Valid animal models for a specific human disease are indispensable for development of new therapeutic agents. The conclusions drawn from animal models largely depend on the validity of the model. Several studies have shown that administration of Abeta into the brain causes some of the pathological events observed in Alzheimer disease (AD). However, the validity of these models has not fully been examined. In this present study, we further characterized and validated Abeta1-40 injected mice as an animal model for AD, based on three major criteria: face, construct and predictive validity. Intracerebroventricular (i.c.v.) injection of Abeta1-40 into mice significantly impaired memory acquisition, but not memory retrieval, which implies similarity to the episodic anterograde memory deficit observed in the early stage of AD. Electrophysiological assessment showed that i.c.v. administration of Abeta1-40 significantly attenuated hippocampal long-term potentiation. Treatment with galantamine, a drug currently in clinical use for AD, significantly improved cognitive dysfunction in this model. These results demonstrate that i.c.v. injection of Abeta1-40 caused specific dysfunction of memory processes, which at least partly fulfills three validity criteria for AD. Symptomatic and pathophysiological similarities of this model to AD are quite important in considering the usefulness of this animal model. This validated animal model could be useful to develop and evaluate potential new drugs for AD.