Conventional animal models of schizophrenia do not pretend to account for the symptoms of schizophrenia; i.e. they lack construct validity. Ellenbroek and Cools (1990) have reviewed a number of other models with construct validity for fundamental psychological processes, which are hypothesised to be deficient in schizophrenia. Two of these, Latent Inhibition (LI) and the Kamin Blocking effect (KB), essentially demonstrate the effects of prior learning on the formation of current associations. Animal studies have now shown that LI and KB are disrupted by increased dopaminergic activity, and restored by dopaminergic blockade. The present paper adds to the database described by Ellenbroek and Cools and discusses some further theoretical and practical issues. Specifically: (a) Kamin's blocking is disrupted in acute, but not in chronic schizophrenic subjects; (b) LI and KB are not simply examples of associative interference; (c) The masking task used in adult human studies of LI does not introduce an element of blocking as defined by Kamin's paradigm; (d) A direct study of "selective attention" in acute schizophrenia suggests that the impairment in LI and KB relates to the utilisation of experience about past regularities, rather than a generalised attentional deficit; (e) Recent studies, especially with nicotine, indicate that LI can be disrupted by drug administration during acquisition only; (0 It is now possible to study the release of dopamine, and of other transmitters in the nucleus accumbens and in other brain areas, during the behavioural paradigm of LI in rats. In this way it should be possible to determine directly the neuronal circuitry involved in LI (and KB): this approach could explain the link between the neuropathology and neuropharmacology of schizophrenia and its symptoms, besides providing more valid test procedures for evaluating potential neuroleptic drugs.