A model of frontal inhibition in hypnosis was elucidated by examining with cortical evoked potentials error negativity and positivity in 23 low versus medium/high hypnotisables during a perceptual task which included trials with conflicting stimulus information. In susceptible subjects the number of correct responses was found to be smaller for trials with incongruent stimulus information with hypnosis when compared with baseline. While the early negative wave to incorrect responses tended to be higher in susceptible subjects, this wave was no longer followed by a late positivity, posited to reflect a failure to process further the error-related information. The results with hypnosis are interpreted as a failure of context updating without a global deficit in supervisory attention.