journal of the mechanical behavior of biomedical materials 142 (2023) 105856

Contents lists available at ScienceDirect

Journal of the Mechanical Behavior of Biomedical Materials

journal homepage: www.elsevier.com/locate/jmbbm

A finite element model of contusion spinal cord injury in rodents

Roman Frantsuzov a, 1, Subrata Mondal a, 1, Ciara M. Walsh b, c, James P. Reynolds b, c,
Dearbhaile Dooley b, c, David B. MacManus a, d, e, f, *
a
School of Mechanical & Manufacturing Engineering, Dublin City University, Dublin, Ireland
b
School of Medicine, University College Dublin, Dublin, Ireland
c
Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Ireland
d
MEDeng Research Centre, Dublin City University, Dublin, Ireland
e
Biodesign Europe, Dublin City University, Dublin, Ireland
f
School of Mechanical & Materials Engineering, University College Dublin, Dublin, Ireland

A R T I C L E I N F O A B S T R A C T

Keywords: Traumatic spinal cord injuries result from high impact forces acting on the spine and are proceeded by an
Spinal cord injury extensive secondary inflammatory response resulting in motor, sensory, and autonomic dysfunction. Experi­
Grey matter mental in vivo traumatic spinal cord injuries in rodents using a contusion model have been extremely useful in
White matter
elucidating the underlying pathophysiology of these injuries. However, the relationship between the patho­
Meninges
Trauma
physiology and the biomechanical factors is still not well understood. Therefore, the aim of this research is to
provide a comprehensive analysis of the biomechanics of traumatic spinal cord injury in a rat contusion model.
This is achieved through the development and validation of a finite element model of the thoracic rat spinal cord
and subsequently simulating controlled cortical impact-induced traumatic spinal cord injury. The effects of
impactor velocity, depth, and geometry on the resulting stresses and strains within the spinal cord are investi­
gated. Our results show that increasing impactor depth results in larger stresses and strains within the spinal cord
tissue as expected. Further, for the first time ever our results show that impactor geometry (spherical versus
cylindrical) plays an important role in the distribution and magnitude of stresses and strains within the cord.
Therefore, finite element modelling can be a powerful tool used to predict stresses and strains that occur in spinal
cord tissue during trauma.

1. Introduction lumbar (11%) regions (Alizadeh et al., 2019).

SCIs lie on a continuum of severity from transient to permanent loss
The global incidence of traumatic spinal cord injury (SCI) is 23 cases of function, varying degrees of paralysis, and death. These injuries also
per one million (Jazayeri et al., 2015) resulting in substantial public often result in multiple organ system dysfunction due to disruption of
health and socio-economic problems (Bárbara-Bataller et al., 2018). In the autonomic nervous system which can subsequently lead to auto­
Ireland, major trauma-related injury represents the third leading cause nomic dysreflexia, respiratory, bowel and bladder dysfunction, and
of death (Trauma Steering Group, 2018) with the SCI incidence rate at accelerated osteoporosis (Bloom et al., 2020). However, to date, there
195.4 cases per one million (Jazayeri et al., 2015), accounting for 17% remains no viable treatment options for improving the neurological
of all major trauma cases (National Office of Clinical Audit, 2020). In the outcomes of people with SCI (Alizadeh et al., 2019; Bloom et al., 2020).
United States there are 17,730 new cases of SCI annually, and it is SCI is intrinsically a mechanical injury resulting from high impact
estimated that up to 2.5 million people are living with chronic SCI forces acting on the spinal cord. Despite this, the relationship between
(Bloom et al., 2020). The most common causes of SCI are motor vehicle SCI biomechanics and pathophysiology is still not well understood.
accidents and falls, with a mean age of 43 years at the time of injury Preclinical rodent central nervous system (CNS) trauma models, such as
(Bloom et al., 2020). SCI occurs most often at the cervical level of the the controlled cortical impact (CCI) model, have proved extremely
spinal cord (50% of all injuries), followed by the thoracic (35%) and useful in recent years as they are cost-effective, easily available and can

* Corresponding author. School of Mechanical & Materials Engineering, University College Dublin, Belfield, Dublin, Ireland.
E-mail address: [email protected] (D.B. MacManus).
1
R. Frantsuzov and S. Mondal contributed equally to this work.

https://doi.org/10.1016/j.jmbbm.2023.105856
Received 17 January 2023; Received in revised form 2 April 2023; Accepted 12 April 2023
Available online 17 April 2023
1751-6161/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
R. Frantsuzov et al. Journal of the Mechanical Behavior of Biomedical Materials 142 (2023) 105856

