Computational biology — Modeling of primary blast effects on the
central nervous system
David F. Moorea,⁎, Antoine Jérusalemb, Michelle Nyeinb, Ludovic Noelsc,
Michael S. Jaffeea, Raul A. Radovitzkyb
aDefense and Veterans Brain Injury Center, Walter Reed Army Medical Center, Building 1, Room B207, 6900 Georgia Avenue NW, Washington DC 20309-5001, USA
bDepartment of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge, MA, USA
cAerospace and Mechanical Engineering Department, University of Liège, Liège, Belgium
a b s t r a c ta r t i c l e i n f o
Received 14 November 2008
Revised 2 February 2009
Accepted 4 February 2009
Available online 24 February 2009
Objectives: Recent military conflicts in Iraq and Afghanistan have highlighted the wartime effect of traumatic
brain injury (TBI). The reason for the prominence of TBI in these particular conflicts as opposed to others is
unclear but may result from the increased survivability of blast due to improvements in body armor. In the
military context blunt, ballistic and blast effects may all contribute to CNS injury, however blast in particular,
has been suggested as a primary cause of military TBI. While blast effects on some biological tissues, such as
the lung, are documented in terms of injury thresholds, this is not the case for the CNS. We hypothesized that
using bio-fidelic models, allowing for fluid–solid interaction and basic material properties available in the
literature, a blast wave would interact with CNS tissue and cause a possible concussive effect.
Methods: The modeling approach employed for this investigation consisted of a computational framework
suitable for simulating coupled fluid-solid dynamic interactions. The model included a complex finite
element mesh of the head and intra-cranial contents. The effects of threshold and 50% lethal blast lung injury
were compared with concussive impact injury using the full head model allowing upper and lower bounds of
tissue injury to be applied using pulmonary injury as the reference tissue.
Results: The effects of a 50% lethal dose blast lung injury (LD50) were comparable with concussive impact
injury using the DVBIC-MIT full head model.
Interpretation: CNS blast concussive effects were found to be similar between impact mild TBI and the blast
field associated with LD50lung blast injury sustained without personal protective equipment. With the
ubiquitous use of personal protective equipment this suggests that blast concussive effects may more readily
ascertained in personnel due to enhanced survivability in the current conflicts.
© 2009 Published by Elsevier Inc.
Military operations in Iraq and Afghanistan have brought into
sharp focus military-related traumatic brain injury (TBI). Some recent
reports suggest that very significant numbers of Servicemembers are
affected by TBI, principally mild traumatic brain injury (mTBI) or
concussion defined by a Glasgow Coma Scale (GCS) N12 h, or loss of
consciousness (LOC) b1 h, or post-traumatic amnesia (PTA) b24 h
(Tanielian, 2008). In particular, the RAND study estimates 320,000
TBIs from a total Servicemember deployment of 1.64 million (19.2% of
the Servicemembers experiencing a probable TBI). More conservative
estimates as determined from the Defense and Veterans Brain Injury
Center (DVBIC) surveillance programs indicate that mTBI represents
∼10–20% of TBI screen positive Servicemembers. TBI is a significant
civilian cause of death and morbidity in the 0–40 year-old range
occurring typically from impact injury such as motor vehicle
accidents. There are huge direct and indirect economic costs to
society at large for TBI through the burden of care imposed on family
members and lost earning potential (Brazarian et al., 2005; Bruns and
The exposure of military personnel to the consequences of blast
wavesmayincreasethe overall TBIburden.Injuryfromblastis defined
as 1) primary blast injury directly due to the propagation of the blast
wave through the tissue, 2) secondary blast injury due to tissue injury
resulting from interactionwith shrapnel or fragments, 3) tertiary blast
injury with tissue injury due to impact with environmental structures,
for example buildings or vehicle roll-over and 4) quaternary blast
injury due to heat, electromagnetic pulses or toxic detonation
products such as carbon monoxide (DePalma et al., 2005; Moore
et al., 2008). Typical military events represent a mixture of all four
categories of blast tissue injury where separation or identification of
the magnitude of the contributing components from primary to
quaternary injury factors may be impractical. For this reason, the term
NeuroImage 47 (2009) T10–T20
⁎ Corresponding author.
E-mail address: email@example.com (D.F. Moore).
1053-8119/$ – see front matter © 2009 Published by Elsevier Inc.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/ynimg
Blast (+) TBI is introduced so as to allow for the recognition of the
concurrence of both blast and impact injury together with other blast
injury mechanisms. The low frequency of primary blast CNS injury is
illustrated by a recent isolated case report (Warden et al., 2009).
Nocurrent patient-based clinical investigativeevidence exists as to
whether blast waves ‘per se’ modify or alter the underlying CNS injury
cascade of TBI compared to blunt or impact injury. However, given the
evidence from animal studies and the parsimony of biological
processes, this would seem unlikely (Petras et al., 1997). A blast
wave is a pressure shockwave of finite amplitude resulting from an
atmospheric (gas fluid) explosion releasing a large amount of energy
in a short period of time. The explosive chemical energy following
detonation is converted to thermal, electromagnetic, and kinetic
energy, the latter being imparted to the surrounding material (such as
soil, explosive casing, and fragments) together with the primary blast
shockwave (DePalma et al., 2005; Iverson et al., 2006; Strehlow and
Baker,1976). An idealized free-field blast is most simply described by
the biphasic Friedlander waveform with a rapid rise to peak pressure
andan exponential fall-offof theover pressurefollowedbya relatively
prolonged underpressure. This results in a combination of compres-
sive and tensile material components when the wave propagates
through biological tissues (Elsayed, 1997; Mayorga, 1997). Several
theories have been suggested to account for TBI after blast exposure
such as direct propagation of the blast wave through the brain,
propagation of the pressure wave through the great vessels and
secondarily to the brain, and cavitation secondary to the blast
underpressure together with any associated blast electromagnetic
pulses (Moore et al., 2008).
