Comparison of diffusion MRI models of the brain microstructure using the generalization error

In this document, we first present the theoretical background needed to understand diffusion-weighted magnetic resonance imaging. In a second time, we describe in detail the four chosen models and expose their limitations and advantages. Then, the metric used to compare the models, called the generalization error, is presented. It turns out that this error needs to be estimated. Consequently, different estimators are exposed as well as their bias. Finally, the generalization error is computed for each model and the results are presented. A comparison between the models is then performed.