help determine the underlying biomechanics and pathophysiology (FE) model of the rat thoracic spinal cord and the results from simulating
(Sharif-Alhoseini et al., 2017; Cheriyan et al., 2014). In the context of preclinical contusion CCI. Specifically, the effects of the impactor ve­
SCI, a CCI induces a transient acute injury onto the spinal cord using a locity, depth, and geometry on the resulting stresses and strains are
weight or an impactor device. This results in an injury response similar presented and discussed in the context of experimental design.
to that seen in humans, including the formation of cystic cavities and an
absence of spontaneous neural regeneration (Kjell and Olson, 2016; 2. Methods
Alizadeh et al., 2019; Bloom et al., 2020). The rat CCI model in partic­
ular, offers a clinically relevant platform for studying the biomechanics 2.1. Ethics and use of animals
and pathophysiology of SCI and can recapitulate the pathophysiology
seen in injuries resulting from motor vehicle accidents, the most com­ All animal experiments were conducted in line with national legis­
mon clinical presentation of SCI (Sharif-Alhoseini et al., 2017; Bloom lation in accordance with European Commission Directive (2010/63/
et al., 2020). However, there is considerable experimental variability EU), as overseen by the Health Products Regulatory Authority of Ireland
which currently exists for in vivo SCI experimental models (Reier et al., and the Animal Research Ethics committee of University College Dublin.
2012), resulting in different levels of injury severity and outcomes be­ Male and female Wistar rats (p56-p98) were maintained in controlled
tween laboratories. This variability is primarily due to the use of environments on a 12hr light/dark cycle, with food and water provided
different impactor geometries, velocities, injury classification, injury ab libitum. As part of ongoing animal studies on contusion SCI, fixed
region, and specimen age and sex. tissue cryosections were collected, with one transverse section selected
Detailed 3D finite element spinal cord models (FESMs) are another and assessed to help design the FE models, as described below.
extremely useful tool for studying the biomechanics of SCI by geomet­
rically and mechanically representing the biological tissues of the spinal
2.2. Finite element model development
cord in computer simulations. Many laboratories have created detailed
3D FESMs to investigate the biomechanics of SCI (Maikos et al., 2008a;
A FE model of the rat spinal cord and contusion model impactor were
Greaves et al., 2008; Russel et al., 2012; Yan et al., 2012; Khuyagbaatar
developed for use in ANSYS FE software package (ANSYS Inc., Penn­
et al., 2016; Stoner et al., 2020). Previously, a positive correlation be­
sylvania, USA). The 3D geometry of the rat spinal cord and impactors
tween strain and axon permeability has been established following
(Fig. 1) were developed in SolidWorks (2021) (DS SolidWorks Corp.,
experimental SCI (Russel et al., 2012). Further, the relationship between
Concord MA, USA) and using an in-house MATLAB (MathWorks, MA,
maximum principal strain and tissue damage, and open-field behav­
United States) code to automatically identify the boundaries of the grey
ioural score have been described (Lam et al., 2014) demonstrating the
(GM) and white matter (WM) regions from a transverse cryosection (10
important role biomechanics plays in SCI. To the best of the authors
μm thickness) taken from thoracic spinal cord (Fig. 1A). The cryosection
knowledge, only maximum principal strain has been correlated with the
was taken approximately 12 mm rostral of a T8 laminectomy and
pathophysiology of SCI. However, it was previously shown in brain
contusion impact site in a female Wistar rat, 4 weeks after injury. The
tissue that the strains, strain rates, and stresses that arise during impact
cryosection was taken at approximately 12 mm rostral of the impact site
correlate well with the pathology of traumatic brain injury (Khosroshahi
to ensure there was no macroscopic damage to the tissue from the
et al., 2021; Ghajari, 2017). Such studies highlight the potential of
impact resulting in a change in spinal cord morphology. Immunolabel­
FESMs to elucidate the relationship between the biomechanics and
ling βIII Tubulin was performed to identify spinal cord neuronal soma
pathophysiology of SCI.
and determine the outline of grey and white matter regions. The
Here, we present the development and validation of a finite element
meninges were not suitably identified using the automated in-house

Fig. 1. Overview of the finite element model development. The transverse cryosection (immunolabelled for βIII-Tubulin) used to develop the spinal cord geometry is
shown in (A), the 3D finite element model regions are shown in (B), the model boundary conditions in (C) and a 3D isometric view of the model in (D).

