Thirty-six patients with multiple sclerosis were evaluated by means of brain-stem trigeminal and auditory evoked potentials. The brain-stem auditory evoked potentials (BAEPs) were abnormal in 26 patients (72.2%). Brain-stem trigeminal evoked potentials (BTEPs) yielded similar results, showing distorted waveforms and/or prolonged latencies in 25 patients (69.4%). As expected, the MRI proved to be the most efficient single test, revealing plaques in 86.4% of the patients evaluated. However, the diagnostic accuracy of MRI was lower than that provided by the combination of the BTEP and BAEP (88.9%). Moreover, in patients having signs of brain-stem involvement, the BTEP, alone and in combination with the BAEP, proved to be more sensitive than the MRI in revealing brain-stem lesions. Correlation between clinical and BTEP findings could be found only in those patients who presented with signs of trigeminal involvement such as trigeminal neuralgia or dysesthesiae. The analysis of the BTEP waveforms showed two distinct types of abnormality-a peripheral type and a central type-suggesting plaques in distinct locations. Both the BTEP and the BAEP demonstrated a correlation with the clinical course of the disease and the condition of the patient at the time of the evaluation. Relapse of the disease was associated with a marked prolongation of the central conduction time in the BTEP and in the BAEP, suggesting the application of such studies to the monitoring of unstable patients in the evaluation of new therapeutic protocols.