Spinal cord injury (SCI) is characterised by permanent loss of motor and sensory function. The primary damage from the initial mechanical insult is exacerbated by the secondary patho-physiological cascade. Research into neuroprotective interventions to preserve tissue and reduce the damage caused by the secondary injury is hampered, in part, due to a lack of understanding of the link between the biomechanics of the primary traumatic injury and the subsequent evolution of the secondary injury. Hence, there is a need to better understand the biomechanics of SCI, the distinct injury patterns produced, and how these affect the evolution of the secondary cascade.
Computational models using finite element methods (FEM) have been established as a useful tool for investigating SCI biomechanics. These may be used to obtain data that is difficult or impossible to capture through in vivo and in vitro experiments, in particular; stress and strain fields within the neural tissue. However, the complexity of these models is limited by difficulties. These include: problems coping with large deformations over short periods of time due to mesh tangling, difficulties in incorporating the fluid structure interactions, and scalability issues when attempting to make use of high performance computing facilities, utilising large numbers of processors.
This work has involved the creation of a computational spinal cord injury using the Material Point Method (MPM) and MPMICE (MPM for Implicit, Continuous Fluid, Eulerian), alternative computational methods that overcome these limitations. The model incorporates the neural spinal cord tissue, the dura mater, and the cerebrospinal fluid. This model has been validated against equivalent experimental and FEM results. MPM/MPMICE was found to be a viable alternative to FEM for modelling SCI computationally, with the potential to enable more complex and anatomically detailed models through the utilisation of increased parallel computation.