2
R. Frantsuzov et al. Journal of the Mechanical Behavior of Biomedical Materials 142 (2023) 105856

MATLAB code. Therefore, the average thickness of the meninges were In this case the difference between the final mesh and previous mesh was
calculated manually using Fiji (Schindelin, 2012) and found to be 0.91% and 0.93% when simulating impacts (i) and (ii), respectively.
approximately 0.08 mm thick. Due to the symmetry of the impactor and Overall, the model consists of 211,540 elements with uniform cell size
assumed symmetry of the spinal cord at the impact site, a one-quarter (125 μm) and nodes in contacts were merged using the contact matching
symmetrical model about the impactor’s centre was generated using tool in ANSYS.
symmetrical boundary conditions (Fig. 1C). Considering the thoracic
portion of the spinal cord is relatively flat longitudinally, the spinal cord 2.3. Finite element model validation
geometry was generated by extruding the right hemisphere’s geometry
10 mm in the axial direction. Given the orders of magnitude higher shear The FESM described above was validated by simulating contusion
moduli values and almost incompressible behaviour of the spinal cord SCI experiments performed by Lam et al. (2014) and comparing the
and meninges tissues, the cerebrospinal fluid was assumed to have a force-time curves from simulations against those recorded experimen­
negligible contribution to the mechanical behaviour of the spinal cord tally and against Lam et al.’s, simulation of the same experiment. The FE
during impact (Maikos et al., 2008b). Also, in order to generalise the model’s impactor geometry was modified to be consistent with the ge­
model and improve computational efficiency, the dura mater, arachnoid ometry of the impactor used in the experiments of Lam et al. (2014), i.e.,
mater, and pia mater were modelled as a single layer (meninges) and it a cylindrical geometry of 2 mm diameter with a 0.1 mm, 45◦ taper. Two
was assumed to have a uniform thickness of 0.08 mm, and the central experimental protocols described in Lam et al. (2014) were simulated:
median fissure was removed. The main impactor geometry used here impactor depths of 0.9 mm (mild injury) and 1.5 mm (moderate injury)
was a cylindrical geometry with a radius of 0.785 mm which tapered out using the mean displacement-time profiles up to the target displacement
at an angle of 117.65◦ degrees to a radius of 1 mm measured from im­ for each experiment. The resulting FE force-time data was downsampled
ages taken under a dissection microscope (Fig. 1B). from >100,000 datapoints to approximately 100 datapoints for
The 3D geometries were subsequently imported into ANSYS and computational efficiency. As shown in Fig. 2 (A-B), for ‘mild’ and
boundary conditions, contacts, constraints, materials, and meshes were ‘moderate’ impactor depths the reaction force-time curve produced by
applied. A rigid ‘floor’ was placed below the spinal cord and was fixed in the model (black lines) closely matched the experimentally recorded
all directions representing the inferior portion of the vertebrae and data (blue lines). Therefore, we can conclude that the model is validated
intervertebral discs below the laminectomy site. Symmetrical boundary and capable of producing reasonable predictions for SCI biomechanics.
conditions were applied to the impactor and the spinal cord, and the end The simulation force-time curves from Lam et al. (2014) are also shown
of the spinal cord opposite the impactor had no boundary conditions for comparison. Note: the data adapted from Lam et al., was considered
applied (Fig. 1C). Fixed contacts were applied between meninges-GM, only up to the maximum force (target displacement) for ease of
meninges-WM, and GM-WM. Frictionless contacts were applied be­ comparison.
tween the impactor and meninges, and the meninges and rigid floor
except for a small section at the bottom centre of the cord that was fixed 2.4. Simulations of dynamic indentation
to the rigid floor (red line Fig. 1B and C). The spinal cord and meninges
tissues were modelled as Ogden-based viscoelastic materials. The ma­ Impacts with impactor depth up to 1, 1.5, and 2 mm at constant
terial model parameters are presented in Table 1. velocities of 1, 1.5, and 2 m/s were applied to the model using the main
Tetrahedral element meshes were applied to the impactor, GM, WM, impactor geometry described above (Fig. 1B) to investigate the biome­
and meninges, and shell elements were used for the rigid floor. A mesh chanics of experimental in vivo spinal cord contusion injury. The
convergence study was performed by performing two simulations impactor geometry and depths were chosen as these are the geometry
increasing depth with (i) constant impactor velocity = 0.84 m/s, and indentation depths corresponding to a ‘mild’, ‘moderate’ and ‘se­
impactor depth = 0.9 mm, and (ii) constant impactor velocity = 0.84 m/ vere’ traumatic spinal cord injury at 1 m/s impactor velocity in our other
s, impactor depth = 1.5 mm. The maximum element size was decreased work. The classifications of ‘mild’, ‘moderate’, and ‘severe’ correspond
until the error between the resulting force-time curves was less than 1%. to the level of functional deficits arising from these contusion SCI ex­
periments. Constant impact velocities of 1, 1.5, and 2 m/s were also
applied to the impactor to investigate the effects of velocity on the
Table 1 resulting injury biomechanics. The effects of impactor geometry were
The material properties used in the spinal cord FE model. μ0 = shear modulus, α
also investigated by creating a spherical impactor of 2 mm diameter.
= nolinear parameter, D = compressibility factor, gi = relaxation modulus, τi =
Impacts using a spherical impactor to depths of 1, 1.5, and 2 mm at 1,
time constant.
1.5, and 2 m/s were also simulated and compared against the results of
Geometry Material Hyperelastic Viscoelastic Density the cylindrical impactor.
Model parameters Prony
parameters
3. Results
Meninges Ogden μ0 = 207.41 g1 = 0.3182, τ1 1000 kg/m3
kPa = 9 ms
α = 16.2 g2 = 0.1238, τ2 3.1. Effects of impactor velocity and depth
7
D = 1.7e− = 81 ms
Pa− 1 g3 = 0.0997, τ3 The von Mises stress (Pa), maximum shear stress (Pa), von Mises
= 564 ms strain, and shear strain generated in the spinal cord tissues by varying
g4 = 0.0997, τ4
= 4.69 s
the impactor velocity (1, 1.5, 2 m/s) and depth (1, 1.5, 2 mm) are
Spinal Ogden μ0 = 40.04 kPa g1 = 0.5282, τ1 1000 kg/m3 presented in Fig. 3. Interestingly, increasing the impactor velocity ap­
cord α = 4.7 = 8 ms pears to have a negligible effect on the distribution and magnitude of the
6
D = 4.9e− g2 = 0.3018, τ2 stresses and strains in the grey and white matter. This is due to the
Pa− 1 = 150 ms
impact durations ranging from 0.5 ms (depth = 1 mm, velocity = 2 m/s)
Impactor Rigid – – Equivalent to an
Body impactor mass of to 2 ms (depth = 2 mm, velocity = 1 m/s) and the first Prony series time
2 g. constant being 8 ms and 9 ms for the meninges and spinal cord tissue,
Floor Rigid – – Fixed. respectively. Therefore, negligible viscoelastic effects take place during
Body the ramp phase of the impacts investigated here. However, as expected
The meninges and spinal cord material properties were adopted from Russel there is a noticeable increase in the distribution and magnitude of the
et al. (2012). stresses and strains when the impactor depth is increased. At 2 mm

3
R. Frantsuzov et al. Journal of the Mechanical Behavior of Biomedical Materials 142 (2023) 105856

Fig. 2. Comparison with previous experimental results in the literature for contusion SCI with a target velocity (v) of 0.84 m/s to a depth (d) of 0.9 mm (A) and 1.5
mm (B) which agree well with experiments of mild and moderate impacts from Lam et al. (2014).

