Localized changes in oxygen consumption related to increased neural activity can result in a small and transient “initial dip” of the blood oxygenation level-dependent (BOLD) signal used in functional magnetic resonance imaging (fMRI). The initial dip has been of great interest to the fMRI community because it may provide a more accurate and localized measure of neural activity than the conventional BOLD signal increase. Although potentially useful as a technique for human brain mapping, the initial dip is not always detected and has been a source of some controversy. In this study, the BOLD response to a 4-s long visual stimulus was measured with a 3-T MRI system in 5 healthy volunteers both before and immediately after a 200-mg oral caffeine dose. The caffeine dose significantly (P< 0.001) reduced or eliminated the initial dip in all subjects. These findings suggest that caffeine usage may be a key factor in the detection of the initial dip in human fMRI studies.