Although the precise nature of the changes in synaptic transmission that ultimately account for the increase in synaptic strength are still the subject of intense interest (and controversy), it is clear that activated protein kinases may affect the AMPA receptors mediating excitatory synaptic transmission in several ways. First, the AMPA receptors may become phosphorylated by one or both kinases, rendering them more sensitive to presynaptically released glutamate. Second, excitatory synapses are located on the ends of small processes known as dendritic spines (F1), and these may be changed in shape or number after induction of long-term potentiation. Finally, clusters of AMPA receptors may be translocated from an intracellular, subsynaptic location and inserted into the postsynaptic membrane at the end of the dendritic spine. The net effect of all three mechanisms is to increase the sensitivity of the postsynaptic cell to glutamate, thereby increasing the chances that activity in the presynaptic cell will successfully elicit action potentials in the postsynaptic cell. Long-term potentiation thus represents a use-dependent strengthening of connections between cells. Because genetic and pharmacological manipulations that prevent induction of long-term potentiation typically block one or more forms of learning and memory in animal experiments, it is likely that the neurobiological changes in glutamatergic synapses underlying long-term potentiation represent the fundamental cellular basis for cognition and the encoding of memories.