Indirect dopamine (DA) agonists disrupt, and antagonists potentiate, latent inhibition (LI), an attentional phenomenon in which pre-exposure (PE) to a stimulus without consequence impairs subsequent learning of associations with that stimulus. It has previously been argued that disrupted LI, which is present in acute schizophrenia, might reflect a core cognitive deficit underlying the attentional deficits of the disorder, and that potentiated LI may be predictive of antipsychotic activity. We have presented animal evidence consistent with a critical role of dopamine function in the n. accumbens (NAC) in the control of latent inhibition (Joseph et al, Schiz Res, 11; 112, 1994). This putative link between NAC DA function and some of the symptoms of schizophrenia and their treatment has been challenged on the basis that intra-NAC AMP does not disrupt LI, using a novel paradigm, and that the apparent systemic effects of DA drugs on LI are actually artefacts of their actions on the strength of learned associations (see Killcross et al, Psychopharmacol, 115; 196-205, 1994).We have now demonstrated, using our established paradigm, (Peters and Joseph, Behav Pharmacol 4; 183-6, 1993) that reducing DA function bilaterally and selectively in NAC, either by local injection of haloperidol (0.5 μg/side) 15 min before conditioning, or by prior lesioning with 6-OH-DA, results in potentiated LI. That is, 10 PE of a novel tone, chosen to be insufficient to result in LI in control or vehicle infused animals, did result in the impairment of subsequent tone-mild footshock conditioning in the treated animals. However, neither treatment reduced the strength, or accelerated the extinction, of the learned association in the non-PE animals. We conclude that it is indeed the NAC in which DA function is critical for the modulation of LI, and that drug effects on the strength of conditioning, for which previous evidence is anyway far from convincing, do not interfere. The failure of repeated intra-NAC AMP to disrupt LI is most likely due to its inability to induce impulse-dependent DA release (see Warburton et al, Behav Pharmacol, 4; 454, 1993 and in press).