In 2001, an international panel suggested new diagnostic criteria for multiple sclerosis (MS). These criteria integrate clinical, imaging (MRI), and paraclinical results in order to facilitate diagnosis. Since then, these so-called McDonald criteria have been broadly accepted and widely propagated. In the meantime a number of publications have dealt with the sensitivity and specificity for MS diagnosis and with implementing these new criteria in clinical practice. Based on these empirical values and newer data on MS, an international expert group recently proposed a revision of the criteria. Substantial changes affect (1) MRI criteria for the dissemination of lesions over time, (2) the role of spinal cord lesions in the MRI and (3) diagnosis of primary progressive MS. In this article we present recent experiences with the McDonald and revised criteria.