5 Modelling of the Brain for Injury Simulation and Prevention
101
5.1.3 Finite Element Modelling of the Head and Brain
Many researchers agree that injury mechanisms and associated thresholds based on
externally measured linear and/or angular acceleration do not address the underlying
biomechanical basis for injury. A number of studies have pointed out that brain
deformation or strain is a principal cause of injury. Unfortunately, measuring strain,
particularly in vivo, is almost impossible during an impact. At present, the best
method for predicting intracranial biomechanical responses is through numerical
modelling. In particular, the FE method is exclusively suitable to model structures
of irregular geometries, multiple material compositions, and complex loading and
boundary conditions. The FE method has been the preferred method to study head
injury since the last decade. Numerical models developed using the FE method can
provide tissue-level responses for correlation with the location and severity of injury
outcomes. Considering the fact that globally, kinematic-based injury criteria have
yet to bring about a further reduction in the incidence rate of TBI, it is perhaps
necessary to seek criteria that are based on brain response, using FE computational
models. In this chapter, emphases will be placed on developing and using FE models
to better understand the injury mechanism.
A number of human and animal FE brain models have been reported since the
1980s when computers became powerful enough to run these models. The general
belief of these studies is that a fully validated human FE head model is needed
to identify the injury mechanism, which in turn can be used as a surrogate to
design countermeasures to mitigate head injury severity or eliminate head injury
altogether. Because traffic-induced TBI accounted for a majority of the TBI cases,
many of these numerical models were published in the Stapp Car Crash Conference
Proceedings or Journals. Yang et al. [103] conducted a comprehensive review of
these models on the occasion of the 50th anniversary of the Conference. It was
concluded that even though significant insight into head injury biomechanics was
attained through exercising these models, there was a glaring lack of human brain
material properties under loading rates relevant to impact-induced TBI. Until now,
most experiments conducted to depict brain properties were conducted at quasistatic conditions, while a limited number of studies were tested at high strain rates
(over 100 s −1 ) on nonhuman brain tissues. For example, Prabhu et al. [80] tested
porcine brain tissues at a strain rate between 50 and 750 s −1 , while Pervin and Chen
[78] tested bovine brains at 1000, 2000, and 3000 s −1 . High-rate material properties
acquired from human brain tissues are rare. For medium-rate testing, Jin et al. [44]
reported tensile, compressive, and shear properties obtained from human cadavers
at rates of up to 30 s −1 .
Additionally, experimental and real-world data needed to properly validate these
numerical models were in short supply. To date, the only cadaveric data available
consist of three sets of intracranial pressure data reported by Nahum et al. [71],
Trosseille et al. [95], and Hardy et al. [30] and a few sets of relative motion data
between the brain and skull reported by Hardy et al. [30, 31]. There are also live
human data available for various extents of model validations. They consisted of
101
5.1.3 Finite Element Modelling of the Head and Brain
Many researchers agree that injury mechanisms and associated thresholds based on
externally measured linear and/or angular acceleration do not address the underlying
biomechanical basis for injury. A number of studies have pointed out that brain
deformation or strain is a principal cause of injury. Unfortunately, measuring strain,
particularly in vivo, is almost impossible during an impact. At present, the best
method for predicting intracranial biomechanical responses is through numerical
modelling. In particular, the FE method is exclusively suitable to model structures
of irregular geometries, multiple material compositions, and complex loading and
boundary conditions. The FE method has been the preferred method to study head
injury since the last decade. Numerical models developed using the FE method can
provide tissue-level responses for correlation with the location and severity of injury
outcomes. Considering the fact that globally, kinematic-based injury criteria have
yet to bring about a further reduction in the incidence rate of TBI, it is perhaps
necessary to seek criteria that are based on brain response, using FE computational
models. In this chapter, emphases will be placed on developing and using FE models
to better understand the injury mechanism.
A number of human and animal FE brain models have been reported since the
1980s when computers became powerful enough to run these models. The general
belief of these studies is that a fully validated human FE head model is needed
to identify the injury mechanism, which in turn can be used as a surrogate to
design countermeasures to mitigate head injury severity or eliminate head injury
altogether. Because traffic-induced TBI accounted for a majority of the TBI cases,
many of these numerical models were published in the Stapp Car Crash Conference
Proceedings or Journals. Yang et al. [103] conducted a comprehensive review of
these models on the occasion of the 50th anniversary of the Conference. It was
concluded that even though significant insight into head injury biomechanics was
attained through exercising these models, there was a glaring lack of human brain
material properties under loading rates relevant to impact-induced TBI. Until now,
most experiments conducted to depict brain properties were conducted at quasistatic conditions, while a limited number of studies were tested at high strain rates
(over 100 s −1 ) on nonhuman brain tissues. For example, Prabhu et al. [80] tested
porcine brain tissues at a strain rate between 50 and 750 s −1 , while Pervin and Chen
[78] tested bovine brains at 1000, 2000, and 3000 s −1 . High-rate material properties
acquired from human brain tissues are rare. For medium-rate testing, Jin et al. [44]
reported tensile, compressive, and shear properties obtained from human cadavers
at rates of up to 30 s −1 .
Additionally, experimental and real-world data needed to properly validate these
numerical models were in short supply. To date, the only cadaveric data available
consist of three sets of intracranial pressure data reported by Nahum et al. [71],
Trosseille et al. [95], and Hardy et al. [30] and a few sets of relative motion data
between the brain and skull reported by Hardy et al. [30, 31]. There are also live
human data available for various extents of model validations. They consisted of