LTP has input specificity and associativity. It appears to be very selective, occurring only at synapses stimulated by a given pattern of afferent activity and not at neighboring synapses. The generation of LTP requires coordination of pre- and postsynaptic activity- synaptic input can only be potentiated if the receptor is active while the postsynaptic cell is depolarized . If both of these conditions are not met, LTP will not take place.
The NMDA receptor acts like a coincidence detector. Specifically, this receptor must detect the coordinated activity to induce LTP. The presynaptic activity, glutamate release and binding to the receptor site must occur with postsynaptic depolarization. This depolarization occurs through the activity of the AMPA receptor, which is a sodium/potassium (Na+/K+) ionotropic receptor. When the postsynaptic membrane is depolarized through Na+ influx, magnesium (Mg2+) is expelled from the channel and the receptor can pass current. The opening of the NMDA channel and the resulting calcium (Ca2+) influx is the critical event that determines whether LTP will occur. NMDA activation by strong input helps activate adjacent neurons. Depolarization by strong input in one neuron will help depolarize adjacent synapses if they are concurrently being activated by weaker input that would normally not be sufficient to activate them.
Much research has been performed to show what individual processes are necessary for LTP . Blocking NMDA receptor activity with an antagonist will prevent the induction of LTP. Additionally, mutant mice lacking functional NMDA receptors are unable to perform LTP. When the postsynaptic cell is filled with a chemical that prevents an increase in Ca2+ levels, LTP induction is prevented. Also, when levels of Ca2+ are directly increased,synaptic transmission is enhanced.
The question arises as to how this whole process leads to changes at the neuronal level as well as larger scale changes. The Ca2+ influx begins the activation of 2nd messenger systems, which broadcast the signal that the NMDA receptor generated throughout the cell. It has been suggested that protein kinases are very important in LTP induction. These two proteins are Ca2+/calmodulin-dependent protein kinase II (CaMKII) and protein kinase C (PKC). They are activated by the Ca2+ influx and are believed to increase the sensitivity to glutamate and number of AMPA receptors through phosphorylation . Additionally, through an unknown mechanism, they also may stimulate increased glutamate release presynaptically. Experimentally, these protein kinases importances have been shown to be necessary in the formation of LTP. When inhibitors were injected into the CA1 field, LTP was blocked. They have both been shown to be active during stimulation that leads to LTP. Direct injection of CaMKII into the CA1 field leads to synaptic enhancement as well. Although the actual role of these protein kinases is unsure, they appear to be instrumental in the process of LTP.
The phosphorylation that CaMKII and PKC can induce will change synaptic efficacy for a period of time, but the long-term changes of LTP require more. Activation of genes and new protein synthesis is likely the cause of these longer changes. These will lead to changes in the synaptic structure, strengthening it and allowing for future LTP to occur more easily. Due to these changes, LTP may result in larger responses to similar stimulation.
Extending LTP to translate into memory is important in showing the link between the two. If LTP is the cellular explanation for memory, preventing LTP should result in prevention or disruption of memory formation. This has been shown to be the case. Mice that have been genetically developed to lack NMDA receptors in the CA1 field demonstrate profound spatial learning deficits. Conversely, mutant mice with enhanced NMDA receptor function show better memory task performance. Much difficulty arises when trying to show individual neural connections being active and associated with learning. However, LTP-like activity has been observed in parallel with learning in fear conditioning paradigms. Experimentally shown, blocking NMDA receptors in the amygdala prevents the development of conditioned fear responses.
LTP occurs in phases. A single loop produces short-term changes (early LTP) that will last only a few hours. This does not require protein synthesis. Four or more loops will produce a more persistent phase (late LTP) that will last up to 24 hours and requires protein and RNA synthesis.