Diffusion Modeling using Bayesian Probability Theory applied to imaging of Multiple Sclerosis
Kroenke C, Sheline Y, Cross A, Neil J, Epstein A, Shimony J, Bretthorst G, Snyder A
Washington University Medical School
The diffusion tensor model has proven extremely valuable for modeling the diffusion properties of water in the brain. The model incorporates a monoexponential function for the diffusion attenuation curve. Since the introduction of this model studies from several research groups have shown that this function does not fit the data well at high gradient values. As a result many researchers have applied a biexponential function to the data. We explore a simple alternative model using Bayesian probability theory that provides a unique contrast mechanism and apply it to normal subjects and patients with multiple sclerosis.