High and low responders to novelty (Wistar rats) were selected with the help of an open-field test and then equipped with intra-accumbens cannulae. They were then tested in a simple four-arm radial maze during 5 successive days, three trials per day, following intra-accumbens injections of distilled water or the dopaminergic D2 antagonist (±)-sulpiride. The injections were given 15 min before the first trial on each day. Both types of drug-naive rats reached the same level of performance on day 5. However, high responders made more visits, more revisits, and needed less time to make the first visit than low responders. Moreover, high responders showed their greatest increase in learning 2 days earlier than low responders. It is discussed that these differences between high and low responders are not due simply to differences in locomotor activity, but are due to a subtle, but important, difference in the mode of learning between both types. Sulpiride significantly attenuated the learning in both rat types; however, its effect in high responders was much less than that in low responders. It is suggested that the effects of sulpiride are not due to changes in locomotor activity, motivation, or perception, but are due to a learning deficit. The data are discussed in view of the genetic variation in the neurochemical and neurobiological makeup of the nucleus accumbens in both types.Key words: ventral striatum, radial maze, sulpiride, individual vulnerability.