Magnetic resonance imaging (MRI) has become an important technique for monitoring the effectiveness of putative treatments for multiple sclerosis (MS) because of its high sensitivity, objectivity, and noninvasive nature. Its importance as a surrogate measure of disease, however, is an issue that is more difficult to validate than might seem to be the case. In this review, we describe the role of MRI in the assessment of putative therapies for MS. New magnetic resonance techniques and methods of image analyses aimed at better demonstrating the nature and extent of disease are discussed, and the role of MRI in published MS therapeutic trials is examined. MRI is a frequently used secondary outcome measure for putative treatment strategies for MS. Although it is sensitive to changes in the inflammatory component of the MS disease process, poor correlation has been noted between MRI findings and long-term patient outcome. There is a widespread expectation that new magnetic resonance techniques--such as fluid-attenuated inversion recovery, magnetization transfer imaging, and magnetic resonance spectroscopy--will ultimately be useful for characterization of pathologic changes within the MS lesion and more generally of the MS disease process. Whether magnetic resonance changes seen in experimental therapies predict the long-term clinical course of the disease remains to be determined.