Active leg prostheses can provide individuals with an amputation with motion support and thereby enable better participation in societal life. However, their versatility of use is still constrained, particularly with respect to volitional control by the user. Promising volitional control methods like employing muscle and joint model simulations to mimic human motor control are actively researched. However, experimental paradigms currently either include costly hardware setups or are limited to very simplistic virtual setups. For rapid control prototyping and facilitating the involvement of healthy participants, we propose an immersive virtual reality environment, focusing on ankle prostheses. Based on a theoretical discussion and a pilot study, we outline the potential of this method for agile development of lower limb volitional controllers.

Colonoscopy plays a pivotal role in early colorectal disease detection. It is hindered by procedural difficulties such as friction between colonoscope and colon wall as well as maneuverability issues. Expanding robots offer a promising alternative, aiming for a reduced risk of tissue damage and reduced forces between colononscope and tissue. In this paper, the usage of expanding locomotion for colonoscopy is investigated to address technical challenges. Preliminary tests on a custom-designed test rig demonstrate the potential of intermitted advancement and retraction modes to overcome challenges such as the difference in pathlength of robot body and working channel as well as the typical buckling during retraction of expanding robots. Future research will focus on refining control systems based on sensor feedback and evaluating additional parameters for clinical application.

We investigate the force-frequency relationship (FFR) representation of human ventricular cardiomyocyte models and point out shortcomings, motivated by discrepancies in whole-heart simulations at increased pacing rates. Utilizing the openCARP simulator, simulations across frequencies ranging from 1 Hz to 3 Hz were conducted. Experimental data on healthy human ventricular cardiomyocytes were collected and compared against simulated results. Results show deviations for all models, with Tomek et al. modeling time sensitive biomarkers the best. For example, the ratio of time to peak tension at 2 Hz and 1 Hz is around 85 % for experiments, 82 % for hybrid data, 95 % for Tomek et al., 98 % for O’Hara et al. and 138 % for ten Tusscher et al.. These discrepancies, highlight not only the need for careful selection of ionic models, but also the importance of refining ventricular cardiomyocyte models for advancing in-silico cardiac research.

Finite Element Models (FEM) for simulating various physical phenomena are indispensable in today's research. They are also applied in high-frequency (HF) surgical applications to simulate Joule heating and thermal necrosis formation in biological tissue. One challenge lies in choosing an appropriate mesh size, balancing between sufficiently accurate simulation results and a coarse mesh for fast computations. In this contribution, we investigate, using three mesh sizes (coarse, medium, and fine), the extent to which regression analysis of discrete simulation results of necrosis dimensions can compensate for mesh refinement in an FEM for HF soft coagulation. We conducted exponential regression analysis on discrete simulation results of the necrosis zone. It has been found that while it is generally possible, if the FEM simulation results already deviate significantly from the true solution, regression analysis cannot compensate for errors arising from the numerical method. Therefore, a coarse mesh with subsequent regression analysis can be utilized to determine necrosis propagation in depth, with a maximum error of only -3.4 % compared to the fine mesh within the considered interval. However, for the necrosis in the lateral direction, the maximum error is 44.3 %, which is too large. Even the mid-size mesh is not accurate enough in the lateral direction, necessitating the use of a fine mesh for the lateral direction.

Introduction

For over four decades, finite element models of the middle ear have been essential for better understanding hearing mechanics, aiding implant and hearing aid development, as well as diagnosing middle-ear diseases. However, existing models are mostly deterministic and neglect individual variations in middle-ear parameters. While material parameterization is straightforward, geometry parameterization like ossicle and ligament arrangement pose challenges in finite element models. This study introduces a stochastic model that adresses these challenges and comprehensively parameterizes middle-ear geometry and material properties.

Methods

The stochastic model, based on an existing finite element model, simplifies ossicles as rigid bodies and ligaments as beam elements. Only the tympanic membrane, tympanic cavity, and ear canal are meshed with finite elements. This approach allows straightforward parameterization of ossicular mass and moments of inertia, joint and ligament attachment points and orientations, and tympanic membrane shape using publicly available or published computed tomography data and material properties. Stochastic variables are defined based on this data and linked in an anatomically plausible parameter chain using principal component analyses and copulas. The model is calibrated using standard data, and global sensitivities of parameters are assessed using Sobol indices. A neural network for differentiating between normal ears and those with otosclerosis and disarticulation is trained using simulated data from the stochastic model and tested on existing literature data.

Results

The calibrated model accurately reproduces the mean and variance of middle-ear measurements like impedance, reflectance, stapes and umbo transfer function. Ligament and joint material parameters have a significant effect on the variance of these measurements, while variations in center of mass positions, for example, have less effect. The neural network trained on the simulated data shows promise for diagnostics, performing similarly to classifiers trained on measured immittance data.

Conclusion

The stochastic model combines finite element modeling benefits with straightforward geometry parameterization. Calibrating the model significantly reduces uncertainties of difficult-to-measure material properties compared to literatur values. This facilitates future model validation for further middle-ear pathologies and makes the model valuable for diagnosing rare pathologies with limited validation and training data.