Serial brain magnetic resonance (MR) imaging detects active lesions 5-10 times more frequently than the occurrence of clinical changes in patients with early relapsing-remitting and secondary progressive multiple sclerosis (MS). Based on monthly unenhanced and gadolinium enhanced MR findings in 23 unselected and untreated patients, the power of an MS treatment trial was calculated, using MR imaging activity as the primary measure of outcome. It was shown that a 80% reduction in the number of active lesions (i.e. an efficacy of 80%) should be detected using a placebo-controlled parallel-groups design with a power of 80%, if either 2 x 20 patients are scanned monthly for 4 months, or 2 x 30 patients monthly for 2 months. Short to medium term studies of new experimental treatments in MS, using MR imaging as the primary outcome measure, provide considerable statistical power in small patient populations studied over a short period of time.