Multiple sclerosis (MS) is after trauma the most important neurological disease in young adults, affecting 1 per 1000 individuals. With currently available medications, most of these targeting the immune system, satisfactory results have been obtained in patients with relapsing MS, but these can have serious adverse effects. Moreover, despite some promising developments, such as with B cell targeting therapies or sphingosine-1-phosphate modulating drugs, there still is a high unmet need of safe drugs with broad efficacy in patients with progressive MS. Despite substantial investments and intensive preclinical research, the proportion of promising lead compounds that reaches the approved drug status remains disappointingly low. One cause lies in the poor predictive validity of MS animal models used in the translation of pathogenic mechanisms into safe and effective treatments for the patient. This disturbing situation has raised criticism against the relevance of animal models used in preclinical research and calls for improvement of these models. This publication presents a potentially useful strategy to enhance the predictive validity of MS animal models, namely, to analyze the causes of failure in forward translation (lab to clinic) via reverse translation (clinic to lab). Through this strategy new insights can be gained that can help generate more valid MS models.