We investigated changes in autonomic nerve activity following caffeine intake by power spectral analysis of R-R intervals of heartbeats in humans. A beverage containing 240 mg of caffeine or a control beverage was given to 10 healthy volunteers, and R-R intervals were measured while subjects were sitting and controlling their respiration at a constant rate. After consumption of the caffeine-containing beverage, a transient and significant increase (P < 0.001) in spectral integrated values (areas under the curve) of high frequency power (high component, HC) was observed, and at 30 min the value was significantly greater than in controls (P < 0.02), suggesting an increase in vagal autonomic nerve activity. The effect of caffeine was also examined using decaffeinated coffee supplemented with exogenous caffeine (2 mg/kg body wt). A transient and significant increase (P < 0.0001) in HC was observed, and the value was significantly greater (P < 0.02) than when subjects consumed decaffeinated coffee without supplemental caffeine. The ratio of HC to total integrated value (which is also used as a selective indicator of vagal activity) was also significantly higher (P < 0.04) after caffeine consumption. Physiological variables accompanying the change in autonomic nerve activity (i.e., blood pressure, surface body temperature and heart rate) were not significantly affected by caffeine intake. These results suggest that power spectral analysis of heartbeat R-R intervals is an effective and noninvasive method for detecting subtle changes in autonomic nerve activity in response to food intake.