126
K. H. Yang and H. Mao
the skull-brain interface, the complex 3D human brain vasculature, which includes
arteries, arterioles, capillaries, venules, and veins, has not been fully represented in
a human head model. In addition, there are efforts being made to represent axonal
fibres inside the brain. However, the application and validity in predicting axonrelated brain damage need to be further investigated, and better representation of
axonal fibres inside the brain is needed. As such, future improvements to human
head models will likely take the form of continued additions of more details.
However, with these added details, and the same requirements of computational
stability, efficiency, and comprehensive validation, the costs of developing a future
head model with these improvements could be huge. Hence, collaboration among
researchers to develop and share a high-quality head model is much preferred for
generating high-quality brain-biomechanics data for the entire research community.
With the computational human head model being available to predict head
responses, the following need is about injury assessment functions that can be
used to interpret model predictions and predict risks of real-world head injury.
With the understanding that cadaver heads do not have biological consequences and
do not represent most brain injuries, studies from live human beings are needed.
One valid method is to use a computational human head model to reconstruct
impacts of live human subjects and then test and evaluate several injury assessment
functions to investigate whether these functions can be used to predict risks that the
human subjects experience after trauma. While many studies have demonstrated that
strain-based injury assessment functions can predict brain injury, one recent work
demonstrated that the maximum principal strain-based injury evaluation functions
predicted about 80 concussions and 14 cases of diffuse axonal injury out of 335
noninjury cases [86]. Such results highlight the great needs for improving the
specificity of any injury prediction functions in the future.
The human brain is a biological organ that reacts to impacts at various levels
(macro, micro, cellular, and molecular levels). A traditional human head model
can provide stresses and strains of neural tissues, but it lacks in detail in terms
of how neural structures respond. Hence, finer scale models to predict neural
responses, such as Ranvier node strains during axonal stretch [114], would be
beneficial. In addition to the various levels, the human brain is a biological organ
that has bioelectrical and biochemical activations due to impact. The mechanisms
of these activations can be better studied with a multi-physics model that couples
strains/stresses to bioelectrical and biochemical equations. In addition, the brain is
filled with fluids, such as blood flow in the vessels, CSF in the ventricles, and fluids
in the extracellular spaces. The fluids’ movements due to trauma and disease, plus
movement-induced shearing loads to nearby vascular and neuronal cells, may cause
damage. Overall, multi-scale, multi-physics computational head models, with the
FE method being an integral part, will be largely needed in the future. Collaborations
across disciplines will be more important in the future for brain injury-related
studies.
K. H. Yang and H. Mao
the skull-brain interface, the complex 3D human brain vasculature, which includes
arteries, arterioles, capillaries, venules, and veins, has not been fully represented in
a human head model. In addition, there are efforts being made to represent axonal
fibres inside the brain. However, the application and validity in predicting axonrelated brain damage need to be further investigated, and better representation of
axonal fibres inside the brain is needed. As such, future improvements to human
head models will likely take the form of continued additions of more details.
However, with these added details, and the same requirements of computational
stability, efficiency, and comprehensive validation, the costs of developing a future
head model with these improvements could be huge. Hence, collaboration among
researchers to develop and share a high-quality head model is much preferred for
generating high-quality brain-biomechanics data for the entire research community.
With the computational human head model being available to predict head
responses, the following need is about injury assessment functions that can be
used to interpret model predictions and predict risks of real-world head injury.
With the understanding that cadaver heads do not have biological consequences and
do not represent most brain injuries, studies from live human beings are needed.
One valid method is to use a computational human head model to reconstruct
impacts of live human subjects and then test and evaluate several injury assessment
functions to investigate whether these functions can be used to predict risks that the
human subjects experience after trauma. While many studies have demonstrated that
strain-based injury assessment functions can predict brain injury, one recent work
demonstrated that the maximum principal strain-based injury evaluation functions
predicted about 80 concussions and 14 cases of diffuse axonal injury out of 335
noninjury cases [86]. Such results highlight the great needs for improving the
specificity of any injury prediction functions in the future.
The human brain is a biological organ that reacts to impacts at various levels
(macro, micro, cellular, and molecular levels). A traditional human head model
can provide stresses and strains of neural tissues, but it lacks in detail in terms
of how neural structures respond. Hence, finer scale models to predict neural
responses, such as Ranvier node strains during axonal stretch [114], would be
beneficial. In addition to the various levels, the human brain is a biological organ
that has bioelectrical and biochemical activations due to impact. The mechanisms
of these activations can be better studied with a multi-physics model that couples
strains/stresses to bioelectrical and biochemical equations. In addition, the brain is
filled with fluids, such as blood flow in the vessels, CSF in the ventricles, and fluids
in the extracellular spaces. The fluids’ movements due to trauma and disease, plus
movement-induced shearing loads to nearby vascular and neuronal cells, may cause
damage. Overall, multi-scale, multi-physics computational head models, with the
FE method being an integral part, will be largely needed in the future. Collaborations
across disciplines will be more important in the future for brain injury-related
studies.