Magnetic resonance imaging (MRI) has significantly extended the understanding of multiple sclerosis (MS), owing to its ability to sensitively depict the dynamics of the disease process in vivo. The subject of this review is the use of MRI in the post-mortem setting, with emphasis on how it may be used to improve the specimen selection process at autopsy. Lesions with active demyelination are highly interesting in the study of MS pathogenesis, but are rare in a typical autopsy material of chronic MS. The yield of MS lesions in autopsy specimen selection can be increased by the use of MRI-guided tissue sampling, as a significant proportion of abnormalities detected by post-mortem MRI are not macroscopically visible/palpable. The majority of these MRI abnormalities have been found to represent either discrete areas of microglial activation with no demyelination (so-called (p)reactive lesions), or active demyelinating MS lesions by further histopathological examination. The presence and extent of MS pathology outside of the focal demyelinated lesions is more readily appreciated by MRI-guided specimen sampling, as has been shown in the study of extensive areas of partial myelin loss in the spinal cord. A further advantage of MRI-guided specimen sampling is the ability to use three-dimensional and quantitative measures. The potential of correlating these with histopathological data may be further exploited in the future. The technical procedure for MRI-guided tissue sampling at autopsy is presented, and the limitations of the technique are discussed.