Characteristics:

• Based on in vivo patient data
• Segmentation tools to aid separation of multiple, interconnecting parts
• 15 structures segmented including hard and soft tissues
• Highly complex, robust, accurate mesh of multiple regions
• FE analysis performed directly on generated mesh using LS-Dyna

Image Processing

       A high resolution T1-weighted MRI scan of a healthy 26yr old male was obtained in vivo. 15 structures were segmented, including; brain (gray + white matter, brain stem, cerebellum), CSF, skull, mandible, cervical vertebrae, intervertebral discs, eye (eyeball, optic nerve, fatty tissue), nasal passage and skin. Semi-automated and manually driven segmentation tools enabled the complex, interlinked structures to be identified.

Mesh Generation

      A number of finite element meshes were produced directly from the segmentation. The meshing parameters were adjusted to produced models ranging in size from 2-10 million elements. Each had high element quality (max aspect ratio of 10), conforming interfaces between parts and coincident nodes and elements across boundaries. The exterior surface of the skin was used to define a contact surface.

Simulation

     The mesh was exported to LS-DYNA® (LSTC, Livermore, USA). An impactor was introduced with a velocity of 7m/s, a mass of 6.8kg and an event duration 15ms. The brain region was set as a viscoelastic material, the CSF as an elastic fluid and everything else was an elastic material. Pressure transients through the brain were observed, as were the Von Mises stresses produced in the intervertebral discs and cervical vertebrae.