A central topic of recent research on multiple sclerosis (MS) concerns the development of sensitive neurophysiological techniques with the ability to detect lesions at the earliest possible stage. Specifically, the recording of evoked potentials has during the past decade demonstrated its diagnostic value by disclosing clinically silent lesions especially of the optic nerve and brainstem region. The visual evoked potential (VEP) was registered in 177 patients with definite, probable and possible MS (McAlpine). Abnormal latencies were obtained in 86/46/34% of these three patient groups, respectively. 44% of all MS-patients without clinical or anamnestic evidence of optic nerve dysfunction showed delayed VEPs. The brainstem auditory evoked potential (BAEP) enables topodiagnostic statements to be made regarding brainstem lesion site. This is very valuable, since the brainstem region is especially prone to demyelination, which clinically as well as morphologically often evades detection. 74/60/48% of our definite, probable and possible MS-patients had pathological BAEPs. 50% of asymptomatic patients had an abnormal BAEP-result. Computer tomography yielded brainstem pathology in only 10% of all MS-patients.