Tissue microstructure characterisation through relaxometry and diffusion MRI using sparse component analysis

摘要: We introduce a new approach to characterise brain tissue microstructure combining diffusion MRI and relaxometry. The tissue is modelled as a composition of water pools with distinct T2 and diffusivities. Mathematically, we establish a blind source separation framework that we solve using sparse component analysis. Results show that this technique is able to extract distinct diffusion and relaxometry properties of the different water pools.