Experimental evidence indicates a role of the N-methyl-D-aspartate receptor in the pathogenesis of brain injury occurring during cardiac surgery with cardiopulmonary bypass (CPB). Dextromethorphan is a noncompetitive antagonist of this receptor with a favorable safety profile. Thirteen children age 3-36 months undergoing cardiac surgery with expected CPB of 60 minutes or more were randomly assigned to treatment with dextromethorphan (36-38 mg/kg/day) or placebo administered by naso-gastric tube. Dextromethorphan was absorbed well and reached putative therapeutic levels in blood and cerebrospinal fluid. Adverse effects were not observed. Mild hemiparesis developed after operation in one child of each group, and severe encephalopathy in one of the placebo group. Sharp waves were recorded in postoperative continuous electroencephalography in all placebo (n = 7) but only in 2/6 dextromethorphan treated children (p = 0.02). Pre- and postoperative cranial magnetic resonance imaging (MRI) revealed less pronounced ventricular enlargement in the dextromethorphan group (not significant). An increase of periventricular white matter lesions was visible in two placebo-treated children only. No elevations of cerebrospinal fluid enzymes were observed in either group. Although children with dextromethorphan showed less abnormalities in electroencephalography and MRI, dissimilarities of the treatment groups by chance diminished conclusions to possible protective effects of dextromethorphan at this time.