An indirect event-related potential (ERP) memory assessment procedure was used to study ERP correlates of suggested hypnotic amnesia. Subjects selected for high or low hypnotic susceptibility learned two 5-word lists prior to hypnosis. They then listened to a recorded hypnotic induction and amnesia suggestion for one of the two lists. After being aroused from hypnosis, they learned a third list. ERPs were then collected during a recognition task in which subjects were presented with a random ordering of the three lists intermixed with seven lists of unlearned words. Subjects were required to press one button if the word was from the third list learned, and another button if the word was unlearned or from words learned pre-hypnosis. Thus response requirements were identical for the first two lists learned. A subgroup of low-hypnotizables were asked to simulate hypnotic amnesia. Only subjects who later demonstrated hypnotic amnesia on a recognition test showed significantly different ERP responses (larger P300 amplitude) to words for which amnesia had been suggested compared to control words. The ERPs of these high-hypnotizable amnesics significantly differed from those of both the other groups (i.e., high- and low-hypnotizables who did not report amnesia, and simulators who reported, but did not experience amnesia). This result indicates that the phenomenon of suggested hypnotic amnesia cannot be explained solely by behavioral compliance or simulation.