Spatial and Temporal Characteristics of Diffusion MRI: New Developments and Applications

Diffusion processes in biological tissues involve intricate spatial and temporal dynamics modulated by the heterogeneous microstructures. In magnetic resonance imaging (MRI), diffusion-weighted gradients induce signal attenuation, typically analyzed via the Gaussian mono-exponential model. However, this model oversimplifies the diffusion process in biological tissues with complex morphology and spatial heterogeneity. Therefore, advanced non-Gaussian models are essential for more precise tissue characterization. In addition to the spatial characteristics, there's a growing acknowledgment of the temporal dependency in non-Gaussian diffusion models. Studies typically use oscillating-gradient spin-echo (OGSE) or pulsed-gradient stimulated-echo (PGSTE) sequence to shorten or prolong diffusion times. However, the time dependency across a broad range of diffusion time and the consistency in the parameters obtained from different sequences have not been well studied. Additionally, practical limitations in investigating diffusion time dependency lies in the acquisition time, necessitating the development of rapid acquisition techniques that enable multiple diffusion times in a single sequence to ensure efficiency, patient compliance, and data reliability while mitigating motion-induced image misregistration. In this dissertation, both spatial and temporal characteristics of diffusion MRI have been investigated to enable a number of challenging clinical and technical applications. Specifically, an advanced fractional order calculus (FROC) model which recognizes spatial heterogeneity was used to provide a new approach to differentiate insignificant and significant prostate cancer. Diffusion temporal dependency in Sephadex gels with varying bead size and permeability were investigated over a wide diffusion times range using multiple sequences. Finally, a novel time-efficient diffusion MRI acquisition technique was proposed to accelerate time dependent diffusion MRI acquisition.