Insulin receptors have been detected in several structures of the brain, yet the biological significance of insulin acting on the brain remains rather unclear. In humans, direct central nervous effects of insulin are difficult to distinguish from alterations in neuronal functions because of insulin-induced decrease in blood glucose levels. Since several intranasally administered viruses, peptides, and hormones have been shown to penetrate directly from the nose to the brain, we tested whether insulin after intranasal administration likewise has access to the brain. After a 60-min baseline period, insulin (20 IU H-Insulin 100 Hoechst) or vehicle (2.7 mg/ml m-Cresol) was intranasally administered every 15 min to 18 healthy subjects according to a double-blind within-subject crossover design. Auditory-evoked potentials (AEP) indexing cortical sensory processing were recorded while the subjects performed a vigilance task (oddball paradigm) during the baseline phase and after 60 min of intranasal treatment with insulin or placebo. Blood glucose and serum insulin levels were not affected by intranasal insulin. Compared with placebo, intranasal administration of insulin reduced amplitudes of the N1 (P < 0.005) and P3 (P < 0.02) components of the AEP and increased P3 latency (P < 0.05). The reduction in P3 amplitude was most pronounced over the frontal recording site (2.42 +/- 1.00 vs. 4.92 +/- 0.79 microV, P < 0.0005). At this site, after insulin administration, a broad negative shift developed in the AEP between 280 and 500 ms poststimulus (area under the curve -166.0 +/- 183.8 vs. 270.8 +/- 138.7 microV x ms after placebo, P < 0.01). The results suggest that after intranasal administration, insulin directly enters the brain and exerts distinct influences on central nervous functions in humans.