Objective: The aim of this article is to provide a review of studies using N-methyl-D-aspartate (NMDA) receptor antagonists to assess the hippocampal long-term potentiation (LTP)/learning hypothesis.
Discussion: In particular, we will re-examine the validity of both (1) the original hippocampal LTP/spatial learning hypothesis of Morris and (2) the sensorimotor account put forward by Cain, among others, both from the point of view of the pharmacological studies on which they were based and with regard to recent studies with genetically modified mice. More specifically, we will review the pharmacological studies in the light of recent work on the glutamate receptor A (GluR-A or GluR1) L-alpha-amino-3-hydroxy-5-methyl-4-isoxazelopropionate (AMPA) receptor sub-unit knockout mouse. We will argue that neither the original hippocampal LTP/spatial learning hypothesis nor a sensorimotor account can adequately explain all of the available data. We argue instead that hippocampal synaptic plasticity, which requires NMDA receptors for its induction and GluR-A-containing AMPA receptors for its continued expression, contributes to a process whereby appropriate behavioural responses are selected rapidly on the basis of conditional information provided by the context. These contextual cues could include not only the spatial context (i.e. the 'where') and the temporal context (the 'when'), but also other aspects of context, such as internal state cues (hunger and fear state), and can be used to rapidly and flexibly alter valences of specific response options.
Recommendations: We also suggest that there is a separate, distinct, NMDA/GluR-A-independent mechanism through which the context can gradually (incrementally or decrementally) alter the valence of a particular response option.