Importantly, it is necessary to realize that an explosion is a non-
linear process characterized by a three-dimensional complex fluid
flow field that may be significantly influenced by ambient and
environmental boundary conditions. This may result in shock wave
reflections with up to an eight-fold amplitude intensification over the
primary shockwave. Such phenomenon indicates the considerable
possibility for variability in military associated blast exposure and TBI
resulting from a primary blast process (Cullis, 2001; Kambouchev et
The propagation of the shockwave within the fluid (air) and
subsequent direct interaction with solid structures such as the skull
and intra-cranial contents is the focus of the current paper. This allows
consideration of the following hypotheses: (a) whether there is a
direct interaction between a primary blast wave and CNS tissue and
(b) whether such an interaction is sufficient to cause concussion or
Recent results on fluid-structure interaction (FSI) underscore
the possibility of mitigating impulse transfer from the blast wave
to the structure by exploiting the FSI effect (Kambouchev et al.,
2006). Blast wave interaction with a material, either biological or
non-biological, results in the development of waves which have
longitudinal, shear (transverse), and surface (Rayleigh) wave
components (Kolsky, 1963). These waves, at high strain-rates
associated with blast, may have a differential pathological effect
on the anisotropic structures of the brain, especially white matter
tracts where directionality is paramount. Such a proposition is
borne out by recent results using Kolsky bar or Split Hopkinson
Pressure bar to experimentally extract material properties at high
strain rate following the development and measurement of
propagation speed of one-dimensional stress waves in solids. The
bulk modulus in different tissues varied non-linearly but was
effectively a monotonic function of the strain rate. By contrast, the
shear modulus exhibited more than one state across the strain rate
domain in some tissues (Saraf et al., 2007). A wide range of
strange rates is relevant to Blast (+) syndrome from low values
characteristic of impact injury typical of motor vehicle accidents to
high values associated with ballistic and blast or shock waves.
A lot of uncertainty exists regarding blast waves and the
subsequent propagation of stress waves in the brain. The relationship
of stress waves to TBI have been well studied using simulations under
impact loading such as in sports or motor vehicular injuries
(Belingardi et al., 2005; Gilchrist et al., 2001; Horgan and Gilchrist,
2004; Raul et al., 2006; Taylor and Ford, 2006; Zhang et al., 2001;
Zhang et al., 2004). For example, Willinger and Baum (2003),
recreating motorcycle accidents, correlated Von Mises stresses (a
scalar derivative function of the deviatoric component of the tissue
stress tensor) with concussion while strain energy in the CSF was
correlated with subdural hematoma formation Zhang et al. (2004),
recreating American football collisions, found high shear stress
concentrations localized in the upper brainstem and thalamus regions
and determined that shear stress response in the upper brainstemwas
a good predictor for mTBI.
The purpose of this paper is to develop a full head finite
element model and simulate the fluid–solid interaction with a
blast shockwave under open conditions. The paper firstly will
examine, whether direct interaction and propagation of a detona-
tion shockwave can occur through the skull and secondly compare
these mechanical events using ‘order of magnitude arguments’ to
other blast tissue damage such as the lung. As previously
suggested, a particular reason as to why TBI has been ascertained
with greater frequency in the current conflicts is likely due to the
significant strides in mitigation of thoracic and abdominal injury
by the use of personal protective equipment (PPE). Prior to the
availability of such PPE blast lung was a significant cause of
military mortality and morbidity. Standardized pressure standoff
(distance) curves or Bowen curves are available for a 70 kg person
allowing assessment of the biological impact of blast on lung tissue.
We utilized the Bowen survivability–lethality curves for unprotected
Fig. 1. DVBIC-MIT full head finite element model showing progressively more internal tissue layers. The skin (a), gray matter (b), and white matter (c) from the 808,766-element
head mesh used in the fluid–solid interface blast simulations.
Parameters for the Mie-Gruneisen/Hugoniot equation of state.
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20
pulmonary blast exposure to derive initial blast conditions that
could then be compared to known impact conditions causing
concussion (Bass et al., 2006; Bowen et al., 1968; Stuhmiller, 1987;
White et al., 1971).
In this way, we hoped to compare significant current survivable
blast injury with known impact mechanical forces causing concussion
to estimate whether a concussive potential might exist from a
survivable primary blast in the context of PPE use. In this work, we
used the Bowen curve derived threshold of lung injury of 5.2 atm and
the lethal dose 50% of lung injury of 18.6 atm to consider an upper and
lower bound of currently survivable blast injury. Results obtained for
these two bounds were then compared to impact injury typically seen
to potentially cause concussion in the sports injury environment.