depth almost the entire dorsal column directly below the centre of the geometry was measured using a histological image taken approximately
impactor is at the maximum stress value (1 MPa), compared to the same 12 mm rostral of a T8 laminectomy and contusion impact site in a female
region at 1 and 1.5 mm. A larger area of the spinal cord tissue including Wistar rat, 4 weeks after injury. The model was validated by simulating
the entire dorsal column and a marked area of the grey matter are also in vivo CCI experiments from Lam et al. (2014) and comparing the
under the maximum observed strains (89–100%) at 2 mm when force-time curves from the simulations and experiments. Once vali­
compared to 1 mm (56–67%) and 1.5 mm (78–89%), demonstrating the dated, the effects of impactor velocity, depth, and geometry on the
sensitivity of spinal cord tissue to changes in the impactor displacement resulting stresses and strains within the cord were investigated. It was
into the cord. Maximum shear stress and maximum shear strain were found that there are distinct differences in the resulting stress and strain
also found in the spinal cord when impact depth reached 2 mm. The fields within the spinal cord depending on impactor depth and geometry
maximum shear stress at 2 mm was 0.44–0.5 MPa which was distributed but not velocity, which are discussed in detail below. These results have
throughout the entire dorsal column and the dorsal column-grey matter implications for establishing the relationship between the biomechanics
boundary, with a similar distribution pattern to the von Mises stress. and pathophysiology of SCI and in the design, interpretation and
Maximum shear strain ranged between 160 and 180% throughout the standardisation of in vivo SCI experiments.
dorsal column, and the dorsal column-grey matter boundary, and the
grey matter-anterolateral system boundary. 4.1. Simulating traumatic spinal cord injury

3.2. Effects of impactor geometry FE models of experimental SCI are useful tools in determining the
correlation between the biomechanics and resulting injury pathology.
Cylindrical and spherical are the two main impactor geometries used Previously, it has been demonstrated that the patterns of FE-predicted
in contusion SCI experiments (Gruner, 1992; Lee et al., 2012). It has stresses and strains match the patterns of primary injury (Maikos
previously been demonstrated in mechanical characterisation studies et al., 2008a) and tissue damage (Russel et al., 2012). Here, we have
that the geometry of the indenter (impactor) can significantly affect the demonstrated the ability to develop anatomically accurate FE models of
stresses and strains produced within materials (Briscoe et al., 1994; Xu the thoracic rat spinal cord which can simulate experimental SCI up to 2
and Li, 2008; Iivarinen et al., 2014). However, hitherto this has not been mm at 2 m/s impacts into the cord. The model presented here will be
investigated in experimental contusion SCI studies. Here, we investi­ used in the future to determine what, if any, correlations exist between
gated such effects by simulating impacts to the spinal cord as described the predicted stresses and strains and likelihood of injury and recovery
above with both cylindrical and spherical impactors of equivalent di­ in rat models of contusion SCI. Furthermore, such models could also be
ameters. The results from cylindrical and spherical simulations are used in the experimental design of in vivo contusion SCI to reduce
presented in Figs. 3 and 4, respectively. There were negligible differ­ variability and number of animals, in accordance with the NC3Rs (https:
ences in the maximum von Mises stress and strain, and maximum shear //www.nc3rs.org.uk/), and improve repeatability of contusion SCI
stress and strain values between the spherical and cylindrical geome­ studies by allowing researchers to select appropriate impactor geome­
tries. However, vastly different spatial distributions for these parameters tries and depths that correlate with specific injury-inducing strains
were observed between the two different geometries. Overall, the cy­ (Maikos et al., 2008a; Lam et al., 2014). However, there are a number of
lindrical geometry produced larger regions of iso-stress and iso-strain open questions regarding the biomechanics of SCI which must be
than the spherical geometry, although the maximum values were the answered for these computational models to provide accurate and reli­
same for both geometries. This is likely due to a focusing effect produced able data that can be used to predict contusion SCI. Here, we provide
by the spherical geometry as the contact area increases with penetration new insights into the biomechanics of contusion SCI by investigating the
depth. potential effects of impactor velocity, depth, and geometry. These as­
pects of experimental and computational SCI are discussed in detail
4. Discussion below.

A FE model of the rat thoracic spinal cord has been developed and
validated for modelling contusion SCI experiments. The FE model’s

4
R. Frantsuzov et al. Journal of the Mechanical Behavior of Biomedical Materials 142 (2023) 105856

Fig. 3. The results from simulating a contusion injury with a cylindrical impactor to a depth of 1 mm, 1.5 mm and 2 m/s at 1 m/s, 1.5 m/s, and 2 m/s are shown. The
von Mises stress (A), von Mises Strain (B), maximum shear stress (C), and maximum shear strain (D) in the spinal cord tissues are presented at the maximum impactor
displacement at the end of the ramp phase.

