All atypical antipsychotics avoid extrapyramidal side-effects yet differ in their propensity to cause other side-effects, like prolactin elevation. We proposed that the atypical antipsychotics with a propensity for prolactin elevation would show a higher pituitary versus striatal D2 receptor occupancy. To investigate this hypothesis, we tested four atypical antipsychotics, two that are commonly associated with prolactin elevation (amisulpride and risperidone) and two that are less frequently associated (quetiapine and olanzapine). In particular, we calculated their ED(50) values to increase plasma prolactin and block peripheral pituitary D2 receptors to their ED(50) values to antagonize apomorphine-induced stereotypy and occupy central striatal D2 receptors. All antipsychotics dose dependently increased prolactin levels and antagonized apomorphine-induced stereotypy. However, the central to peripheral potency (ED(50) for apomorphine antagonism to ED(50) for prolactin elevation) differed remarkably across these drugs: amisulpride (21764), risperidone (14), quetiapine (10), and olanzapine (1.7). Compounds displaying a higher peripheral potency brought about higher prolactin levels for a given level of functional central antagonism. This dissociation between central and peripheral effects was explained by the differential occupancy of D2 receptors in the striatum versus in the pituitary [ratio of striatal/pituitary ED(50) values (milligram per kilogram) for D2 occupancy): amisulpride (17/0.026 = 654), risperidone (0.89/0.081 = 14), quetiapine (24/4.1 = 6), olanzapine (0.30/0.43 = 0.7). These results indicate that dissociation between central and peripheral D2 receptor occupancy is a major determinant of the degree of prolactin elevation observed at therapeutic doses.