Twenty-eight volunteers were instructed to attend stimuli presented at one side of the computer screen and to ignore stimuli presented at the other side. Both attended and unattended stimulus series consisted of targets (25%) and nontargets (75%) defined on the basis of stimulus shape. Attended targets required a binary choice based on stimulus color. Selective attention led to the expected increase in both midlatency (N2b) and late (P3) brain potential components. Furthermore, selective attention led to increased anticipatory cardiac slowing preceding the target stimulus and to increased primary bradycardia. Correlational analyses revealed a positive relation between the effects of selective attention on N2b amplitude and primary bradycardia suggestive of cortical involvement in the chronotropic control of heart rate.