The influence of a single dose of caffeine was evaluated in focused and divided attention conditions of a visual selective search task in which subjects had to perform controlled search processes to locate a target item. Search processes were manipulated by varying display load. A dose of 3 mg/kg body weight caffeine or lactose, dissolved in a cup of decaffeinated coffee, was administered double blindly and deceptively to overnight abstinent coffee drinkers. Behavioral measures were supplemented by event-related potentials (ERPs). Subjects reacted faster in the caffeine condition. The P3b peak latency decreased after caffeine in the low display load condition and in the focused attention condition, indicating that the effects of caffeine are dependent on the number of relevant display items, not on the total number of display items presented. Search processes, as reflected in a negative ERP deflection, were not affected by caffeine.