In order to examine and compare the effects of blast shock waves
on the human head, simulations were run with a full head mesh in
three different contexts: (1) a blast with overpressure of 5.2 atm or
threshold lung injury, equivalent to a free air explosion of 0.0648 kg
TNT at a 0.6 m standoff distance; (2) a blast with overpressure of
18.6 atm — the 50% lethal dose (LD50) for lung injury survival,
equivalent to a free air explosion of 0.324 kg TNT at a 0.6 m
standoff distance; and (3) an impact between a head traveling at
5 m/s and a stationary/immovable boundary likely to result in
concussive injury based on comparable impact studies in the
literature (Casson et al., 2008; Zhang et al., 2004). The over-
pressures for the two blast simulations were selected based on the
Bowen curves, which give the estimated tolerance to a single blast
at sea level for a 70-kg human oriented perpendicular to the blast
(Bowen et al., 1968).
High-resolution T1 MR images were downloaded from the
Montreal Neurological Institute at an isotropic voxel dimension of
1×1×1 mm (Collins et al., 1998). These images were merged with a
bone windowed CT of head allowing skull reconstruction using a
mutual information algorithm. The resulting volume set of images
was then semi-automatically segmented using Amira (http://www.
amiravis.com/) into topological closed regions of interest followed
by export as VRML (Virtual Reality Modeling File) files. This format
was chosen because it is one of the solid model file formats
accepted by the software employed for mesh generation. Amira is
an imaging software analysis suite allowing structured regional
labeling of image data together with filtering and co-registration. A
variety of input formats and output formats are available. These
VRML files were then imported into the mesh generation capabil-
ities in the computational fluid mechanics software ICEMCFD
(http://www.ansys.com/products/icemcfd.asp). This software pro-
vides a variety of meshing algorithms capable of importing CAD
models of high topological and geometrical complexity and
producing volumetric conformal computational meshes required
by the blast simulations presented in this paper. In addition, the
software provides mesh decimation, refinement and smoothing
algorithms that can be used to optimize the mesh for computational
efficiency. An unstructured finite element mesh was constructed
using the Octree and Delaunay tetrahedral mesh generation
algorithms. The meshes were further refined by isolation of poorly
meshed areas and poorly shaped tetrahedra prior to running the
computational fluid–solid dynamics code (Deiterding et al., 2006).
A variety of computational meshes with different resolutions
were created and it was found that meshes with fewer than
700,000 elements were too coarse to describe the intricate
topology of some human head anatomical structures relevant for
blast injury analysis. Among the computational models of the
human head reported in the literature, the one with highest
resolution model appears to be the model developed by Zhang et
al. (2004) comprising 314,500 elements. As part of the model
development process, we produced several computational meshes
ranging from 800,000 to 5,000,000 elements as coarser models did
not capture either the geometric intricacies of brain structures (e.g.
gyrencephalic cortex and white matter fascicular structure) or, in
turn, the mechanical fields associated with the blast wave. In order
to balance mesh resolution and computational requirements, a
mesh with 808,766 elements was used in the current simulations
(Fig. 1). The computational model differentiates 11 distinct
structures characterized by mechanical properties summarized in
Tables 1 and 2.
Material models and properties
The constitutive response of brain tissue encompasses a variety of
complex mechanisms including nonlinear viscoelasticity, anisotropy
and strong strain rate dependence (Shen et al., 2005; Velardi et al.,
2006). Several investigations have focused on characterizing this
response experimentally (Coats and Marguilies, 2006; Gefen and
Marguilies, 2004; LaPlaca et al., 2005; Miller, 2005; Millerand Chinzei,
Fig. 2. Propagation of blast compressivewave through cranial cavity in 5.2 atm simulation, viewed through mid-coronal plane at 0.132–0.685 ms. The propagation of the blast wave is
from left to right in the external observer frame of reference. Scale is from 0 to 500 kPa.
Parameters for the Tait equation of state.
K (Pa)B (Pa)Γ0
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20
2002; Morrison et al., 2006; Prange and Marguilies, 2002; Velardi
et al., 2006) and on developing a variety of constitutive models to
capture the behavior of the brain as a material (Brands et al., 2004;
Drapaca et al., 2006; Miller, 1999; Shen et al., 2005). Owing to the
complexities and inherent variabilityassociated with biological tissue
(especially the brain where marked regional variation exists), there
is significant uncertainty in quantifying tissue response to material
stresses, particularly at high strain-rates such as those occurring
with blast wave propagation. In consideration of these limitations,
computational models have usually favored simpler (i.e. elastic)
models with few parameters that can be quantified with less
uncertainty, instead of more sophisticated models with many
parameters that are harder to estimate with confidence. In the
impact TBI modeling work, isotropic elastic models (where the
material properties are both elastic and spatially uniform) have
been used to derive the volumetric (bulk modulus) response
whereas linear viscoelastic effects have been considered in the
shear or deviatoric response via a time-dependent shear modulus
evolving from an instantaneous to the long-term value (Belingardi,
2005; Willinger and Baumgartner, 2003; Zhang et al., 2001). In this
context, the shear modulus is the physical parameter determining
the characteristics (speed of propagation, intensity) of shear waves.
These are associated with disturbances in the displacement field
that are transverse to the direction of propagation of the wave.
