Some abnormalities in the P100 on monocular stimulation have been said to indicate retrochiasmal abnormalities. This study was undertaken to evaluate the reliability of visual evoked potentials (VEPs) for this purpose. We retrospectively reviewed VEPs that could be interpreted as indicative of a retrochiasmal lesion, based on an amplitude asymmetry of >50% between the P100 recorded on the left and right occipital regions. The abnormality had to be present on both left and right monocular stimulation (uncrossed pattern). Studies were conducted in accordance with the guidelines of the American Electroencephalographic Society. Findings on magnetic resonance (MR) imaging were analyzed. For comparison, we also studied a control group of 30 patients with normal VEPs. During a 16-month period, we identified 15 patients who fulfilled the aforementioned criteria and who had undergone an MR scan of the brain. Eight scans were normal and seven were abnormal. Of the seven abnormal scans, five revealed multiple periventricular white matter lesions in both hemispheres suggestive of multiple sclerosis: one a single punctate white matter lesion in the left frontal corona radiata, and one multiple white matter lesions in the basal ganglia, brainstem, and cerebellum but not in the hemispheres. In the control group with normal VEPs, nine had at least one hemispheric lesion on MR scans. This yielded a positive predictive value of 33%. The negative predictive value was 70%, sensitivity 36%, and specificity 68%. Because of a high rate of false positives, full-field VEPs were not reliable in predicting retrochiasmal lesions.