Skip to main content

Lower Extremity Amputee Dynamics: Simulating the Motion of an Above-Knee Amputee’s Stump by Means of a Novel EMG-Integrated 3D Musculoskeletal Forward-Dynamics Modelling Approach

Final Report Summary - LEAD (Lower Extremity Amputee Dynamics: Simulating the Motion of an Above-Knee Amputee’s Stump by Means of a Novel EMG-Integrated 3D Musculoskeletal Forward-Dynamics Modelling Approach)

The overall aim of LEAD was to develop new modelling philosophies and methodologies to achieve, as the first research group worldwide, forward-dynamics simulations of three-dimensional, continuum-mechanical musculoskeletal systems consisting of at least one agonist-antagonist muscle pair and to use these methods to achieve EMG-driven, subject-specific, 3D simulations of above-knee amputee dynamics. To achieve these ambitious goals, LEAD pursued three research objectives: enhancing existing skeletal muscle modelling techniques, developing new computational techniques enabling simulations of complex system models, and implementing novel experimental techniques to provide input for the simulations and for validation. Most of the research aims were fulfilled.
With respect to the modelling objectives, we developed new skeletal muscle constitutive laws, in particular, of the active behaviour for gaining a better understanding of a muscle’s activation dynamics. By starting a new collaboration with AvH Humboldt Laureate, Prof. Ponte Castaneda (UPenn, USA), we were able to replace existing state-of-the-art skeletal muscle models that rely on fitting experimental data to phenomenological constitutive laws by bottom-up approaches based on novel homogenisation techiques. This research is the basis for new ideas in achieving subject-specific skeletal muscle constitutive laws. Further, we developed new material descriptions describing muscle fatigue. In particular interpreting and analysing individual muscular activities while wearing a prosthesis posed to be challenging. To overcome these limitations and for gaining a better understanding of the activation dynamics, LEAD teamed up with Prof. Dario Farina and his ERC-AdG DEMOVE. In addition, one focus of LEAD was to develop novel computational methods to determine in silico EMG signals. This did not only lead to new insights into the active state of a skeletal muscle but also led to an improved understanding of the neuromuscular system – however only at the price of substantial compute times. To overcome this limitation, we developed as part of the computational objectives novel efficient computational methodologies based on model order reduction. After developing and implementing these novel methods for computing in silico EMG results, we modified these methods in such a way that they can be applied to solids, i.e. for predicting skeletal muscle force exertion and deformation. For computing in silico EMG data, we were already able to achieve great speedups. Addressing the experimental objectives led to the establishment of an entire new experimental-oriented and sustainable subgroup. We developed new methodologies for detecting the movement of the residual femur within an above knee amputee using A-mode ultrasound, new EMG-measurement techniques, novel methods for calibrating and reducing the error of motion capture, and novel image processing workflows for model generation and fibre tracking based on DTI-images.
Unifying the results achieved by addressing the different objectives led to the first forward-dynamics simulation of a 3D, continuum-mechanical musculoskeletal systems, we first achieved a two-muscle agonist-antagonist system – initially with one degree of freedom followed by new method development for solving overdetermined systems through optimisation. These newly developed methodologies were then used to achieve a patient-specific, above-knee musculoskeletal system simulation for investigating socket-stump interaction. In summary, by LEAD, we were able to establish an entirely new research direction, i.e. forward-dynamics simulation of musculoskeletal systems in which all components are modelled as continuum-mechanical, 3D objects and enabled me to consolidate my research group. Moreover, these newly developed methods will have in the future significant impact on new research projects and real-world applications in the fields of prosthetics, ergonomics, and safety.