4.2. Effects of impactor velocity and depth velocity did not cause any noticeable changes in the predicted stresses
and strains (Figs. 3 and 4). In contrast to the literature on brain tissues’
Experimental SCI and traumatic brain injury (TBI) have been used to mechanical properties, there are limited experimental mechanical
model varying levels of injury e.g. mild, moderate, severe. These injury characterisation studies of spinal cord tissues that show an increase in
levels are typically defined in the experiments by the amount of pene­ the apparent tissue stiffness with increasing rate of deformation (Bilston
tration into the tissue e.g. 1, 1.5, and 2 mm representing mild, moderate, and Thibault, 1996; Fiford and Bilston, 2005; Clarke et al., 2009; Fradet
and severe injury, respectively, with ‘mild’, ‘moderate’, and ‘severe’ et al., 2016). Furthermore, there is a dearth of mechanical properties for
typically referring to the probability of structural and functional dam­ spinal cord tissue characterised under large strains at high strain rates
age. Further, it is well documented that CNS tissues such as the brain and similar to those experienced during traumatic SCI. The need for such
spinal cord are sensitive not only to the amount of deformation but also data is further highlighted by the work of Maikos and Shreiber (2007)
to the rate of deformation (Bilston and Thibault, 1996; Fradet et al., that demonstrated a stronger correlation between spinal cord micro­
2016; MacManus et al., 2017a; 2018). Therefore, the effects of impactor vasculature injury and strain rate than with depth of spinal cord
velocity on the resulting stress and strain fields were also investigated. compression. Of note is the concentration of maximum stresses in the
However, due to the Prony series’ first time constant for spinal cord grey matter (See: Fig. 3). Indeed, Fiford et al. (2004) demonstrated a
tissue being larger than the duration of impacts, increasing the impactor concentration of vascular injury in spinal cord grey matter and

5
R. Frantsuzov et al. Journal of the Mechanical Behavior of Biomedical Materials 142 (2023) 105856

Fig. 4. The results from simulating a contusion injury with a spherical impactor to a depth of 1 mm, 1.5 mm and 2 m/s at 1 m/s, 1.5 m/s, and 2 m/s are shown. The
von Mises stress (A), von Mises Strain (B), maximum shear stress (C), and maximum shear strain (D) in the spinal cord tissues are presented at the maximum impactor
displacement at the end of the ramp phase.

Khosroshahi et al., 2021 have recently shown the important relationship generation and progression of stresses and strains as the initiating fac­
between shear stresses and vascular injury in TBI. Therefore, in the tors of SCI pathology. The two main impactor geometries used for
future it may be possible to predict microvasculature injury in the spinal indentation mechanical characterisation and experimental contusion
cord using computational models such as the one presented here. SCIs are cylindrical and spherical geometries (Gruner, 1992; Lee et al.,
2012; Finan et al., 2014; MacManus et al., 2017b; Menichetti et al.,
4.3. Effects of impactor geometry 2020). The results presented in Figs. 3 and 4 demonstrate the markedly
different patterns of stress and strain that are produced in spinal cord
Erbayraktar et al. (2013) suggested that experimental contusion in­ tissue for cylindrical and spherical impactors of equivalent diameter at
juries, such as the CCI model, are considered biomechanically similar to equal velocity and indentation depth. These findings highlight the dif­
vertebral fractures and dislocations and therefore provide the most ficulty in comparing results in SCI literature when different impactor
relevant and realistic experimental conditions to test potential neuro­ geometries are used. While the maximum values of von Mises and shear
protective and regenerative therapeutics. Considering the similarity of stresses are similar for both geometries, larger distributions are observed
the biomechanics of CCI to vertebral fractures and dislocations when using the cylindrical impactor compared with the spherical
described by Erbayraktar et al. (2013), the CCI model is also likely to impactor. Additionally, the maximum values of the von Mises and shear
provide the most relevant experimental conditions to investigate the strains are similar for both geometries with markedly different

