The present studies were designed to test the learning and memory capacities of transgenic mice with central overexpression of corticotropin-releasing factor in a forced alternation water T-maze task and in the Morris water maze. In T-maze testing, littermate control mice reached a criterion of 70% correct responses after five days of trials, while the performance of transgenic subjects was still random after the same training. In Morris maze testing, control subjects reached the submerged platform significantly faster (F(1.48) = 4.51, P < 0.05) after three days of trials, while the performance of transgenic mice was unimproved over the same period. The deficit in Morris maze performance in transgenic mice was reversed when the platform was visible above the surface of the water. Pre-test administration of the benzodiazepine anxiolytic, chlordiazepoxide (10 mg/kg), before acquisition training also produced a significant (F(4.40) = 16.61, P < 0.001) and persistent improvement in Morris maze performance in transgenic mice when compared to vehicle-treated transgenic litter mates. Finally, there was no evidence of hippocampal cell loss in transgenic brains. The results suggest that corticotropin-releasing factor-overexpressing mice exhibit a profound learning deficit without sensory or motor-related impairments, and that memory plasticity can be restored by anxiolytic pre-treatment. Thus, constitutive overabundance of brain corticotropin-releasing factor may produce hyperemotionality that interferes with learned behaviors. Stress-related disorders characterized by co-morbid deficits in learning/memory may benefit from pharmacological normalization of brain corticotropin-releasing factor systems.