Magnetic resonance imaging (MRI) allows the assessment of structural changes in subjects with Alzheimer's disease (AD). Early studies used visual assessments of MRI or manual measurements of structures of interest, although these methods were limited by inter-rater variability. Techniques have now been developed which allow automated analysis of both cross-sectional and longitudinal MRI data and have provided valuable information concerning the patterns and progression of atrophy in subjects with AD. It is also now possible using machine learning-based techniques to provide individual-level diagnostic information from MRI scans. Various analysis techniques have been applied to validate the use of MRI to capture subtle structural changes due to atrophy in AD and its usefulness in providing diagnostic and prognostic information, as well as tracking the disease progression in AD.