Such waves are always present in 3-dimensional wave propagation
in solid materials and accompany longitudinal (pressure) wave. The
material relaxation times involved are often on the order of tens to
hundreds of milliseconds and higher for low deformation rates. It
is thus reasonable to expect that deferred deformation or stress
relaxation due to viscoelastic effects play a secondary role under
blast loading, where the characteristic times seldom, if ever, exceed
a few milliseconds. Therefore the linear blast stress wave proper-
ties are on a time-scale of several orders of magnitude less than
where viscoelastic effects may be of significance. However,
potential nonlinear viscoelastic effects may occur with relaxation
times that can be much shorter at blast deformation rates, and,
thus, may be relevant for the analysis. For simplicity and due to the
unavailability of high-rate tissue properties, these effects were
neglected in the model proposed.
For these reasons, and as a first approximation, we have
adopted a simplified constitutive modeling strategy of brain
material properties emphasizing effects pertinent to blast condi-
tions. Specifically, the description of the pressure wave propagating
through the brain was parameterized through suitable equations of
state. To this end, the volumetric response of brain tissue has been
described by the Tait equation of state with parameters adjusted to
fit the bulk modulus of the various tissue types. The Tait equation
allows modeling of material densities over a wide range of
Fig. 3. Pressure histories of selected nodes in the mid-coronal plane. Clockwise from top
left, the nodes are located in the skull, skull and gray matter. The scale is from −0.2 to
Fig. 4. Pressure histories of selected nodes in the mid-sagittal plane. (a) Shows the pressure histories of nodes within the cranial cavity. Clockwise from top right, the nodes are
located in the glia, ventricles, csf, whitematter, gray matter,and venous sinuses. The scale is from −700 to 500 kPa.(b) Shows the pressurehistoriesof nodes on oroutside the cranial
cavity. Clockwise from top right, the nodes are located in the skull, skull, air sinuses, skin/fat, and skull. The scale is from −150 to 300 kPa.
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20
pressure such as might occur during propagation of a blast
shockwave through the brain. The deviatoric response has been
described via a neo-Hookean elastic model with properties
adjusted to fit reported values of the instantaneous shear modulus
to allow modeling of large elastic deformations as may arise under
blast conditions. The Mie-Gruneisen/Hugoniot equation of state
Fig. 5. Maximum overall compressive and tensile pressures for 5.2 atm, 18.6 atm, and 5 m/s impact simulations.
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20
was used to describe the volumetric response of the skull under
high strain rate conditions. The constitutive properties of the
tissues were determined from a literature review. Details about
these constitutive models are summarized in Tables 1 and 2.
Tait and Mie-Gruneisen equations of state
Theshock responseofmanysolidmaterialsiswell describedbythe
Hugoniot relation between the shock wave velocity Us and the
Fig. 6. Maximum overall compressive and tensile pressures reached by differentiated structures in 5.2 atm, 18.6 atm, and 5 m/s impact simulations.
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20
material velocity Up by the simple equation form (1) below
(Drumheller, 1998; Meyers, 1994; Zel'dovich and Raizer, 1967).
Us= C0+ sUp
In this expression, C0and s are material parameters that can be
obtained from experiments. By considering Eq. (1) and conservation
of mass and momentum in a control volume at the shock front, the
final pressure can be calculated explicitly as a function of the Jacobian
behind the shock front JHand the reference densityahead of the shock
ρ0(Drumheller, 1998; Meyers, 1994; Zel'dovich and Raizer, 1967):
where JHis related to the density ρH, the specific volume VH, and
the deformation gradient tensor FH, defined behind the shock front,
= det FH
The relation Eq. (2), also called the “shock Hugoniot,” relates
any final state of density to its corresponding pressure. The
deformation path taken by the material between the initial state
(P0, V0) and the final state (PH, VH) is then defined by a straight
line in the (σ1, V) plot where σ1 is the axial stress in the shock
direction: the Rayleigh line (Drumheller, 1998; Meyers, 1994;
Zel'dovich and Raizer, 1967). The parameters for the Hugoniot/
Mie-Gruneisen equation of state used in the simulations are given
in Table 1. The values for C0and S are the same as those used for
the skull in Taylor and Ford (2006).
The Tait equation of state, which is commonly used to model fluids
under large pressure variations, is given by:
P = B
where B and Γ0are constants. The Tait equation of state provides a
reasonable representation of the volumetric response of soft tissues
embedded in a fluid mediumandwas, therefore,employed to describe
the pressure response of all the head structures except the skull. To
obtain the necessary parameters, Γ0was taken to be the value for
water ∼6.15. Appropriate bulk modulus values K were selected from
the literature, and B was computed for each structure from Γ0and K
using the relation:
K0= B C0+ 1
Table 2 contains the parameters for the Tait equation of state used
in the simulations.
As mentioned above, an important component of a stress wave
relevant in the case of the response of brain tissue to blast as it causes
transverse or shear strains and potential damage to axons. In addition,
increase the potential for injury. This is particularly true in the white
increase tissue vulnerability to shear stresses.
Inordertodescribetheshearwave component, themodelneeds to
account for the deviatoric response. Toward this end, an elastic
deviatoric model was combined with the volumetric equation of state
following a stress-strain relation of the following form (Cuitiño and
Ortiz, 1992; Holzapfel, 2001; Ortiz and Stainer, 1999):
σB= − PI + J−1Feμ log
where Ce=FeTFeis the elastic Cauchy–Green deformation tensor,
log√Ceis the logarithmic elastic strain, μ is the shear modulus, and the
pressure P follows from the equations of state defined in Eqs. (2) and
(4). Explicit formulae for the calculation of the exponential and
logarithmic mappings, and the calculation of their first and second
linearizations, have been given by Ortiz et al. (2001). For the blast and
impact simulations, values of μ were selected from the literature
Fig. 7. Propagation of compressive wave through cranial cavity in 18.6 atm simulation, viewed through mid-coronal plane over a time range of 0.051–0.601 ms. Blast wave is
propagated from left to right in the external frame of reference. Scale is from 0 to 5 Mpa.
