1. Long-term potentiation (LTP), long-term depression (LTD), and depotentiation of synaptic activity have been suggested to model synaptic plastic changes that occur during learning. Recent reports however show that neither LTP induced by high frequency stimulation (HFS) in the dentate, CA3, or CA1, nor depotentiation in area CA1 of the hippocampus, are reliable models of the learning abilities of rats or mice. LTD cannot reliably be obtained in the hippocampus in vivo and might be an artefact caused by altered inhibitory transmission. 2. Experiments with gene deletion ('knock out') mice strains show that mice that do not express HFS-induced LTP in the dentate are able to learn spatial tasks. 3. Studies of the effect of NMDA receptor blockers also showed that HFS-induced LTP in the dentate is not a model for processes that occur during learning. Studies using drugs that act on metabotropic glutamate receptors showed that HFS-induced LTP or depotentiation of LTP in area CA1 are not models for learning mechanisms either. 4. Neither in vivo recording of naturally-occurring LTP in the dentate nor synaptic saturation experiments in the hippocampus was able to support the theory that LTP occurs during learning. 5. While in vitro experiments are essential tools to investigate cellular and subcellular mechanisms that underlay synaptic transmission, measurements of LTP, LTD, or DP are not reliable models for learning processes and cannot replace experiments with intact animals that learn spatial tasks.