Several large studies have demonstrated the improved safety record of nonionic versus ionic contrast agents for intravenous administration. However, nonionic agents are much more expensive than ionic agents. The author addresses whether, given this large cost differential, nonionic contrast agents should always be used in neuroangiography (cerebral and spinal cord angiography and intravascular neurointerventional procedures). The answer could come from a closer examination of the effects of contrast agents on the brain. There have been a number of animal experiments and clinical trials performed using a variety of available intravascular contrast agents. In an attempt to arrive at some reasonable conclusions regarding the use of contrast agents today, the author reviews several of these studies. In the human studies, three areas were analyzed: 1) cerebral angiography, 2) spinal cord angiography, and 3) intravascular neurointervention. The author explains why demonstrating the effect of a contrast agent on the brain or spinal cord in the clinical setting is more difficult than studying the effect of this agent on the liver, heart, or kidney. For example, obtaining objective measurements of altered cerebral physiology following intravascular injection of a contrast agent may itself alter the physiology. In lieu of objective measurements, investigators must rely on apparent changes in behavior, mentation, or the production of a focal neurologic deficit. However, it is extremely difficult, if not impossible, to separate the effect of the contrast agent from the effects of the arteriographic procedure, or from the disease process being evaluated. The neuronal environment is protected by the blood-brain barrier. A number of animal experiments have demonstrated that nonionic agents produce breakage of the blood-brain barrier less frequently than do ionic agents. In these studies, nonionic agents also produced fewer neurologic effects than did ionic agents. The human studies showed no statistically significant differences in neurologic effects when ionic and nonionic agents were compared. Cerebral: minor changes in heart rate were more common with ionic than with nonionic agents; there were no significant electroencephalogram changes in any of the patients studied. Spinal cord: the effects of intravascular injections of contrast material into the spinal cord of experimental animals have been rarely evaluated; direct comparisons of contrast agents in human spinal cord angiography have not been performed. Neurointervention: there have been no comparative studies of different contrast agents used during intravascular neurointerventional procedures in humans. Extensive animal data demonstrate that nonionic contrast agents are safer than ionic for cerebral angiography. Animals in the cited studies show less blood-brain barrier disruption, fewer direct neuronal effects, and fewer neurobehavioral deficits. However, the overwhelming conclusion from the human studies is that, while there is evidence in the experimental animal that nonionic agents produce fewer neurologic effects than do ionic agents, no study to date has been able to translate these findings into an apparent clinical difference in humans, mainly because it's so difficult to detect and measure neurologic changes in human trials. In addition, differences in neurologic effects between contrast agents used in human studies may be relatively small. Thus, one must make an educated guess as to the appropriate use of contrast agents in the context of their apparent clinical safety and cost-benefit ratio.