Fig. 8. Pressure histories of selected nodes in the mid-coronal plane. The scale is from
−1 MPa to 8 MPa.
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20
The two blast fluid–solid interaction simulations were run on 20
processors, 14 of which were assigned to the solid solver and 6 to the
fluid. This proved sufficient to obtain the results in a reasonable time.
Two levels of grid subdivision were employed in the fluid to resolve
the blast front and the fluid–solid interface with enoughfidelity. In the
simulations, the lower region of the head was fixed where the neck
would ordinarily be attached to the head in order to avoid the blast
engulfing the bottom of the head. The solid-only impact simulation
was run on 20 processors.
5.2 atm simulation
Figs. 2a–e illustrate the propagation of the compressive blast
wave through the coronal sections of the head in the 5.2 atm
simulation. The blast wave is incident on the right temporal region
(radiological convention). The compressive wave is seen propagat-
ing through the cranial cavity from the right to the left with some
minor reflection from the left side of the cavity, leading to a
pocket of concentrated pressure in the skull on the right-hand side
of the head. In the coronal plane, pressure amplitude–time
curvesatindividual nodesin differenttissueregions are
illustrated in Fig. 3. This shows differential and decremental
tissue responses depending on the location with the potential for
significant differential strain even without the development of
transverse or shear waves. Similar effects are seen in the sagittal
plane as illustrated in Figs. 4a and b.
The maximum tensile and compressive pressures and Von Mises
stress for the entire head were then extracted for each time step and
plotted. These curves, which are the envelopes of the pressure and
stress histories of all points within the head, are given in Figs. 5a–f
for blast and impact simulations. The maximum tensile and
compressive pressures were then extracted and plotted at each
time step for each of the 11 distinct structures. These curves are
shown in Figs. 6a and f. Overall, the maximum compressive
pressure reached was 6.5 MPa at 0.00045 s and the maximum
tensile pressure reached was 0.89 MPa at 0.00048 s. The maximum
impact stresses appeared to develop with a more monotonic
quality. The highest compressive pressures were experienced by
the skull and muscle, followed by the subarachnoid CSF. The skull
experienced higher stresses due to its more rigid (higher stiffness)
material properties. The structures can be crudely divided into two
groups: muscle, skull, CSF, gray matter, skin/fat, and air sinuses all
experienced high compressive pressure, while the venous sinuses,
ventricle, glia, white matter, and eyes tissue experienced a differing
Fig. 9. Pressure histories of selected nodes in the mid-sagittal plane. In (a), the scale if from −2.5 MPa to 3.5 MPa. In (b), the scale is from −1.5 MPa to 1.5 MPa.
Fig. 10. Propagation of compressive wave through cranial cavity in 5 m/s impact simulation, viewed through mid-coronal plane. Scale range (0 to 500 kPa).
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20
time course and a lower intensity. The different structures also
experienced peak compressive pressure at different times, with gray
matter intermediate, between the muscle, skull, CSF, and skin/fat.
The structures that experienced the highest tensile pressure were
sinus, gray matter, CSF, skull, and white matter. In this simulation,
the nodes that experienced the highest pressures and stresses were
all located on the right side of the head, in the concentrated pocket
of stress created by the reflection of the blast wave from the left-
side of the head.
18.6 atm simulation
At the LD50blast wave overpressures of ∼18.6 atm, the compres-
sive wave propagation is seen in Fig. 7. Differential stresses are
generated as the wave propagates through the brain tissue. Wave
reflection is more apparent at this overpressure as seen by the
multiphasic pressure curves shown in Figs. 5 and 6. Similar coronal
and sagittal reconstructions at selected nodes are seen in Figs. 8 and 9.
The 18.6 atm simulation reached a maximum compressive pressure of
39 MPa at 0.00042 s in the skull and muscle and a maximum tensile
pressure of 4.5 MPa at 0.00041 s in the gray matter area. High
compressive pressures were experienced by the skull, muscle, CSF,
gray matter, and skin/fat material elements. The skull peaked first,
then the CSF and muscle and finally the gray matter. In the case of the
tensile pressure, the highest pressure values were experienced by the
skull, CSF, and gray matter followed by the white matter. Compared to
the 5.2 atm simulation, the 18.6 atm simulation generated pressures
that were 4–6 times higher than those at 5.2 atm. Further the
pressures and stresses also peaked earlier in the 18.6 atm simulation.
However the locations in the head that experienced the highest
stresses were very similar in both the 5.2 atm and 18.6 atm
The 5 m/s impact simulation ran to 0.000634 s, reaching a
maximum compressive pressure of 27.2 MPa and a maximum tensile
pressure of 7.1 MPa. The impact was delivered in the mid-coronal
plane in a lateral direction, from left to right (Figs. 10–12). Significant
differences are seen between blast and impact injury specifically in
the monotonic form of the impact compression and tension curves.
