Despite vast clinical experience with antipsychotics, there is no broad consensus on the doses of these substances that should be administered. Currently, most antipsychotics are administered empirically according to clinical dose-finding studies, in which arbitrarily selected doses were tested to find the "most efficient" dose range in a patient population, with no regard for the molecular effects of the tested drug. Brain imaging studies using nuclear medical techniques, such as positron emission tomography (PET) or single photon emission computed tomography (SPECT), can now provide a rationale for doses, directly derived from the central effects of the drugs on neurotransmitter receptors measured in vivo. PET results indicate that occupancy of at least 65% of dopamine D(2) receptors is needed for clinical response to antipsychotics, and that occupancy rates exceeding 72 and 78% are associated with a high risk for elevation of prolactin levels and motor adverse effects, respectively. For example, clinical studies with haloperidol do not point to an advantage of dosages exceeding 5 mg/day. The relevance of D(2) receptor occupancy for drug administration is also borne out by studies relating the effects of antipsychotics to their D(2) receptor occupancy in relevant animal models. Taken together, neuroimaging and clinical studies, as well as animal models, provide a rationale for the use of relatively low doses of typical antipsychotics and equivalent doses of novel antipsychotics. The lower risk of adverse effects with appropriate doses of antipsychotics may further enhance compliance and outcome. This seems to be particularly important in individuals experiencing a first episode of schizophrenia, as they appear to be especially responsive to pharmacotherapy and quite sensitive to adverse effects.