To gain insight into the relationship between pathological alterations and memory deficits observed in Alzheimer's disease (AD), a number of amyloid precursor protein (APP) transgenic animal models have been generated containing familial AD mutations. The most commonly utilized method involves a cDNA-based approach, utilizing heterologous promoters to drive expression of specific APP isoforms. As a result of the assumptions inherent in the design of each model, the different cDNA-based transgenic mouse models have revealed different relationships between the biochemical, pathological and behavioral alterations observed in these models. Here we provide further characterization of a genomic-based, amyloid precursor protein yeast artificial chromosome transgenic mouse model of AD, R1.40, that makes few assumptions regarding disease pathogenesis to study the relationship between brain pathology and altered behavior. Aged R1.40 transgenic and control mice were tested for learning and memory in the Morris water maze and for working memory in the Y maze. Results from the water maze demonstrated intact learning in the both control and R1.40 mice, but impairments in the long-term retention of this information in the transgenic mice, but not controls. Interestingly, however, long-term memory deficits did not correlate with the presence of Abeta deposits within the group of animals examined. By contrast, age-related working memory impairments were also observed in the Y maze in the R1.40 mice, and these deficits correlated with the presence of Abeta deposits. Our results demonstrate unique behavioral alterations in the R1.40 mouse model of AD that are likely both dependent and independent of Abeta deposition.