6
R. Frantsuzov et al. Journal of the Mechanical Behavior of Biomedical Materials 142 (2023) 105856

distributions. The cylindrical impactor produces larger distributions of or cavitation (Wrede et al., 2020).
von Mises and shear strains than the spherical impactor’s more localised
distributions. These results could be used to guide the design and 4.5. Limitations
establish future protocols for experimental contusion SCI. For example,
the spherical geometry may be more suitable for investigating injury to Although the work presented here provides a number of novel in­
specific spinal cord regions due to its focusing effect on strain. Whereas, sights into the biomechanics of SCI, there are several limitations which
the cylindrical geometry may be more suitable for investigating diffuse need to be addressed and discussed. The geometry of the spinal cord is
injury at the impact site. Further analysis should be conducted to generated from a single transverse cryosection (10 μm thickness) taken
investigate the effects of different impactor sizes and geometries on from the thoracic portion of the spinal cord approximately 12 mm
resulting injury severity and location. rostral of a contusion injury site. While this provides an accurate
geometrical representation of this specific area of the spinal cord, it does
4.4. Implications for experimental traumatic spinal cord injury not capture how the true morphology of the spinal cord changes along
the length of the cord considered here. Similarly, in order to simplify the
These data implicate two key considerations in the design and model and improve computational efficiency, the mechanical contri­
interpretation of experimental contusion SCI studies. Firstly, impact bution of the CSF was assumed to be negligible, and the meninges were
parameters (such as tip geometry) need to be accounted for, both when in direct contact with the WM and GM regions. However, the model has
standardising experiments and consolidating data across studies. In SCI recapitulated the experimentally measured force-time values seen in the
studies performing controlled contusion impacts, details on the induced literature and validated its use (Fig. 2). The boundary conditions
injury parameters, including force, depth or impactor geometry, are imposed on this model may not reflect the in vivo boundary conditions
often limited or missing despite recommendations within the field to during experimental SCI. During in vivo contusion SCI experiments, a
provide detail of the biomechanics of a contusion device (Lemmon et al., portion of the vertebrae is removed which will impose different
2014). Many publications report the impact force (kDyne) as a measure boundary conditions upon the cord. Here, we have assumed that the
of contusion severity without reporting tip geometry (Donnelly et al., region of the cord being impacted is directly in the centre of the lam­
2011; Fenn et al., 2014; DePaul et al., 2017; Freria et al., 2017; Nishi inectomy region with the spinal cord supported inferiorly by the rigid
et al., 2020). When tip geometry is reported, a wide range of tip di­ floor, with a small inferior portion of the meninges rigidly fixed to the
ameters are evident, including 1 mm tips (Nishi et al., 2020) and 2.5 mm rigid floor, representing the boundary conditions imposed by the
tips (DePaul et al., 2017). Some mention the use of tips with chamfered vertebrae and anchoring the cord in place. Furthermore, the dentate
edges (Lee et al., 2012) without details on the degree of chamfering or ligaments that attach the pia mater to the dural sac and thus anchor the
honing involved. Given the differences in distribution observed in strain spinal cord in place, were not modelled here. Instead the function of the
and stress when modelling a spherical tip (Fig. 4), increasing the degree dentate ligaments were assumed to be captured within the contact
of chamfering is likely to shift this distribution toward that of a spherical condition that fixed the meninges to the spinal cord. Symmetry
impactor, with implications for the severity of motor function deficits boundary conditions were used to enforce a one-quarter model (Fig. 1).
observed as well as the downstream effects of inflammatory infiltration, Future work should include analysis on the effects of including the
wound compaction and re-vascularisation. vertebrae and other anatomy of the spine to ensure boundary conditions
Secondly, because downstream strain and stress arising from impacts are reflective of in vivo conditions. Indeed, Stoner et al. (2020) and
depend on the biomechanical nature of the spinal cord (namely, Russel et al. (2012) included the vertebrae and curvature in their models
compressive stiffness of the meningeal layers, and the differential me­ of human and rat cervical spinal cord, respectively. Future work will
chanical stiffness of GM and WM), putative age-, sex- and region-specific address these limitations by developing a model with more
variations in these biomechanical properties (Okada et al., 2009; region-specific detail to account for changes in the grey and white
Takamura et al., 2020) will impart cryptic effects on the segmentation of matter morphology, and include the CSF and vertebrae to obtain a more
strain and stress across the various subregions of the spinal cord. These biofidelic response of the spinal cord during experimental contusion SCI.
need to be considered when designing and performing SCI experiments, Further, the same mechanical properties were used to describe grey and
but also present an opportunity in further delineating how age and sex white matter in this model with a Prony series first time constant being
ultimately effect SCI pathology and recovery. Indeed, a full characteri­ more than four times larger than the duration of the impact due to the
sation of the mechanical properties of spinal cord tissues remains a dearth of region specific mechanical properties for spinal cord tissue
priority (Fournely et al., 2020), and such a characterisation will ulti­ characterised under strains at strain rates suitable for modelling trau­
mately improve the applicability of FE models to experimental design matic SCI. Region-specific mechanical properties of spinal tissue char­
and interpretation. acterised under conditions suitable for modelling trauma are urgently
One critical aspect of modelling data such as these is that they pro­ needed to more accurately describe the mechanical behaviour of spinal
vide predictive value when determining how the initial mechanical cord tissue during SCI.
forces during a contusion give rise to distinct biological signalling
events. Secondary injury is a hallmark of SCI (Bradbury and Burnside, 5. Conclusion
2019), but whether different aspects of secondary injury (e.g., extra­
cellular matrix remodelling and fibrosis, astrogliogenesis, immune cell The current work presents a new FE model of the rat thoracic spinal
infiltration and activation) segment with the distribution of stress/strain cord for modelling contusion SCI. This new FE model was developed
during impact remains unresolved. For instance, unilateral contusions using an immunohistochemistry image of the intact rat cord and was
are sometimes accompanied by contralateral motor function deficits validated with existing experimental in vivo CCI data from the literature.
despite the absence of severe tissue damage on the contralateral side The effects of impactor velocity, depth, and geometry on the resulting
(Streijger et al., 2013; Watson et al., 2014). In an FE model of unilateral biomechanics were investigated providing several insights into the
contusion, strain was predicted to propagate into the contralateral cord, mechanical behaviour of the spinal cord during impact. The impactor
in a manner dependent on the inhomogeneity of white and grey matter geometry plays an important role in the distribution and magnitude of
stiffness (Fournely et al., 2020). Through the iteration of different stresses and strains within the cord and should be considered during in
impact parameters, FE models can help determine whether deficits like vivo experimental design in order to improve repeatability and
these, even where tissue damage is not evident, arise from non-localised reproducibility.
events (such as generalised secondary inflammation) or more directly
from difficult-to-observe phenomena such as strain (Russel et al., 2012)