High compressive pressures were also found in the skull, CSF, muscle,
skin/fat and the gray matter with the highest compressive pressure
being experienced by the skull itself. High tensile pressures were
experienced by the skull, muscle, skin/fat, and gray matter, with the
highest tensile pressures again being experienced by the skull. The
magnitudes of the maximum pressures reached in the impact
simulation are comparable to the maximum pressures reached in
the 18.6 atm simulation. This suggests that blasts that would result in
50% lethality due to unarmored parenchymal lung injury would lead
to concussive injury in the brain since the impact simulation was
based on the known parameters for sports concussive injury (Casson
et al., 2008).
The use of personal protective equipment (PPE), particularly body
armor, is considered to be protective against lethal blast and
Fig. 11. Pressure histories of selected nodes in the mid-coronal plane. Scale is from
−100 kPa to 4.5 MPa.
Fig. 12. Pressure histories of selected nodes in the mid-sagittal plane. In (a), the scale is from −150 kPa to 1.8 MPa. In (b), the scale is from −150 kPa to 80 kPa.
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20
fragmentation lung injury. The goal of this work was to use the
standardized Bowen curves for threshold and 50% lethal lung injury to
examine and compare the effects of similar currently survivable blast
exposure on the head. This was then subsequently compared with
impact decelerations brain injury often seen within sports concus-
sions such as in the National Football League (Casson et al., 2008;
Zhang et al., 2004). The analysis indicated that a blast consistent with
lung LD50i.e. a blast in the pre-PPE era that would result in 50%
lethality from lung injury would also be associated with a concussion
equivalent to impact injury that may result from a sports concussion
or mild TBI (mTBI).
From the model it is observed that direct propagation of blast
waves into the brain occurs, confirming one of the original stated
hypotheses.Further withexposure totheequivalentof 0.324kg TNTat
a 0.6 m standoff distance (18.6 atm overpressure), a similar order of
magnitude stress develops within CNS tissues comparable to those
known to be significant in the development of sport concussive injury
or mTBI. The American Academy of Neurology grades sports
concussion from grades 1–3. Only grade 3 indicates a loss of
consciousness while grades 1 and 2 represent transient confusion or
loss of situational awareness (Quality Standards Committee of the
American Academy of Neurology, 1997). In the military context, such
transient confusion or loss of situational awareness within a combat
environment may slow reaction and response time to adverse events
with potentially devastating results.
The blast–solid interaction simulation technology employed in
this work is well established and has been validated in the case of
the response of engineering materials such as steel and aluminium
plates (Cummings et al., 2002; Kambouchev et al., 2007a). Also, for
the case of impact conditions in which blast is not involved, there
have been efforts to validate this type of simulations (Zhang et al.,
2004). Although there are currently no validated models for the
case of blast effects on the human head, there are current efforts
focusing on the experimental test programs using animal models
that will be used for model validation in an extension of the present
work. This ongoing work will combine field blast tests and a porcine
full head model simulation. The blast events will be monitored
using external pressure sensors together with animal instrumenta-
tion. The number of internal and external pressure experimental
sensors will necessarily be limited and will fall far short of the
theoretical number of sensors required to completely and uniquely
capture the complex 3-D fluid flow field associated with a blast. This
crucially emphasizes the importance and value of simulation work
where the complete dataset is available for interrogation. Both
approaches, the field investigational and the numerical/computa-
tional must be regarded as complementary in enhancing under-
standing of these complex fluid–solid interaction blast events in
biological tissue such as the CNS.
The constitutive model and properties used in the current blast
simulations will be refined as experimental tissue response char-
acterization results become available at blast-relevant strain rates.
However, the models and properties used represent “good first order
approximations” that are extrapolated from impact conditions into
the blast strain-rate domain. The three contexts compared in this
paper are: (a) a peak blast over pressure of 5.2 atm equivalent to free
air explosion of 0.0648 kg TNTat 0.6 m stand-off and corresponding to
the Bowen threshold for unarmored lung parenchymal injury (b) a
peak blast over pressure of 18.6 atm equivalent to free air explosion of
0.324 kg TNT at 0.6 m stand-off and corresponding to the Bowen 50%
lethality for unarmored lung parenchymal injury and (c) impact
deceleration of the full head model from a velocity of 5 m/s to 0 m/s
following impact with a stationary immovable boundary. While other
comparative combinations could have been considered, these speci-
fically provide calculable evidence for concussive brain injury
following primary blast exposure at a specific standoff and TNT
equivalence. The results indicate a potential for concussive effects
from blast under the strict model conditions applied. This suggests, at
a minimum, that such a potential threat deserves further investigation
and consideration. CNS stress waves mayalso be generated by ballistic
injury and the effect of a stress wave generated by ballistic impact
conditions and TBI have been recently discussed (Courtney and
Blast associated TBI is probably rarely isolated or primary but more
likely to be combined with secondary shrapnel or fragment injury, or
tertiary blast injuryduetoimpactsuch asoccurswithvehicle roll-over
or hitting a stationary wall (Warden et al., 2009). The biological
independent and synergistic effects of detonation products on brain
injury cannot be excluded and are part of an ongoing scientific
investigation scientific investigation (Moore et al., 2008). The model
described in this paper can also be used to explore such multi-physics
effects. Other areas of current investigation are examination of the
relative contribution of direct pressure effects and indirect transmis-
sion of pressure effects on the brain through the great vessels (Cernak
et al., 1999; Moore et al., 2008).
