Alzheimer's disease starts as an almost imperceptible malady, first observed clinically as a mild memory problem. Accumulating genetic and biochemical data have suggested that amyloid beta-protein (Abeta) plays an important role in this memory loss, and Abeta has been shown to suppress long-term potentiation (LTP), a cellular model for memory and learning. Here we show that a very brief (3 min) swimming, twice daily for 2 weeks, rescues LTP inhibition in the CA1 region of hippocampal slices caused by Abeta(42) or Abeta(40) carrying the Arctic mutation using a theta burst stimulation (TBS) protocol. Whereas the input-output curve was not affected, the paired-pulse ratio was reduced in mice receiving our repeated swimming protocol, suggesting a possible involvement of presynaptic facilitation. Similar to swimming, Abeta's inhibition of LTP could be rescued with the adenylyl cyclase, forskolin. Interestingly, this swimming protocol produced conditions in which a weak-TBS could invoke LTP not observed in naïve mice, which again was mimicked by forskolin. In contrast, the protein kinase A (PKA) inhibitor, H89, blocked both the forskolin and swimming potentiation of LTP; these data implicate cAMP/PKA signaling in the protective effect of swimming and mediating Abeta' detrimental effects. Our data add a new simple behavior paradigm that shows the importance of an environmental factor in reversing the pathophysiological effects of Abeta, and suggest new therapeutic avenues.