Behavioral Roles

Long-term potentiation (LTP) refers the enhanced ability of a neuron to excite a neuron with which it is connected as a result of previous successful activation in the same pathway. LTP is typically produced by electrically stimulating a group of input neurons in a way that produces simultaneous and repetitive activation of their target neurons. Following this high frequency repetitive activation, future stimulation of the input neurons produces larger and more rapid activation of the target neuron. The phenomenon of LTP has many of the features of memory, including a lasting change in the responsiveness of neurons following brief experience with specific inputs, suggesting that LTP may be the cellular basis of memory. At the same time, one should not confuse LTP with memory. LTP is a laboratory phenomenon that involves massive levels of activation never observed in nature. Thus, the best we can hope is that LTP and memory share a common basis in cellular mechanisms.

This section will summarize some of the recent research on the possible linkage between LTP and memory, using two main approaches: demonstrations of changes in synaptic efficacy consequent to a learning experience and attempts to prevent learning by pharmacological or genetic manipulation of the molecular mechanisms of LTP induction.

Do Conventional Learning Experiences Produce Changes in Synaptic Efficacy Similar to Those That Occur after LTP?

Tying changes in synaptic physiology to learning seems like a daunting task because of the expectation that the magnitude of such observable changes in gross field potentials would be vanishingly small following any normal learning experience. In addition, it is likely that learning involves both positive and negative changes in synaptic efficacy, that is, both LTP and long-term depression (LTD). Thus learning could result in changes in the distribution of potentiated and depressed synapses with little or no overall shift and consequently no change or even an overall reduction in the averaged evoked potentials commonly used to measure LTP.

Despite these concerns, an early study reported enhanced excitability of the hippocampal perforant pathway in rats who had been exposed for prolonged periods to an "enriched" environment (group housing with toys) as compared to "impoverished" environment (solitary housing without toys). Environmental enrichment resulted in an increased slope of the synaptic potential and larger population action potentials, consistent with the pattern of increased synaptic efficacy observed following. Perhaps the strong and long duration of learning involved in environmental enrichment overcame the "needle-in-the-haystack" problem by enhancing the excitability of many hippocampal neurons.

More recently, Joseph LeDoux and his colleagues demonstrated LTP-like changes in neural responses in the amygdala. They trained rats to fear tones by presenting repeated pairings of auditory stimuli and foot shocks. Subsequently, in trained rats, conditioned tones produced evoked potentials of greater slope and amplitude similar to the characteristics of LTP observed in the auditory pathway to the amygdala. There was no observable enhancement when the same tones and foot shocks arrived separately, and are therefore not associated with one another. Furthermore, the synaptic enhancements observed in trained rats were enduring, lasting as long as the behavioral response.

John Donoghue and his colleagues extended this approach to the motor cortex and another form of learning. They trained rats to reach with a paw through a small hole in a chamber to retrieve food pellets. Following the training, Donoghue and his colleagues removed the brain and measured the strength of connections between cells within the area of the motor cortex that controls hand movements. They used an in vitro preparation and evoked synaptic potentials (EPSPs) in a principal cell layer of the motor cortex by stimulating horizontal fibers that connect neighboring cells to one another. They found that for the same or lower intensity of input stimulation, the magnitude of the EPSPs on the side of the brain that controlled the trained paw (i.e., in the contralateral hemisphere) were consistently larger than those on the side of the brain that controlled the untrained paw. Furthermore, they found it difficult to induce LTP by electrical stimulation in the trained hemisphere but not in the untrained hemisphere. Thus, training produced an anatomically localized increase in synaptic efficacy that occluded the capacity for LTP. These observations show that synaptic potentiation results from motor learning and that the real plasticity phenomenon shares common resources with the artificial one. This study provides strong evidence for common cellular mechanisms of LTP and learning.

Do Treatments That Block LTP Prevent Memory?

The major limitation of the foregoing approach is that the experiments only provide correlations between aspects of LTP and memory. The converse approach is to establish causal links between the LTP and memory by seeing whether memory is disrupted by blocking LTP with drugs or genetic manipulations. This approach appeared fruitful because of the assumption that the manipulations would target plasticity, not normal information processing in the brain, and that they would knock out a critical kind of plasticity. This assumption arose from the observation that drugs such as D-2-amino-5-phosphonovalerate (AP5) selectively block the NMDA receptor and thus prevent hippocampal LTP while sparing normal synaptic transmission—hence the expectation that, to the extent that the role of the NMDA receptor is restricted to plasticity, these drugs would indeed block new learning without affecting nonlearning performance or retention of learning normally accomplished prior to drug treatment.

Some of the earliest and strongest evidence supporting a connection between LTP and memory came from studies on spatial learning by Richard Morris and his colleagues. These studies exploited a watermaze task in which rats learn to find an escape hidden in a pool. Initially Morris and his colleagues showed that AP5 prevented new spatial learning in the water maze. Drug-treated rats swam normally but did escape as rapidly as the normal rats did. To assess the rats' knowledge of the escape locus, the researchers used probe tests in which they removed the escape site and measured swimming near the location of the former escape site. Untreated rats showed a distinct preference for swimming in the vicinity of the former escape locus, but drug-treated rats showed little or no such bias, indicating the absence of memory of the escape location. Further experiments showed no effect of AP5 on memory when training was accomplished prior to drug treatment. This is the expected result, because NMDA receptors are necessary only for the induction of LTP, not for its maintenance.

