Changes in clinical symptoms and MRI lesions of multiple sclerosis (MS) were evaluated on two occasions, one month apart, in 30 patients. Seventeen patients (group 1) with acute exacerbation were treated with methylprednisolone in high, then decreasing doses during a total of 30 days. MRI examinations were performed before and at the end of treatment. The remaining 13 patients (group 2) had been clinically stable for more than 6 months and received no treatment; here again, MRI was performed at 30 days' interval. All patients in group 1 showed functional improvement. The MRI lesions remained stable in 7 of group 1 patients and in 4 of group 2 patients. In the remaining 19 patients (10 in group 1 and 9 in group 2), the number, size and location of MRI lesions were found to have changed over 1 month. There was no correlation between clinical changes and the modification observed at MRI. This study confirms that high-dose corticosteroids are effective, at least clinically, in acute exacerbations, but the main results are that MS is a continually evolving disease, that changes rapidly occur in the lesions observed at MRI and that corticosteroids do not seem to influence the course of MRI lesions. Our study also suggests that MRI is inadequate to evaluate the effectiveness of short-time treatments of MS.