Alzheimer's disease (AD) is the most common cause of senile dementia, for which there is presently no disease-based treatment. The identification of genetic factors contributing to this disease, and the intense investigation into the cell biology of amyloid precursor protein (APP) and, to some extent, tau, has led to the development of several transgenic mouse models of this disease. These mice show some of the characteristic AD pathology, such as an age-dependent formation of amyloid plaques consisting of Abeta peptides. However, they usually lack both the tau pathology, i.e. neurofibrillary tangle formation, and the neurodegeneration also associated with AD. Importantly, many of these transgenic lines develop age-dependent deficits in some relevant behavioural tests and thus provide an animal model not only for amyloidosis but also for the cognitive deficits of AD patients. Incorporation of additional disease genes may lead to models that show a more complete disease phenotype. This review attempts to summarize much of this work, and describes how the availability of these models should assist in the understanding of AD aetiology and the identification of effective treatments for this disease. The review also considers the role behavioural testing may have in future AD drug discovery research.