The effects of fear/anticipatory anxiety on the acoustic startle reflex were investigated in humans using a paradigm involving anticipation of electric shocks. The eyeblink component of the startle reflex, elicited by an abrupt auditory stimulus, was measured in 9 normal volunteers during either the anticipation of electric shocks (anticipatory anxiety) or periods in which no shocks were anticipated (safe period). The eyeblink was consistently higher in amplitude, and shorter in latency, during periods when the subjects anticipated shocks, compared to the safe periods. This effect could not be attributed solely to a reduction in habituation and was statistically significant before the subjects actually received any shock (a single 30 mA stimulation on the median nerve). These results indicate that anticipatory anxiety can be measured objectively in humans using the fear-potentiated startle reflex in a paradigm not actually requiring any shock. Because a great deal is known about the neuroanatomical and pharmacological mechanisms of fear-potentiated startle in laboratory animals, this test procedure may be especially useful in humans to investigate the neurobiological substrates of anxiety disorders and their pharmacological treatments.