Skip to main content

A deformable, multi-domain, numerical muscular neck model for orthopaedics and ergonomics applications

Final Report Summary - DEMU2NECK (A deformable, multi-domain, numerical muscular neck model for orthopaedics and ergonomics applications)

This 4-years project was set in the global framework of designing detailed numerical models of the human body for orthopaedics and ergonomics applications. In many of these applications, modelling both musculoskeletal interactions and the muscles’ activation appears to be key to investigate the links between posture, stability and mobility of the cervical spine. In an orthopaedic context, it is also necessary to account not only for differences between e.g. healthy, pathologic or ageing subjects, but also to model inter-individual, subject-specific differences. In this context, the aim of the project was to develop a subject-specific deformable, contractile numerical Finite Element (FE) muscular neck model.

To reach this aim, the research work concentrated on three main objectives:

1) The development and validation of a reference (generic) passive FE neck model.
The rationale behind this first objective was to build a FE model that would be representative of the geometry and of the passive biomechanical behaviour of an average male subject. Such model is state-of-the-art for the purpose of comparative numerical assessment of medical devices. A geometry of the head, neck and upper torso skeleton and muscles was reconstructed based on existing imaging data (Visible Human Project®) that was scaled and deformed to match average morphometric parameters from the literature. The geometry was then meshed for FE solving with the explicit FE solver LS-Dyna. Mechanical properties of the tissues were gathered from an extensive review of the literature. The project achieved its objective, by building a functional, detailed passive reference FE neck model that was then validated against an extensive set of boundary conditions. At this stage, this model may be used in the context of predictive and comparative studies to assess e.g. the influence of a medical device’s design on spinal loads.

2) The development and implementation of a contractile FE muscle model and its transfer to each muscle of the reference FE neck model to create a reference active FE neck model.
A single generic active muscle model was first built based on existing 1D FE elements already implemented in LS-Dyna, based on the Hill muscle model. The project achieved most of its objectives for this first part; in particular the generic muscle model (i) can be activated and controlled through a single activation parameter, (ii) yields acceptable resulting tensile forces (iii) models the swelling of the muscle during e.g. concentric contraction and (iv) allows accounting for the forces resulting from lateral stiffening effects associated with a contraction. Next, the transfer of this single model to all the muscles of the reference FE neck model constituted a scientific challenge that resulted in a partial achievement of the associated task. First, task-specific muscle activation patterns had to be estimated. For this purpose, an existing rigid-body of the neck was updated and used in the OpenSim software. With this software, an inverse dynamics approach followed by a static optimisation was used to estimate the highly redundant muscles forces/activations associated to a given set of head movements and loading conditions. These conditions include isometric tasks (e.g. resisting a load applied to the forehead) and isokinetic tasks (e.g. performing a full flexion of the neck), for which published experimental EMG (ElectroMyoGram) data could be used to validate the calculated activations. In a second time, these activations were then used as input to control the active muscles of the reference FE neck model while the same tasks where simulated. These simulations showed the feasibility of the modelling approach, but they also proved to be complex to control, each muscle’s behaviour being extremely dependant on both the definition of its passive properties and on its initial fibres/sarcomere’s length. This also highlighted that such direct input of activation levels to the muscles needed to be constrained to achieve the dynamic equilibrium, and that further development of this model may have to consider alternate strategies such as the coupling with a rigid-body model, which would allow optimising the muscles’ forces/activations at each step of the simulation. A third level of validation, which concerned the use of the postural MRI data from volunteers, was also expected to be performed in the grant proposal, but could not be achieved (Cf. Objective 3). Nevertheless, it is to be emphasised that these first results constitute an advancement beyond the state of the art in the field, as the active FE neck model allows estimating spinal loads that showed to be significantly influenced by the presence of the muscles, and that they thus clearly answered the objective of verifying the feasibility of an assessment of the influence of the muscles and of their functional capacity on the spinal loads. At this stage intervertebral loads calculated with our model could only be compared with published results obtained from rigid-body models, e.g. in the case of axial head tension. More validation will be required to validate these loads, which will prove to be complex owing to a lack of extensive relevant in-vivo data (such as can be found for e.g. the lower limb with research actions such as the grand challenge – Cf. Perspectives for the use of this active (as opposed to passive) reference model in a context of medical device design assessment are however extremely encouraging in the short-term thanks to the improvement in the bio-fidelity of the spinal loads representation that is expected to be achieved through the modelling of the active 3D muscles.

3) The development of a subject-specific modelling approach to create a subject-specific FE neck model.
The creation of a subject-specific FE neck model was based on the deformation of the reference FE neck model through the adaptation into a robust toolbox of existing deformation tools. An experimental protocol was first designed to collect subject-specific imaging data that could be used to (1) create the subject-specific models and (2) evaluate the models during simple postural tasks. Detailed in-vivo subject-specific imaging data was obtained by imaging the upper body of 10 volunteers within a closed 3 Tesla MRI scanner for the first purpose. The same volunteers were then imaged within an open 0.6 Tesla MRI scanner in a range of controlled head and neck postures corresponding to some of postural tasks already presented. A previously developed toolbox of deformation tools was then adapted in order to be able to directly deform the reference FE neck model to any subject-specific medical imaging data. The feasibility of this approach was tested by reconstructing the head and neck of 2 different subjects from the in-vivo MRI data obtained during the project and by simulating and comparing the spinal loads resulting from a passive flexion-extension movement for each subject. Different spinal loads were found between the subjects. Although it was not possible to validate these local results at this stage, these comparative results highlighted the possible influence of morphometric parameters on the behaviour and load sharing within the spine, and emphasise the strong potential for such subject-specific modelling approach.

Such model will be used to explore the influence of the neck’s muscular contribution to the dynamic balance of the cervical spine in the healthy, pathological or ageing subject, allowing a better evaluation of inter-individual differences (pathological or not). Including the 3D interactions and active component of the muscles is expected to improve the bio-fidelity of previous modelling approaches, in particular in terms of the local assessment of the distribution of the loads shared between the neck structures.

Contact details:
Organisation: Université Claude Bernard Lyon 1
Name: Bertrand FRECHEDE
Département de Mécanique, Bât. Oméga
43 Bd. du 11 Novembre 1918
69622 Villeurbanne Cedex, France
Phone: +33 4 72448093