Developing novel diffusion MRI methods for comprehensive analysis of restricted and anisotropic self-diffusion system

Sammanfattning: Diffusion MRI is a non-invasive imaging technique used to study the microstructural properties of biological tissues by observing the self-diffusion of water molecules. Traditional diffusion MRI methods, based on the pulsed gradient spin-echo sequence, employ magnetic field gradients to encode information about translational motion. However, this approach combines various aspects of diffusion, such as restriction, anisotropy, and flow, into a single observable, leading to interpretation ambiguities, especially in complex heterogeneous materials like living biological tissues.In this thesis, we address these challenges and push the boundaries of diffusion MRI by introducing innovative techniques for studying biological tissue microstructure. Our approach centers around the "double-rotation" technique borrowed from solid-state NMR, which generates modulated gradient waveforms, enabling us to explore the 2D frequency-anisotropy domain in-depth. By integrating this technique with oscillating gradients and tensor-valued encoding, we create a comprehensive methodology for data acquisition. Drawing inspiration from the "model-free" analytical strategies originally designed for studying rotational dynamics in macromolecules, we extend its applicability to MRI techniques for understanding diffusion in biological tissues.Through a series of proof-of-principle experiments, we validate our novel acquisition and analysis strategy across various samples. These experiments encompass the study of isotropic and anisotropic Gaussian diffusion in simple liquids, characterizing anisotropic Gaussian diffusion in a lyotropic liquid crystal with lamellar microstructure, and exploring restricted diffusion in a yeast cell sediment. Additionally, we showcase the effectiveness of our methods on ex vivo mouse brain and tumor tissue, highlighting the practical potential of our approach.Our proposed double-rotation gradient waveforms enable comprehensive sampling of both the frequency and "shape" dimensions of diffusion encoding, providing detailed insights into restriction and anisotropy in heterogeneous materials. The implications of our work extend to model-free investigations, allowing us to understand microstructural changes linked with pathology or normal brain development.