In other research by Morris and his colleagues have shown how NMDA-receptor-dependent LTP might play a continuing role in updating memory. To this end they varied the water-maze task by changing the location of the escape platform every day. The rats consistently found the platform very rapidly on the second trial on a given day. The animals were then tested with different memory delays inserted between the first and second trial on each day. On some days, AP5 was infused into the hippocampus, and on other days a placebo was given. AP5 treatment resulted in a deficit on the second trial. Moreover, this deficit varied with the duration between the first and second trials: there was no impairment with a fifteen-second intertrial interval, but significant deficits ensued with a delay of at least twenty minutes. These data suggest that memory for specific episodes of spatial learning remains dependent on NMDA receptors and LTP, even after the animals have learned the environment and the general rules of the spatial task.

Other studies suggest that the cascade of molecular events occasioned by LTP may also mediate the cortical plasticity that underlies memory. Yadin Dudai and his colleagues have focused on taste learning mediated by the gustatory cortex of rats. When rats are exposed to a novel taste and subsequently become ill, they develop a conditioned aversion to that taste, and this learning is known to depend on the gustatory cortex. AP5 produced an impairment in taste-aversion learning, whereas the same injections given prior to retention testing or into an adjacent cortical area had no effect. It is likely, then, that modifications in cortical taste representations rely on NMDA-dependent LTP.

Furthermore, the blockade of protein synthesis in the gustatory cortex by infusion of an inhibitor prior to learning also prevents development of the conditioned-taste aversion. MAP kinase and a downstream protein kinase were activated selectively in the gustatory cortex within ten minutes of exposure to a novel taste; activation peaked at thirty minutes, whereas exposure to a familiar taste had no effect. Conversely, an MAP kinase inhibitor retarded conditioned-taste aversion. This combination of findings strongly implicate NMDA-mediated plasticity and subsequent specific protein synthesis as critical factors in the cortical modifications that mediate this type of learning.

Other research has used targeted genetic manipulations to show that blocking the cascade of molecular triggers for LTP also results in severe memory impairment. In one such early study, mice with a mutation of one form of CaMKII had deficient LTP and were selectively impaired in learning the Morris water maze. Manipulation of biochemical mechanisms by interference with specific genes has allowed a highly specific identification of some of the critical molecular events. One study by Alcino Silva and his colleagues showed that substituting a single amino acid in CaMKII that prevents its autophosphorylation resulted in a severe learning and memory deficit. Other new genetic approaches are providing a greater temporal-and region-specific blockade of gene activation. Susumu Tonegawa and his colleagues created a genetic block that was limited to postdevelopment activation of the genes for the NMDA receptor in the CA1 subfield of the hippocampus. This mutation selectively blocked LTP in that region. Despite these highly selective temporal and anatomical restrictions, the mice with this mutation were severely deficient in spatial learning and other types of memory dependent on hippocampal function. A complementary recent study showed that a mutation that results in overexpression of NMDA receptors can enhance several kinds of memory dependent on the hippocampus. Molecular genetic manipulations increasingly indicate that interference with other aspects of the LTP molecular cascade, specifically PKC and MAPK, also impair memory. Thus it seems likely that the full set of cellular events that mediate LTP play critical roles in memory.

Conclusion

LTP and memory are not the same thing, and there is no universal acceptance of evidence for shared mechanisms between LTP and memory. And, notwithstanding some contradictory evidence not covered in this article, there is, nevertheless, compelling evidence that learning enhances synaptic potentials in circuits relevant to memory. There is correspondingly strong evidence that blocking LTP with drugs or genetic manipulations can impair memory and destabilize relevant neural representations.

See also:GENETIC SUBSTRATES OF MEMORY: HIPPOCAMPUS; GLUTAMATE RECEPTORS AND THEIR CHARACTERIZATION; LONG-TERM POTENTIATION; NEURAL SUBSTRATES OF EMOTIONAL MEMORY; TASTE AVERSION AND PREFERENCE LEARNING IN ANIMALS

Bibliography

Green, E. J., and Greenough, W. T. (1986). Altered synaptic transmission in dentate gyrus of rats reared in complex environments. Evidence from hippocampal slices maintained in vitro. Journal of Neurophysiology 55, 739-750.

Martin, S. J., Grimwood, P. D., and Morris, R. G. M. (2000). Synaptic plasticity and memory: An evaluation of the hypothesis. Annual Review of Neuroscience 23, 649-711.

Rioult-Pedotti, M.-S., Friedman, D., Hess, G., and Donoghue, J. P. (1998). Strengthening of horizontal cortical connections following skill learning. Nature Neuroscience 1, 230-234.

Rogan, M. T., Staubli, U. V., and LeDoux, J. E. (1997). Fear conditioning induces associative long-term potentiation in the amygdala. Nature 390, 604-607.

Rosenblum, K., Berman, D. E., Hazvi, S., Lamprecht, R., and Dudai, Y. (1997). NMDA receptor and the tyrosine phosphorylation of its 2B subunit in taste learning in the rat insular cortex. Journal of Neuroscience 17, 5,129-5,135.

Silva, A. J., Smith, A. M., and Giese, K. P. (1997). Gene targeting and the biology of learning and memory. Annual Review of Genetics 31, 527-547.

Steele, R. J., and Morris, R. G. M. (1999). Delay dependent impairment in matching-to-place task with chronic and intrahippocampal infusion of the NMDA-antagonist D-AP5. Hippocampus 9, 118-136.

Stevens, C. F. (1998). A million dollar question: Does LTP = memory? Neuron 20, 1-2.

Tang, Y-P., Shimizu, E., Dube, G. R., Rampson, C., Kerchner, G. A., Zhuo, M., Liu, G., and Tsien, J. Z. (1999). Genetic enhancement of learning and memory in mice. Nature 401, 63-69.

Tsien, J. Z., Huerta, P. T., Tonegawa, S. (1996). The essential role of hippocampal CA1 NMDA receptor-dependent synaptic plasticity in spatial memory. Cell 87, 1,327-1,338.

HowardEichenbaum