutive behaviors of the skeletal muscle

Fast soft-tissue deformations coupled with mixed reality toward the next-generation childbirth training simulator

High-quality gynecologist and midwife training is particularly relevant to limit medical complications and reduce maternal and fetal morbimortalities. Physical and virtual training simulators have been developed. However, physical simulators offer a simplified model and limited visualization of the childbirth process, while virtual simulators still lack a realistic interactive system and are generally limited to imposed predefined gestures. Objective performance assessment based on the simulation numerical outcomes is still not at hand. In the present work, we developed a virtual childbirth simulator based on the Mixed-Reality (MR) technology coupled with HyperMSM (Hyperelastic Mass-Spring Model) formulation for real-time soft-tissue deformations, providing intuitive user interaction with the virtual physical model and a quantitative assessment to enhance the trainee's gestures. Microsoft HoloLens 2 was used and the MR simulator was developed including a complete holographic obstetric model. A maternal pelvis system model of a pregnant woman (including the pelvis bone, the pelvic floor muscles, the birth canal, the uterus, and the fetus) was generated, and HyperMSM formulation was applied to simulate the soft tissue deformations. To induce realistic reactions to free gestures, the virtual replicas of the user's detected hands were introduced into the physical simulation and were associated with a contact model between the hands and the HyperMSM models. The gesture of pulling any part of the virtual models with two hands was also implemented. Two labor scenarios were implemented within the MR childbirth simulator: physiological labor and forceps-assisted labor. A scoring system for the performance assessment was included based on real-time biofeedback. As results, our developed MR simulation application was developed in real-time with a refresh rate of 30-50 FPS on the HoloLens device. HyperMSM model was validated using FE outcomes: high correlation coefficients of [0.97-0.99] and weighted root mean square relative errors of 9.8% and 8.3% were obtained for the soft tissue displacement and energy density respectively. Experimental tests showed that the implemented free-user interaction system allows to apply the correct maneuvers (in particular the "Viennese" maneuvers) during the labor process, and is capable to induce a truthful reaction of the model. Obtained results confirm also the possibility of using our simulation's outcomes to objectively evaluate the trainee's performance with a reduction of 39% for the perineal strain energy density and 5.6 mm for the vertical vaginal diameter when the "Viennese" technique is applied. This present study provides, for the first time, an interactive childbirth simulator with an MR immersive experience with direct free-hand interaction, real-time soft-tissue deformation feedback, and an objective performance assessment based on numerical outcomes. This offers a new perspective for enhancing next-generation training-based obstetric teaching. The used models of the maternal pelvic system and the fetus will be enhanced, and more delivery scenarios (e.g. instrumental delivery, breech delivery, shoulder dystocia) will be designed and integrated. The third stage of labor will be also investigated to include the delivery of the placenta, and the clamping and cutting of the umbilical cord.

An Inverse Method to Determine Mechanical Parameters of Porcine Vitreous Bodies Based on the Indentation Test

The vitreous body keeps the lens and retina in place and protects these tissues from physical insults. Existing studies have reported that the mechanical properties of vitreous body varied after liquefaction, suggesting mechanical properties could be effective parameters to identify vitreous liquefaction process. Thus, in this work, we aimed to propose a method to determine the mechanical properties of vitreous bodies. Fresh porcine eyes were divided into three groups, including the untreated group, the 24 h liquefaction group and the 48 h liquefaction group, which was injected collagenase and then kept for 24 h or 48 h. The indentation tests were carried out on the vitreous body in its natural location while the posterior segment of the eye was fixed in the container. A finite element model of a specimen undertaking indentation was constructed to simulate the indentation test with surface tension of vitreous body considered. Using the inverse method, the mechanical parameters of the vitreous body and the surface tension coefficient were determined. For the same parameter, values were highest in the untreated group, followed by the 24 h liquefaction group and the lowest in the 48 h liquefaction group. For C₁₀ in the neo-Hookean model, the significant differences were found between the untreated group and liquefaction groups. This work quantified vitreous body mechanical properties successfully using inverse method, which provides a new method for identifying vitreous liquefactions related studies.

Model of calf muscle tear during a simulated eccentric contraction, comparison between ex-vivo experiments and discrete element model

The tearing of the muscle-tendon complex (MTC) is one of the common sports-related injuries. A better understanding of the mechanisms of rupture and its location could help clinicians improve the way they manage the rehabilitation period of patients. A new numerical approach using the discrete element method (DEM) may be an appropriate approach, as it considers the architecture and the complex behavior of the MTC. The aims of this study were therefore: first, to model and investigate the mechanical elongation response of the MTC until rupture with muscular activation. Secondly, to compare results with experimental data, ex vivo tensile tests until rupture were done on human cadavers {triceps surae muscle + Achilles tendon}. Force/displacement curves and patterns of rupture were analyzed. A numerical model of the MTC was completed in DEM. In both numerical and experimental data, rupture appeared at the myotendinous junction (MTJ). Moreover, force/displacement curves and global rupture strain were in agreement between both studies. The order of magnitude of rupture force was close between numerical (858 N for passive rupture and 996 N-1032 N for rupture with muscular activation) and experimental tests (622 N ± 273 N) as for the displacement of the beginning of rupture (numerical: 28-29 mm, experimental: 31.9 mm ± 3.6 mm). These differences could be explained by choices of DEM model and mechanical properties of MTC's components or their rupture strain values. Here we show that he MTC was broken by fibers' delamination at the distal MTJ and by tendon disinsertion at the proximal MTJ in agreement with experimental data and literature.

OpenColab project: OpenSim in Google colaboratory to explore biomechanics on the web

OpenSim is an open-source biomechanical package with a variety of applications. It is available for many users with bindings in MATLAB, Python, and Java via its application programming interfaces (APIs). Although the developers described well the OpenSim installation on different operating systems (Windows, Mac, and Linux), it is time-consuming and complex since each operating system requires a different configuration. This project aims to demystify the development of neuro-musculoskeletal modeling in OpenSim with zero configuration on any operating system for installation (thus cross-platform), easy to share models while accessing free graphical processing units (GPUs) on a web-based platform of Google Colab. To achieve this, OpenColab was developed where OpenSim source code was used to build a Conda package that can be installed on the Google Colab with only one block of code in less than 7 min. To use OpenColab, one requires a connection to the internet and a Gmail account. Moreover, OpenColab accesses vast libraries of machine learning methods available within free Google products, e.g. TensorFlow. Next, we performed an inverse problem in biomechanics and compared OpenColab results with OpenSim graphical user interface (GUI) for validation. The outcomes of OpenColab and GUI matched well (r≥0.82). OpenColab takes advantage of the zero-configuration of cloud-based platforms, accesses GPUs, and enables users to share and reproduce modeling approaches for further validation, innovative online training, and research applications. Step-by-step installation processes and examples are available at: https://simtk.org/projects/opencolab.