The biomechanics of the spinal cord during traumatic spinal cord injury

Cervical spinal cord injury (SCI), although not a very common injury, carries a very high socioeconomic cost in modern society. This motivates a detailed study of the biomechanical factors associated with cervical SCI, in order to formulate tolerance criteria which may be used to evaluate protective safety devices for both automotive and recreational applications. This research project consists of four interrelated studies, which examine the mechanical properties, kinematics and dynamics of the cervical spinal cord and spinal column during normal and injurious motions. The first of these characterized the tensile elastic and time dependent mechanical properties of the human cervical spinal cord in vitro. The spinal cord was found to be much stiffer than brain tissue (E$\sb{\rm av}$ = 1.0-2.0MPa), and material was found to have a non-linearly elastic stress-strain behaviour, with stiffness increasing with increasing applied strain. In addition, the relaxation behaviour indicates that the spinal cord may be expected to behave essentially elastically at strain rates in excess of approximately 2-5s$\sp{-1}$. The second study is an MRI study which characterized the axial motion and stretch of the spinal cord, in vivo, during normal flexion and extension motions, and the displacements and rotations of the individual vertebrae relative to T1 during the same maneuvers. It was found that quasistatic stretches of up to twenty percent and axial displacements of up to 10 mm relative to the adjacent vertebrae occur in the cervical cord region. These data, together with the mechanical properties of the spinal cord, were used in the construction and validation of an anatomically and mechanically accurate physical model of the head and neck, which includes a surrogate spinal cord and brain material. This model was loaded dynamically to simulate hyperflexion and hyperextension injuries, while high speed video captured the deformations of a grid embedded within the surrogate spinal cord tissue. Strains of ten to twenty percent at strain rates of up to 20 s$\sp{-1}$ were seen in the mid cervical region in the worst case (hyperflexion). Previous isolated tissue data suggests that axons exposed to this level of loading may be reversibly or irreversibly damaged. The fourth and final study involved the construction and validation of a finite element model of the cervical spinal cord and brain tissues. This model was used to investigate the effect on the spinal cord deformation fields of loading and material parameters. It was found that an increase in the peak acceleration, or a decrease in the pulse duration had the effect of increasing the strain rates, but leaving the peak strains relatively unchanged. A decrease in cord stiffness resulted in a small increase in the cord strain, while the use of a hyperelastic material model instead of a linear elastic model reduced the cord strains. Together, these four studies form a detailed biomechanical investigation of hyperflexion and hyperextension injuries to the cervical spinal cord, and suggest that in traumatic cervical spinal cord injury sufficiently large tensile deformations occur at high strain rates in the cord to account for the loss of neurological function seen, even in the absence of vertebral fractures