Understanding the effects of Alzheimer's disease in brain microstructure: a comparison study of models in diffusion MRI

Diffusion Weighted Imaging (DWI) is an MRI technique that quantifies the random movement of water molecules within tissues, offering valuable information about tissue architecture and helping to identify various pathologies, particularly in the brain and other soft tissues. A new model called Neurite Orientation Dispersion and Density Imaging (NODDI) has been proposed for characterizing brain tissue at a microscopic level, since it is a two-level multi-compartment model in which all compartments are considered non-exchanging, distinguishing between the tissue and non-tissue components within the brain. This thesis compares two versions of it, the NODDI-Bingham and NODDI-Watson diffusion MRI models in assessing brain microstructure across Alzheimer’s Disease, Mild Cognitive Impairment, and cognitively normal individuals [Alzheimer’s Disease Neuroimaging Initiative (ADNI), ]. Microstructural metrics (ODI, partial volumes, beta fraction) were extracted and mapped to several brain regions of white and grey matter, followed by statistical analysis, including three-way ANOVA, Tukey’s HSD test, and Pearson correlation with clinical cognitive scores. Key findings indicate that both models detect significant microstructural changes linked to cognitive decline, with NODDI-Bingham showing slightly higher sensitivity. Strong correlations in regions like the cortex and hippocampus underscore their role in Alzheimer’s progression. While promising as early biomarkers, NODDI metrics might be more effective when combined with other diagnostic tools (e.g. PET, genetic testing) to improve predictive accuracy. This work highlights the potential of advanced diffusion MRI models in understanding brain changes associated with Alzheimer’s and their possible role in early disease diagnosis.