Spinal cord injury is one of the most devastating medical conditions known. These injuries result from a variety of spinal column injury mechanisms;​ however, the mechanical relationship between column and cord injuries is not known. The objectives of this research project were to develop and validate a finite element model of the spinal cord and the surrounding structures and to use this model to compare the strain distributions in the spinal cord for three simple injury mechanisms:​ compression (contusion), distraction, and dislocation. The final goal was to compare linear elastic and hyperelastic constitutive models for the spinal cord in response to distraction loading.

Geometry of the model was obtained from the Visible Human Project of the National Library of Medicine and developed into a three-dimensional FE model in Ansys (Ansys User's Manual version 7.1). Three vertebrae (C4-6) and the spinal cord were modelled with brick elements, dura mater was modelled with shell elements, and ligaments and discs were modelled with cable elements. Material properties for the tissues were based on experimental data in the literature. Boundary conditions and loading were designed to simulate experimental data in the literature. The model was transferred to Abaqus (Abaqus User's Manual version 6.3 2002) for the comparison of constitutive models.

The model of spinal cord compression was validated with previous experimental data based on the reaction force at the indentor tip (Hung, Lin et al. 1979;​ Hung, Lin et al. 1982;​ Tencer, Allen et al. 1985). The distraction injury model was validated based on relative displacements of the column and cord (Maiman, Coats et al. 1989). No experimental data was available for validation of the dislocation injury mechanism.

Different strain distributions were found for compression, distraction, and dislocation injury mechanisms. The highest strains within the compression injury mechanism were in the dorsal, ventral, and central columns with critical regions in the dorsomedial and ventral white matter between the indentor and the opposing vertebral body. Strains for distraction were more uniformly distributed throughout the cord. The dislocation injury mechanism seemed to spare the lateral columns of normal strains, however all columns experienced elevated shear strains. Critical regions of strain were in the dorsolateral and ventral white matter at the level of the contacting lamina and vertebral body. Linear elastic and hyperelastic constitutive models produced similar strain distributions (difference of less than 1% strain between each maximum component strain) for axial loading of the spinal cord up to 8% axial strain. Assuming the hyperelastic model best represents the behaviour of spinal cord tissue, the linear elastic model overestimated the range of axial strains and underestimated the dorsal-ventral and transverse shear strains for axial strains between 17 and 33%.

The compression results reveal the most extensive damage peripherally, which has not been found experimentally. This highlights the possibility of a mechanical or biological susceptibility of grey matter. Results from this study suggest that dislocation injuries may result in more extensive shear strains than burst fracture injuries, which apply onesided contusion to the cord. Extensive distraction of the spinal cord resulted in substantial shear strains. For static loading up to 8% strain, the linear elastic model may be used for FE modeling of the spinal cord.