Magnetic resonance imaging (MRI) has been a valuable tool in investigating the pathological cascade of Alzheimer's disease (AD) and its progression, which are still open questions. Although some MRI-derived hallmarks in terms of functional connectivity and white matter degeneration have been revealed, the temporal involvement and interplay between them as well as other hallmarks such as amyloid load and neuronal density remain poorly understood. On the other hand, brain glucose hypometabolism is gradually taking center stage as a key player in the onset of AD, which had been described as a form of "type-3 diabetes". In this context, this thesis work presents a comprehensive study to characterize longitudinal changes in functional connectivity using resting-state functional MRI and microstructure using diffusion MRI, and to correlate them to glucose hypometabolism and neuronal density in a rat model of AD.

This study is mainly focused on establishing an optimized image processing pipeline dedicated for rat fMRI data, optimizing a computational model of diffusion in white matter using a novel deep learning approach, and assessing the spatiotemporal relationships between biomarkers regarding to microstructure and function in the brain by using advanced statistical methods as well as state-of-the-art machine learning approaches in order to provide a comprehensive characterization of the pathological cascade and progression of neurodegeneration resulting from brain glucose metabolism disruption.