It is hoped that a direct application of the modeling presented in
this paper will be to shorten the design cycle of engineering
modifications for the development of personal protective equipment.
Currently head PPE tends to be optimized for impact or ballistic
protection with little or no consideration for blast mitigation or
protection. The complexity of issues required to consider in engineer-
ing optimization of head PPE should not be underestimated and may
further represent an avenue where models such as the one developed
in the current work could substantially contribute.
Conflict of interest statement
The authors declare that there are no conflicts of interest.
This work was supported in full by financial aid from the Joint
Improvised Explosive Device Defeat Organization (JIEDDO) through
the Army Research Office. We also thank the Editorial Staff of
NeuroImage for the helpful critique and comment on this manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.neuroimage.2009.02.019.
Bass, C., et al., 2006. Pulmonary Injury Risk Assessment For Short-Duration Blasts.
Proceedings Of The Personal Armour Systems Symposium (Pass). Leeds, UK.
Belingardi, G., Chiandussi, G., Gaviglio, I., 2005. Development And Validation Of A New
Finite Element Model Of Human Head. 19th International Technical Conference on
the Enhanced Safety of Vehicles Conference, Washington, D.C.
Bowen, I., Fletcher, E.R., Richmond, D.R., Hirsch, F.G., White, C.S., 1968. Biophysical
mechanisms and scaling procedures applicable in assessing responses of the thorax
energized by air-blast overpressures or by nonpenetrating missiles. Ann. N.Y. Acad.
Sci. 152, 122–146.
Brands, D., Peters, G.W.M., Bovendeerd, P.H.M., 2004. Design and numerical imple-
mentation of a 3-D non-linear viscoelastic constitutive model for brain tissue
during impact. J. Biomech. 37, 127–134.
Brazarian, J., McClung, J., Shah, M., Cheng, Y., Flesher, W., Kraus, J., 2005. Mild traumatic
brain injury in the United States. Brain Inj. 18, 85–91.
Bruns, J., Hauser, A.W., 2003. The epidemiology of traumatic brain injury: a review.
Epilepsia 44, 2–10.
Casson, I., Pellman, E.J., Viano, D.C., 2008. Concussion in the national football league: an
overview for neurologists. Neurol. Clin. 26, 217–241.
Cernak, I., Savic, J., Ignjatovic, D., Jevtic, M.,1999. Blast injury from explosive munitions.
J. Trauma 47, 96–103 discussion 103–104.
Coats, B., Margulies, S.S., 2006. Material properties. Biomechanics 39, 2521–2525.
Collins, D., Zijdenbos, A.P., Kollokian, V., et al., 1998. Design and construction of a
realistic digital brain phantom. IEEE Trans. Med. Imag. 17, 463–469.
Courtney, A., Courtney, M., 2007. Links between traumatic brain injury and ballistic
pressure waves originating in the thoracic cavity and extremities. Brain Inj. 21.
Cuitiño, A., Ortiz, M.,1992. A material-independent method for extending stress update
algorithms from small-strain plasticity to finite plasticity with multiplicative
kinematics. Eng. Comput. 9, 437–451.
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20
Cullis, I., 2001. Blast waves and how they interact with structures. J. R. Army Med. Corps Download full-text
Cummings, J., Aivazis, M., Samtaney, R., Radovitzky, R., Mauch, S.D., Meiron, D., 2002. A
virtual test facility for the simulation of dynamic response in materials. J.
Supercomput. 23, 39–50.
Deiterding, R., Radovitzky, R., Mauch, S.P., Noels, L., Cummings, J.C., Meiron, I., 2006. A
virtual test facility for the efficient simulation of solid material response under
strong shock and detonation wave loading. Eng. Comput. 22, 325–347.
DePalma, R., Burris, D.G., Champion, H.R., Hodgson, M.J., 2005. Blast injury. N. Eng. J.
Med. 353, 1335–1342.
Drapaca, C., Tenti, G., Rohlf, K., Sivaloganathan, S., 2006. A quasi-linear viscoelastic
constitutive equation for the brain:application to hydrocephalus. J. Elast. 85, 65–83.
Drumheller, D.,1998. Introduction To Wave Propagation In Nonlinear Fluids And Solids.
Cambridge University Press, Cambridge.
Elsayed, N., 1997. Toxicology of blast overpressure. Toxicology 121, 1–15.
Gefen, A., Margulies, S.S., 2004. Are in vivo and in situ brain tissues mechanically
similar? J. Biomech. 37, 1339–1352.
Gilchrist, M., O'Donoghue, D., Horgan, T., 2001. A two-dimensional analysis of the
biomechanics of frontal and occipital headimpact injuries. Int.J. Crashworthiness 6,
Holzapfel, G.A., 2001. Nonlinear Solid Mechanics: A Continuum Approach For
Engineering. John Wiley & Sons Ltd, Chichester, England.
Horgan, T., Gilchrist, M.D., 2004. Influence of FE model variability in predicting brain
motion and intracranial pressure changes in head impact simulations. Int. J.
Crashworthiness 9, 401–418.
Iverson, G., Lange, R.T., Gaetz, M., Zasler, N.D., 2006. Mild TBI. In: Zasler, N, Katz, DI,
Zafonte, RD (Eds.), Brain Injury Medicine. Demos, New York.
