An increasing number of structurally heterogeneous compounds, which may act via very different categories of neuronal mechanisms, have been proposed to facilitate attentional abilities and acquisition, storage and retrieval of information, and/or to attenuate the impairments of such cognitive functions associated with age or dementia. In this article, Martin Sarter briefly reviews the data on putative cognition enhancers and examines the possible bases for the discrepancy between preclinical predictions of efficacy and the fact that unequivocal demonstration of drug-induced cognition enhancement in humans has only rarely been reported. Previous preclinical research strategies appear to have focused on the demonstration of drug effects in a wide variety of tests of uncertain validity, rather than on determination of the specific psychological and neurobiological processes affected by putative cognition enhancers. Some criteria are proposed for evaluating the validity of preclinical tests for cognition enhancers.