Effects of caffeine were studied in a visual focused selective search task in well-rested and fatigued subjects. A dose of 200 + 50 mg caffeine or placebo, dissolved in decaffeinated coffee, was administered in a double-blind and deceptive fashion. The task was to detect a target letter on one diagonal of a visual display designated as relevant and ignore stimuli presented on the irrelevant diagonal. Behavioral measures were supplemented by event-related potential (ERP) measures. Subjects reacted faster in the caffeine condition. Caffeine enhanced the N1 and the N2b components. Selection of relevant information apparently was more adequate in this condition. Search negativity was not affected by caffeine. Caffeine effects on the P3 elicited...