102
K. H. Yang and H. Mao
some 58 cases reconstructed from National Football League games with 22–25
among them being concussion cases [77, 97], a total of 300,977 sub-concussive and
57 concussive head impacts head acceleration data [83], 13 cases of reconstructed
motorcycle data [50, 64], and four sets of graded AIS scale head injury data derived
from real-world car crashes [16].
With ever-increasing computational power, more and more human head models
are being reported. Several institutions have devoted their resources in creating
better head models. Mao et al. [62] at WSU partially validated their head model
against experimental data obtained from 35 different loading cases. The developed
and validated model was then used to determine the maximum principal strain
(MPS) and other parameters for studying the injury mechanism and tolerance
for brain contusion. This model is currently available through the Global Human
Body Modelling Consortium (GHBMC). Takhounts et al. [92] at the National
Highway Traffic Safety Administration (NHTSA) developed a simulated injury
monitor (SIMon) head model and associated injury criteria known MPS, cumulative
strain damage measure (CSDM), brain injury criteria (BrIC), etc.
Kleiven [50] at the Royal Institute of Technology (Kungliga Tekniska Högskolan,
KTH) in Sweden found it necessary to introduce different constants needed by the
Ogden material law in order to properly reflect the tension-compression asymmetry
reported by Franceschini et al. [14] and Miller and Chinzei [67, 68]. The KTH
group also reported a head model in which fractional anisotropy (FA) calculated
from diffusion tensor imaging (DTI) was averaged to obtain the ‘axonal orientation’
for each element and to implement element-specific anisotropic properties [21].
Additionally, Giordano and Kleiven [22] evaluated the ‘axonal strain’ and found
that this response variable had a much higher area under the receiver operating
characteristic (ROC) curve than other response variables. While these results
sounded promising, the inherited deficiencies of using a coarse-mesh FE model
(about 7000 elements for the whole brain in this KTH model) to address the
directional dependency of white matter are further discussed in Sect. 5.3.1 of this
chapter. It is worthy mentioned that Ho and Kleiven [34] developed two fine-mesh
FE models with an average element size of 1 mm that was quite different from other
coarse-mesh models reported from KTH. The authors recommended that future FE
head models should include sulci as it altered the strain distribution in an FE model.
Perhaps this fine-mesh model was computationally too expensive, this suggestion
was not followed in other papers came from the same group.
Sahoo et al. [85] at the University of Strasbourg discussed the use of their
relatively coarse-mesh model (about 13,000 elements or 4–10 mm typical element
size for the whole head) to study the direction-dependent axonal strain so that the
risk of sustaining diffuse axonal injury (DAI) could be predicted. Ji et al. [40] at
Dartmouth College developed a Dartmouth Head Injury Model (DHIM) from a
concussed athlete that included consideration of axonal directions. The same group
then used it to pre-compute a number of potential impact scenarios to determine the
model-predicted responses rapidly [41]. As a side note, Dr. Ji has since moved to
Worcester Polytechnic Institute (WPI) to continue his brain modelling efforts.
K. H. Yang and H. Mao
some 58 cases reconstructed from National Football League games with 22–25
among them being concussion cases [77, 97], a total of 300,977 sub-concussive and
57 concussive head impacts head acceleration data [83], 13 cases of reconstructed
motorcycle data [50, 64], and four sets of graded AIS scale head injury data derived
from real-world car crashes [16].
With ever-increasing computational power, more and more human head models
are being reported. Several institutions have devoted their resources in creating
better head models. Mao et al. [62] at WSU partially validated their head model
against experimental data obtained from 35 different loading cases. The developed
and validated model was then used to determine the maximum principal strain
(MPS) and other parameters for studying the injury mechanism and tolerance
for brain contusion. This model is currently available through the Global Human
Body Modelling Consortium (GHBMC). Takhounts et al. [92] at the National
Highway Traffic Safety Administration (NHTSA) developed a simulated injury
monitor (SIMon) head model and associated injury criteria known MPS, cumulative
strain damage measure (CSDM), brain injury criteria (BrIC), etc.
Kleiven [50] at the Royal Institute of Technology (Kungliga Tekniska Högskolan,
KTH) in Sweden found it necessary to introduce different constants needed by the
Ogden material law in order to properly reflect the tension-compression asymmetry
reported by Franceschini et al. [14] and Miller and Chinzei [67, 68]. The KTH
group also reported a head model in which fractional anisotropy (FA) calculated
from diffusion tensor imaging (DTI) was averaged to obtain the ‘axonal orientation’
for each element and to implement element-specific anisotropic properties [21].
Additionally, Giordano and Kleiven [22] evaluated the ‘axonal strain’ and found
that this response variable had a much higher area under the receiver operating
characteristic (ROC) curve than other response variables. While these results
sounded promising, the inherited deficiencies of using a coarse-mesh FE model
(about 7000 elements for the whole brain in this KTH model) to address the
directional dependency of white matter are further discussed in Sect. 5.3.1 of this
chapter. It is worthy mentioned that Ho and Kleiven [34] developed two fine-mesh
FE models with an average element size of 1 mm that was quite different from other
coarse-mesh models reported from KTH. The authors recommended that future FE
head models should include sulci as it altered the strain distribution in an FE model.
Perhaps this fine-mesh model was computationally too expensive, this suggestion
was not followed in other papers came from the same group.
Sahoo et al. [85] at the University of Strasbourg discussed the use of their
relatively coarse-mesh model (about 13,000 elements or 4–10 mm typical element
size for the whole head) to study the direction-dependent axonal strain so that the
risk of sustaining diffuse axonal injury (DAI) could be predicted. Ji et al. [40] at
Dartmouth College developed a Dartmouth Head Injury Model (DHIM) from a
concussed athlete that included consideration of axonal directions. The same group
then used it to pre-compute a number of potential impact scenarios to determine the
model-predicted responses rapidly [41]. As a side note, Dr. Ji has since moved to
Worcester Polytechnic Institute (WPI) to continue his brain modelling efforts.