In vivo animal models are indispensable both for clarifying the pathological bases of schizophrenia, and for evaluating the potential benefits and disadvantages of novel therapy. Procedures that model mnemonic impairment are of particular interest since currently-available drugs do little to improve cognitive symptoms: this is hardly surprising, in fact, since most of them potently antagonise histamine H(1), muscarinic, and/or alpha(1)-adrenergic receptors. Further, their blockade of D(2) receptors likewise compromises cognitive performance. By contrast, D(3) receptor antagonism improves certain cognitive domains, suggesting that preferential antagonism of D(3) vs. D(2) receptors may permit enhanced effectiveness against cognitive dysfunction. The novel agent, S33138, possesses such an "optimised" profile and shows a unique and broad-based pattern of pro-cognitive properties in rodents and primates, in the absence of extrapyramidal and metabolic side-effects. The present article surveys the preclinical pharmacology of S33138. It also reviews developmental and genetic risk factors for schizophrenia and their experimental modeling in rodents, with a particular emphasis on sensorimotor gating and cognitive deficits.