Exploring natural variation in tendon constitutive parameters via Bayesian data selection and mixed effects models

James Casey, Jessica Forsyth, Timothy Waite, Simon Cotter, Tom Shearer · 2024

Read Exploring natural variation in tendon constitutive parameters via Bayesian data selection and mixed effects models by James Casey, Jessica Forsyth, Timothy Waite, Simon Cotter, Tom Shearer (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-12983.txt.

Introduction

James Casey and co-authors model how tendon stiffness traits vary across a horse population, in a 2024 paper. Tendons differ between energy-storing and position-holding roles. Prior work fit stiffness parameters one experiment at a time, using a Bayesian method on pull tests of two horse tendons. The new study asks how those parameters spread across many horses. A mixed effects model estimates the population pattern.

Careful data choice matters because the model ignores tendon damage. A two-stage Bayesian selection method weights each experiment's data by quality instead of dropping data by hand. Results show the position-holding tendon is stiffer than the energy-storing one. Higher collagen share likely explains the gap. The authors give numeric estimates for the combined collagen stiffness product in each tendon. They frame the method as reusable for other soft tissues.

Catalog record

CREATORJames Casey, Jessica Forsyth, Timothy Waite, Simon Cotter, Tom Shearer
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE61 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-12983

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Exploring natural variation in tendon constitutive parameters via Bayesian data selection and mixed effects models James Casey, Jessica Forsyth, Timothy Waite, Simon Cotter, Tom Shearer Abstract Combining microstructural mechanical models with experimental data enhances our understanding of the mechanics of soft tissue, such as tendons. In previous work, a Bayesian framework was used to infer constitutive parameters from uniaxial stress-strain experiments on horse tendons, specifically the superficial digital flexor tendon (SDFT) and common digital extensor tendon (CDET), on a per-experiment basis. Here, we extend this analysis to investigate the natural variation of these parameters across a population of horses. Using a Bayesian mixed effects model, we infer population distributions of these parameters. Given that the chosen hyperelastic model does not account for tendon damage, careful data selection is necessary. Avoiding ad hoc methods, we introduce a hierarchical Bayesian data selection method. This two-stage approach selects data per experiment, and integrates data weightings into the Bayesian mixed effects model. Our results indicate that the CDET is stiffer than the SDFT, likely due to a higher collagen volume fraction. The modes of the parameter distributions yield estimates of the product of the collagen volume fraction and Young’s modulus as 811.5 MPa for the SDFT and 1430.2 MPa for the CDET. This suggests that positional tendons have stiffer collagen fibrils and/or higher collagen volume density than energy-storing tendons. ID JC 0000-0001-9853-3566, JF 0000-0002-5839-9160, TW 0000-0002-8868-3146, SC 0000-0001-5974-7393, TS 0000-0001-7536-5547 Keywords— Tendon, Data Selection, Mixed Effects, Bayesian Inference, Nonlinear Elastic 1 Introduction Accurate description of the nonlinear mechanical behaviour of collagen is important for understanding and predicting the properties of a wide range of soft tissues, including arterial walls [ 1 , 2 ] , skin [ 3 , 4 ] , and tendons [ 5 , 6 , 7 ] . This knowledge is vital for designing artificial tissues for grafts and surgical interventions [ 8 ] . In this paper, we focus on a model for tendons; however, the underlying approach could be extended to other soft tissues. Tendons are fibrous tissues that connect, and transfer forces between, muscles and bones [ 5 , 9 ] . Their complex microstructure gives rise to anisotropy and non-linear stress-strain profiles [ 10 , 11 , 12 ] . An example of a typical stress-strain profile for a tendon undergoing a uniaxial stretch along its longitudinal axis is given in Figure 1 . Tendons consist of collagen fibrils of varying lengths that assume a crimped waveform within collagen fibres. The fibres are embedded within a non-collagenous matrix (NCM), and reinforce the tendon along a preferred axis [ 5 , 13 ] , which causes anisotropy. A schematic representation of a tendon’s microstructure is shown in Figure 2 . As the tendon is stretched, the fibrils begin to straighten and contribute to the stress response. Due to the fibrils having varying lengths, their recruitment is gradual, which results in the nonlinearity of the stress profile. Figure 1: Idealised stress profile of a tendon. The four regions are: I) Toe, II) Heel, III) Linear, IV) Damage. In reality, damage may begin occurring in region III, or even region II, due to the shorter fibrils breaking. Figure 2: Microstructure of a typical tendon. Bundles of fibrils form fibres, which are embedded within fascicles, which make up the tendon. The NCM is the non-collagenous portion of the tendon. Adapted from Jull et al. [ 14 ] . The characteristic shape of the stress-strain graph has four regions: I) Toe II) Heel, III) Linear, and IV) Damage, with each region corresponding to a different physical phenomenon. The toe region corresponds to the fibrils being slack, resulting in a stress profile that is approximately linear, as governed by the NCM. In the heel region, fibrils are gradually becoming taut and contributing to the stress response, resulting in a non-linear profile. In the linear region, all of the fibres have been recruited and the stress profile is primarily governed by the fibril stress response. In the final region, failure occurs, the tendon is damaged, fibrils begin to break gradually and the tendon no longer deforms elastically. In Figure 1 , the stress response is idealised. In reality, fibrils may begin breaking in earlier regions (III or even II), resulting in data in region III being recorded which is not consistent with models which neglect damage. Typically, the data is manually trimmed to a specified limit [ 15 ] . This approach tends to include a large portion of the linear region (region III), and fitting models to this data can lead to inaccurate estimates of various parameters in the presence of early fibril damage. For this reason, when fitting models to tendon stress-strain data, it is beneficial to use a more sophisticated data selection method to include as much data as possible that is valid under the assumptions of the mechanical model used, whilst down-weighting the contribution of data points that do not satisfy the model’s assumptions [ 16 ] . Models for the stress response of tendons often describe the deformation in terms of the physical parameters of their constituents, such as the Young’s modulus and shear modulus. These models are known as microstructural models, and have been used by several authors for various soft tissues [ 5 , 17 ] . One of the advantages of microstructural modelling is that the model parameters have a direct physical meaning, and can be measured experimentally. However, a problem that arises is that a wide range of values is often reported for the same parameter. The Young’s modulus of Type 1 collagen fibrils, for example, has been reported to be as low as 32 MPa [ 18 ] , and as high as 2.8 GPa [ 19 ] . This spread of reported values could be due to several factors, such as differences in the ages of the samples, inter-species variation, model misspecification within the inference, or even that the parameters may be poorly identifiable from the data. Due to this, it is often not entirely clear what the credible range for a given parameter value is, nor which values are most likely. Quantification of this spread is key to understanding the natural behaviour of soft tissues and their constituents. In this paper, we tackle the problem of intra -species variation. That is, within a given species, individuals can yield varying values for constitutive parameters due to naturally occurring heterogeneity within the population. For context, the individuals in this study are horse tendons, and the species are the SDFT and CDET types. To tackle the uncertainty in the parameter values, we employ a Bayesian approach. This provides a rigorous framework with which to combine prior knowledge, mechanical models, and data. Typically, inference is performed on data from a single individual from the population at a time, as in previous work [ 13 ] . In this paper, we instead use data from multiple individuals to infer the population-level variation in the constitutive parameters of a microstructural, nonlinear elastic tendon model, through a statistical mixed effects model. Since fibril damage occurs at different points for different individuals, we are prompted to use a more sophisticated data selection technique than uniform trimming of all data

Read the complete text