The neural activity associated with conscious recollection and habit was examined using event-related brain potentials. In a training phase, participants learned A-B, A-C word associations in which the probability of specific responses was varied. Once a habit was established, participants studied and were tested on a series of short lists consisting of word pairs seen in training. The process-dissociation procedure was used to estimate the contribution of habit and recollection to memory performance. Habit estimates reflected the probability with which information was presented in training but recollection estimates did not show this effect. Recollection was associated with sustained negativity over the parieto-occipital region, which was opposite in polarity over the frontal regions. Indices of habit strength were associated with a sustained positivity over left fronto-temporal regions and a sustained negativity over right fronto-central regions. Partial-least squares analyses revealed two significant latent variables that distinguished recollection and habit, consistent with the distinction between consciously controlled and automatic influences of memory.