Alzheimer’s disease and related disorders are heterogeneous, age-related neurodegenerative diseases characterized by the co-occurrence of multiple pathologies that manifest themselves decades before a patient becomes symptomatic. Clinical trials have focused on preclinical stages of AD to detect and diagnose a patient using in vivo imaging biomarkers. These biomarkers have been sensitive in detecting structural changes, such as cortical thickness and volumetry, during the early stages of AD. However, their potential as an imaging biomarker is limited due to their lack of sensitivity in differentiating co-morbid neurodegenerative pathologies, which simultaneously affect brain morphometry. Studies have shown that different brain regions are differentially affected by such co-morbid pathologies. Therefore, there is a need for region-specific in vivo biomarkers that can detect and quantify mixed pathologies. The work presented in this dissertation is built on the hypothesis that linking ex vivo MRI-derived morphometry measurements with neuropathology would reveal stronger morphometry/pathology signatures than antemortem MRI, which would help inform the development of disease-specific in vivo biomarkers. To discover AD-specific structural patterns associated with neuropathology/neurodegeneration, we introduce and leverage a postmortem whole-hemisphere MRI dataset to link neuropathological markers derived from gold-standard histopathology examination with morphometry measurements derived from structural MRI. This dissertation presents a set of computational image analysis methods and pipelines to analyze high-resolution postmortem MRI consistently. Specifically, we present a fast, robust, and automated voxel-based tissue segmentation and surface-based anatomical parcellation pipeline for whole cerebral hemisphere 7 tesla postmortem MRI in diseased populations. Next, by constructing a population-level postmortem ex vivo MRI template, we conduct both voxel-based and surface-based point-wise statistical studies within a common coordinate system to help discover disease-specific patterns of neurodegeneration by linking morphometry measurements with neuropathology markers derived from MRI and histopathology, respectively. We then present a framework to align within-subject postmortem MRI to its corresponding antemortem MRI, which would enable the exchange of information from postmortem to corresponding antemortem MRI. This dissertation presents a one-of-a-kind analysis that compares morphometry/pathology in matched postmortem and antemortem MRI specimens. Our results suggest that high-resolution postmortem 7 tesla MRI yields localized atrophy measures that are more sensitive to tau pathology and neuronal loss in Alzheimer’s disease than corresponding measures on antemortem 3 tesla MRI. The computational tools, methods, and pipelines presented in this dissertation are released as a stand-alone software, “purple-mri: Penn Utilities for Registration and ParcelLation of Ex vivo MRI”, to enable systematic and continued development of the next generation of image analysis algorithms for postmortem imaging. The techniques and tools presented in this dissertation will help inform the development of AD-specific in vivo biomarkers for clinical trials.