The potential neurotoxic effects of gadolinium (Gd)-based compounds for enhanced MRI are not completely understood. We investigated electroencephalography changes induced by ionic and non-ionic Gd-based compounds administered intravenously in patients affected by lesions of the central nervous system (CNS) characterized by breakdown of the blood-brain barrier. This double-blind, randomized, study of two parallel groups involved 40 patients scheduled for an MRI examination with contrast medium for known CNS lesions. Twenty patients were randomly allocated to receive non-ionic Gd-DTPA-BMA/gadodiamide and 20 patients were randomly allocated to receive ionic Gd-DTPA/gadopentetate. For both groups the intravenous dose was 0.1 mmol/kg body weight. Three electroencephalography recordings were performed: immediately before, during, and 15 min after contrast medium injection. Mean and peak frequencies of the beta band and absolute power of the delta and/or theta bands of the electroencephalograms (EEGs) were noted. Each EEG was also evaluated to detect any alterations. The values of the 8-12 Hz band showed a significant increase during and after injection versus baseline in the gadopentetate group (P<0.05) and a significant decrease during injection in the gadodiamide group (P<0.05). The values of the 12-16 Hz band showed a significant increase versus baseline during and after injection in the gadopentetate group (P<0.05). The electrophysiological method based on computerised spectral analysis is a sensitive tool for evaluating effects of contrast media on brain bio-electric activity. EEG changes are detectable, even in the absence of any clinical evidence. It would appear that there might be clinical advantages in the use of non-ionic compounds.