Decreases in the 67 kDa isoenzyme of brain glutamic acid decarboxylase (GAD(67)) expression have been consistently found in patients with bipolar disorder and schizophrenia. In animals GAD(67) expression is diminished by chronic, but not acute stimulation of dopamine D(2) receptors and by short-term blockade of NMDA receptors. In contrast, chronic treatment with D(2) receptor antagonists enhances GAD(67) expression. Thus, antipsychotic treatment cannot explain the reduction in GAD(67) levels in patients with psychotic disorders. Rather, pathophysiological findings such as reduced viability of cortical glutamatergic neurones (in schizophrenia) or enhanced dopamine sensitivity (in bipolar disorder) might explain the observed reduction in GAD(67). Since reduction in GAD(67) expression leads to reduced levels of GABA, the GABAergic inhibitory control over glutamatergic cells is reduced. Psychosis could result from AMPA receptor activation caused by overactivity of the glutamatergic system. GAD(67) levels would thus be a surrogate marker for psychosis liability. Pharmacological principles that raise GAD(67) expression levels could represent novel targets for antipsychotic therapy.