7
R. Frantsuzov et al. Journal of the Mechanical Behavior of Biomedical Materials 142 (2023) 105856

CRediT authorship contribution statement Fiford, R.J., Bilston, L.E., Waite, P., Lu, J., 2004. A vertebral dislocation model of spinal
cord injury in rats. J. Neurotrauma 21 (4), 451–458.
Finan, J.D., Fox, P.M., Morrison, B., 2014. Non-ideal effects in indentation testing of soft
Roman Frantsuzov: Writing – review & editing, Visualization, tissues. Biomech. Model. Mechanobiol. 13 (3), 573–584.
Validation, Methodology, Investigation, Formal analysis, Data curation. Fournely, M., Petit, Y., Wagnac, E., Evin, M., Arnoux, P.-J., 2020. Effect of experimental,
Subrata Mondal: Writing – review & editing, Methodology, Investiga­ morphological and mechanical factors on the murine spinal cord subjected to
transverse contusion: a finite element study. PLoS One 15, e0232975.
tion, Formal analysis, Data curation. Ciara M. Walsh: Writing – review Fradet, L., Cliché, F., Petit, Y., Mac-Thiong, J., Arnoux, P., 2016. Strain rate dependent
& editing, Writing – original draft, Visualization, Methodology, Inves­ behaviour of the porcine spinal cord under transverse dynamic compression. Proc.
tigation, Funding acquisition. James P. Reynolds: Writing – review & IMechE Part H: J. Engineering in Medicine 230 (9), 858–866.
Freria, C.M., et al., 2017. Deletion of the fractalkine receptor, CX3CR1, improves
editing, Writing – original draft, Visualization, Methodology, Investi­ endogenous repair, axon sprouting, and synaptogenesis after spinal cord injury in
gation, Data curation. Dearbhaile Dooley: Writing – review & editing, mice. J. Neurosci. 37, 3568–3587.
Writing – original draft, Supervision, Resources, Project administration, Ghajari, M., et al., 2017. Computational modelling of traumatic brain injury predicts the
location of chronic traumatic encephalopathy pathology. Brain 140 (2), 333–343.
Methodology, Funding acquisition, Conceptualization. David B. Mac­ Greaves, C.Y., Gadala, M.S., Oxland, T.R., 2008. A three-dimensional finite element
Manus: Writing – review & editing, Writing – original draft, Visuali­ model of the cervical spine with spinal cord: an investigation of three injury
zation, Supervision, Project administration, Methodology, Investigation, mechanisms. Ann. Biomed. Eng. 36 (3), 396–405.
Gruner, J.A., 1992. A monitored contusion model of spinal cord injury in the rat.
Formal analysis, Data curation, Conceptualization. J. Neurotrauma 9 (2), 123–128.
Iivarinen, J.T., Korhonen, R.K., Jurvelin, J.S., 2014. Experimental and numerical analysis
of soft tissue stiffness measurement using manual indentation device – significance
Declaration of competing interest of indentation geometry and soft tissue thickness. Skin Res. Technol. 20, 347–354.
Jazayeri, S.B., Beygi, S., Shokraneh, F., Hagen, E.M., Rahimi-Movaghar, V., 2015.
Incidence of traumatic spinal cord injury worldwide: a systematic review. Eur. Spine
The authors declare that they have no known competing financial J. 24, 905–918.
interests or personal relationships that could have appeared to influence Khosroshahi, S.F., et al., 2021. Multiscale modelling of cerebrovascular injury reveals the
role of vascular anatomy and parenchymal shear stresses. Sci. Rep. 11, 12927.
the work reported in this paper.
Khuyagbaatar, B., Kim, K., Park, W.M., Kim, Y.H., 2016. Biomechanical behaviors in
three types of spinal cord injury mechanisms. J. Biomech. Eng. 138, 081003.
Data availability Kjell, J., Olson, L., 2016. Rat models of spinal cord injury: from pathology to potential
therapies. Dis. Model Mech. 9 (10), 1125–1137.
Lam, C.J., Assinck, P., Liu, J., Tetzlaff, W., Oxland, T.R., 2014. Impact depth and the
Data will be made available on request. interaction with impact speed affect the severity of contusion spinal cord injury in
rats. J. Neurotrauma 31 (12), 1985–1997.
Lee, J.H.T., et al., 2012. A contusive model of unilateral cervical spinal cord injury using
Acknowledgements the infinite horizon impactor. J. Vis. Exp. (65), 3313.
Lemmon, V.P., et al., 2014. Minimum information about a spinal cord injury experiment:
Subrata Mondal is funded from the European Union’s Horizon 2020 a proposed reporting standard for spinal cord injury experiments. J. Neurotrauma 31
(15), 1354–1361.
research and innovation program under grant agreement number
MacManus, D.B., Gilchrist, M.D., Murphy, J.G., 2017a. An empirical measure of
814410. nonlinear strain for soft tissue indentation. R. Soc. Open Sci. 4 (11), 170894.
James P. Reynolds is funded by Science Foundation Ireland (19/FFP/ MacManus, D.B., Pierrat, B., Murphy, J.G., Gilchrist, M.D., 2017b. Region and species
dependent mechanical properties of adolescent and young adult brain tissue. Sci.
6642 awarded to DD).
Rep. 7, 13729.
Ciara M. Walsh is funded by UCD School of Medicine and an Irish MacManus, D.B., Murphy, J.G., Gilchrist, M.D., 2018. Mechanical characterisation of
Research Council Government of Ireland Postgraduate Scholarship brain tissue up to 35% strain at 1, 10, and 100/s using a custom-built micro-
(GOIPG/2021/304). indentation apparatus. J. Mech. Behav. Biomed. Mater. 87, 256–266.
Maikos, J.T., Shreiber, D.I., 2007. Immediate damage to the blood-spinal cord barrier
due to mechanical trauma. J. Neurotrauma 24 (3), 492–507.
References Maikos, J.T., Qian, Z., Metaxas, D., Shreiber, D.I., 2008a. Finite element analysis of
spinal cord injury in the rat. J. Neurotrauma 25, 795–816.