Kambouchev, N., Noels, L., Radovitzky, R., 2006. Nonlinear compressibility effects in
fluid-structure interaction and their implications on the air-blast loading of
structures. J. Appl. Phys. 100, 063511–063519.
Kambouchev, N., Noels, L., Radovitzky, R., 2007a. Numerical simulation of the fluid–
structure interaction between air blast waves and free-standing plates. Comput.
Struct. 23, 325–347.
Kambouchev, N., Radovitzky, R., Noels, L., 2007b. Fluid–structure interaction effects in
the dynamic response of free-standing plates to uniform shock loading. J. Appl.
Mech. 74, 1–5.
Kolsky, H., 1963. Stress Waves in Solids. Dover Publications, New York.
LaPlaca, M., Cullen, D.K., McLoughlin, J.J., Cargill, R.S., 2005. High rate shear strain of
three-dimensional neural cell cultures: a new in vitro traumatic brain injury model.
J. Biomech. 38, 1093–1105.
Mayorga, M.A., 1997. The pathology of primary blast overpressure injury. Toxicology
Meyers, M., 1994. Dynamic behavior of materials. Wiley Interscience, New York.
Miller, K., 1999. Constitutive model of brain tissue suitable for finite element analysis.
J. Biomech. 32, 531–537.
Miller, K., Chinzei, K., 2002. Mechanical properties of brain tissue in tension. J. Biomech.
Moore, D.F., Radovitzky, R.A., Shupenko, L., Klinoff, A., Jaffee, M.S., Rosen, J.M., 2008.
Blast physics and central nervous system injury. Future Neurol. 3, 243–250.
Morrison, B., Cater, H.L., Benham, C.D., Sundstrom, L.E., 2006. An in vitro model of
traumatic brain injury utilising two-dimensional stretch of organotypic hippo-
campal slice cultures. J. Neurosci. Methods 150, 192–201.
Ortiz, M., Stainier, L., 1999. The variational formulation of viscoplastic constitutive
updates. Comput. Methods Appl. Mech. Eng. 171, 419–444.
Ortiz, M., Radovitzky, R.A., Repetto, E.A., 2001. The computation of the exponential and
logarithmic mappings and their first and second linearizations. Int. J. Numer.
Methods Eng. 52, 1431–1441.
Petras,J., Bauman, R.A., Elsayed, N.M.,1997. Visual system degeneration induced byblast
overpressure. Toxicology 121, 41–49.
Prange, M., Margulies, S.S., 2002. Regional, directional, and age-dependent properties of
the brain undergoing large deformation. J. Biomech. Eng. 124, 244–252.
Quality Standards Committee of the American Academy of Neurology, 1997. Practice
parameter: the management of concussion in sports. Neurology 48, 581–585.
Raul, J., Baumgartner, D., Willinger, R., Ludes, B., 2006. Finite element modelling of
human head injuries caused by a fall. Int. J. Leg. Med. 120, 212–218.
Saraf, H., Ramesh, K.T., Lennon, A.M., Merkle, A.C., Roberts, I.C., 2007. Mecha-
nical properties of soft human tissues under dynamic loading. J. Biomech. 40,
Shen, F., Tay, T.E., Li, J.Z., Nigen, S., 2005. Modified Bilston nonlinear viscoelastic model
for finite element head injury studies. J. Biomech. Eng. 128, 797–801.
Strehlow, R., Baker, W.E., 1976. The characterization and evaluation of accidental
explosions. Pror. Energy and Combust. Sci. 2, 27–60.
Stuhmiller, J., 1987. Biological response to blast overpressure: a summary of modeling.
Toxicology 121, 91–103.
Tanielian, T.L., 2008. In: Tanielian, T., Jaycox, L.H. (Eds.), Invisible Wounds Of War:
Psychological And Cognitive Injuries, Their Consequences, And Services To Assist
Recovery. RAND Corporation, Center for Military Health Policy Research, Santa
Taylor, P., Ford, C.C., 2006. Simulation of head impact leading to traumatic brain injury
Army Science Conference Orlando, FL.
Velardi, F., Fraternali, F., Angelillo, M., 2006. Anisotropic constitutive equations and
Warden, D.L., French, L.M., Shupenko, L., Fargus, J., Reidy, G., Erickson, M.E., Jaffee, M.S.,
Moore, D.F., 2009. Case report of a soldier with primary blast brain injury.
Neuroimage. 47, T153–T154.
White, C., Jones, R.K., Damon, E.G., Fletcher, E.R., Richmond, D.R.,1971. The biodynamics
of airblast. In: T.R. (Ed.), Defense Nuclear Agency. Washington, D.C.
Willinger, R., Baumgartner, D., 2003. Numerical and physical modelling of the human
head under impact-towards new injury criteria. Int. J. Veh. Des. 32, 94–115.
Zel'dovich, T., Raizer, Y.P., 1967. Physics of shock waves and high-temperature
hydrodynamic phenomena. Academic Press, New York.
Zhang, L., Yang, K.H., Dwarampudi, R., Omori, K., Li, T., Chang, K., Hardy, W.N., Khalil, T.,
King, A.I., 2001. Recent advances in brain injury research: a new human head model
development and validation. Stapp Car Crash J. 45, 369–394.
Zhang, L., Yang, K.H., King, A.I., 2004. A proposed injury threshold for mild traumatic
brain injury. J. Biomech. Eng. 126, 226–236.
D.F. Moore et al. / NeuroImage 47 (2009) T10–T20