We prospectively examined 58 patients with suspected or confirmed multiple sclerosis (MS) to evaluate the impact of paraclinical studies (evoked potentials (EPs) and magnetic resonance imaging (MRI] in the diagnostic evaluation of these patients. All patients had cranial MRI studies, brain-stem auditory (BAEP), visual (VEP), posterior tibial (PTN-SSEP) and median somatosensory (MN-SSEP) EPs. Patients were categorized according to criteria of Poser et al. initially without knowledge of paraclinical studies. On the basis of clinical history, neurological examination and CSF chemical analysis, there were 18 (31%) clinically definite (CD). 10 (17%) laboratory supported definite (LSD), 13 (22%) clinically probable (CP) MS cases; 17 (29%) cases could not be classified. By providing electrophysiological and/or anatomical evidence for a 'second lesion,' paraclinical studies permitted 25 patients to move from one diagnostic category to another, reflecting an increased certitude of the diagnosis. Twelve moved to CDMS by either MRI or one EP study. Four moved to CDMS on the basis of MRI only, and four moved to CDMS by VEP only. Five moved from an unclassified status to either LSDMS or CPMS by PTN-SSEP (2), by PTN-SSEP or MRI (1), by MRI or VEP (1) and by VEP only (1). Thus, the diagnosis of MS was refined in 28% of the patients (7/25) by incorporation of EPs alone whereas cranial MRI by itself increased the diagnostic certainty in only 16% (4/25). In the remaining 56%, the diagnosis was refined by use of either EPs or MRI. Overall, EP provided evidence for a 'second lesion' in 36% of the patients (21/58) and MRI 31% (18/58). While MRI may ultimately prove the single most useful paraclinical study in the diagnosis of MS. VEPs, and PTN-SSEPs by assessing areas currently not imaged by MRI, remain essential in patient evaluation. Changes in EPs and MRI technology will undoubtedly redefine the role of these diagnostic techniques.