A Spin Glass Based Framework to Untangle Fiber Crossing in MR Diffusion Based Tracking

We propose a general approach to the reconstruction of brain white matter geometry from diffusion-weighted data. This approach is based on an inverse problem framework. The optimal geometry corresponds to the lowest energy configuration of a spin glass. These spins represent pieces of fascicles that orient themselves according to diffusion data and interact in order to create low curvature fascicles. Simulated diffusion-weighted datasets corresponding to the crossing of two fascicle bundles are used to validate the method.

• Type: Conference
• Year: 2002
• Where: MICCAI
• Authors: Yann Cointepas, Cyril Poupon, Denis Le Bihan, Jean-Francois Mangin
• Keywords: Low Curvature Fascicles, Medical Imaging, MICCAI 2002, Optimal Geometry Corresponds, White Matter Geometry