Maikos, J.T., Elias, R.A.I., Shreiber, D.I., 2008b. Mechanical properties of dura mater
Alizadeh, A., Dyck, S.M., Karimi-Abdolrezaee, S., 2019. Traumatic spinal cord injury: an
from the rat brain and spinal cord. J. Neurotrauma 25, 38–51.
overview of pathophysiology, models and acute injury mechanisms. Front. Neurol.
Menichetti, A., MacManus, D.B., Gilchrist, M.D., Depreitere, B., Vander-Sloten, J.,
10, 282.
Famaey, N., 2020. Regional characterization of the dynamic mechanical properties
Bárbara-Bataller, E., Méndez-Suárez, J.L., Alemán-Sánchez, C., et al., 2018. Change in
of human brain tissue by micro-indentation. Int. J. Eng. Sci. 155, 103355.
the profile of traumatic spinal cord injury over 15 years in Spain. Scand. J. Trauma
National Office of Clinical Audit, 2020. Major Trauma Audit National Report 2018.
Resuscitation Emerg. Med. 26 (27).
Ireland.
Bilston, L.E., Thibault, L.E., 1996. The mechanical properties of human cervical spinal
NC3Rs, The 3Rs, National Centre for the Replacement Refinement and Reduction of
cord in vitro. Ann. Biomed. Eng. 24, 67–74.
Animals in Research, Retrieved 09/08/2021..
Bloom, O., Herman, P.E., Spungen, A.M., 2020. Systemic inflammation in traumatic
Nishi, R.A., et al., 2020. The effects of mouse strain and age on a model of unilateral
spinal cord injury. Exp. Neurol. 325, 113143.
cervical contusion spinal cord injury. PLoS One 15, e0234245.
Bradbury, E.J., Burnside, E.R., 2019. Moving beyond the glial scar for spinal cord repair.
Okada, E., et al., 2009. Aging of the cervical spine in healthy volunteers: a 10-year
Nat. Commun. 10, 3879.
longitudinal magnetic resonance imaging study. Spine 34, 706–712.
Briscoe, B.J., Sebastian, K.S., Adams, M.J., 1994. The effect of indenter geometry on the
Reier, P.J., et al., 2012. Chapter 26 – translational spinal cord injury research: preclinical
elastic response to indentation. J. Phys. Appl. Phys. 27 (6), 1156.
guidelines and challenges. Handb. Clin. Neurol. 109, 411–433.
Cheriyan, T., et al., 2014. Spinal cord injury models: a review. Spinal Cord 52 (8),
Russel, C.M., Choo, A.M., Tetzlaff, W., Chung, T., Oxland, T.R., 2012. Maximum
588–595.
principal strain correlates with spinal cord tissue damage in contusion and
Clarke, E.C., Cheng, S., Bilston, L.E., 2009. The mechanical properties of the neonatal rat
dislocation injuries in the rat cervical spine. J. Neurotrauma 29, 1574–1585.
spinal cord in vitro and, comparisons with adult. J. Biomech. 42, 1397–1402.
Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., et al.,
DePaul, M.A., Lin, C.-Y., Silver, J., Lee, Y.-S., 2017. Combinatory repair strategy to
2012. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9
promote axon regeneration and functional recovery after chronic spinal cord injury.
(7), 676–682.
Sci. Rep. 7 (9018).
Sharif-Alhoseini, M., et al., 2017. Animal models of spinal cord injury: a systematic
Donnelly, D.J., et al., 2011. Deficient CX3CR1 signaling promotes recovery after mouse
review. Spinal Cord 55 (8), 714–721.
spinal cord injury by limiting the recruitment and activation of Ly6Clo/iNOS+
Stoner, K.E., et al., 2020. A comprehensive finite element model of surgical treatment for
macrophages. J. Neurosci. 31, 9910–9922.
cervical myelopathy. Clin. BioMech. 74, 79–86.
Erbayraktar, Z., Gökmen, N., Yılmaz, O., Erbayraktar, S., 2013. Experimental traumatic
Streijger, F., et al., 2013. Characterization of a cervical spinal cord hemicontusion injury
spinal cord injury. In: Ghezzi, P., Cerami, A. (Eds.), Tissue-Protective Cytokines.
in mice using the infinite horizon impactor. J. Neurotrauma 30, 869–883.
Methods in Molecular Biology (Methods and Protocols), 982. Humana Press, Totowa,
Takamura, T., et al., 2020. Influence of age on global and regional brain stiffness in
NJ.
young and middle-aged adults. J. Magn. Reson. Imag. 51, 727–733.
Fenn, A.M., Hall, J.C.E., Gensel, J.C., Popovich, P.G., Godbout, J.P., 2014. IL-4 signaling
Trauma Steering Group, A trauma system for Ireland, Published February 2018,
drives a unique arginase+/IL-1β+ microglia phenotype and recruits macrophages to
Available online: https://www.hse.ie/eng/about/who/acute-hospitals-division/t
the inflammatory CNS: consequences of age-related deficits in IL-4Rα after traumatic
rauma-services/further-information-and-documentation/..
spinal cord injury. J. Neurosci. 34, 8904–8917.
Watson, J.L., Hala, T.J., Putatunda, R., Sannie, D., Lepore, A.C., 2014. Persistent at-level
Fiford, R.J., Bilston, L.E., 2005. The mechanical properties of rat spinal cord in vitro.
thermal hyperalgesia and tactile allodynia accompany chronic neuronal and
J. Biomechanics 38 (7), 1509–1515.

8
R. Frantsuzov et al. Journal of the Mechanical Behavior of Biomedical Materials 142 (2023) 105856

astrocyte activation in superficial dorsal horn following mouse cervical contusion Xu, Z.-H., Li, X., 2008. Effects of indenter geometry and material properties on the
spinal cord injury. PLoS One 9, e109099. correction factor of Sneddon’s relationship for nanoindentation of elastic and elastic-
Wrede, A.H., et al., 2020. Characterization of astrocytic response after experiencing plastic materials. Acta Mater. 56, 1399–1405.
cavitation in vitro. Glob Chall 4, 1900014. Yan, Y.-B., Qi, W., Wu, Z.-X., Qiu, T.-X., Teo, E.-C., Lei, W., 2012. Finite element study of
the mechanical response in spinal cord during the thoracolumbar burst fracture.
PLoS One 7 (9), e41397.