2.1. INTRODUCTION

Experience-dependent changes in behavior are thought to derive from lasting changes in synaptic strength and neuronal excitability. This chapter attempts an overview of the mechanisms underlying induction and maintenance of long-lasting activity-dependent synaptic modifications as at least one of the key mechanisms by which experiences modify neural circuit behavior. We consider the important yet complex regulation of long-term potentiation and depression by prior synaptic activity as well as the influence of some neurotrophins as regulators of synaptic efficacy crucial issues concerning the contribution of long-lasting forms of synaptic modification to learning and memory.

2.2. LONG-LASTING FORMS OF SYNAPTIC MODIFICATION: METHODOLOGICAL APPROACH

2.2.1. Long-Term Potentiation (LTP)

Tim Bliss and Terje Lomo^1,2 first reported the phenomenon of long-term potentiation (LTP), an increase in synaptic efficacy following synaptic activity, over 30 years ago. Since then, LTP has generated enormous interest as a potential mechanism of memory, primarily because it exhibits numerous properties expected of a synaptic associative memory mechanism, such as rapid induction, synapse specificity, associative interactions, persistence, and dependence on correlated synaptic activity. Given these important features, LTP remains only a model of the synaptic and cellular events that may underlie memory formation. Although LTP-like phenomena are appealing as models of the synaptic changes in learning and memory, one question crucial to this research is the contribution of LTP to learning and memory. While this issue has been the subject of numerous behavioral studies since LTP was discovered, more recent studies combining advanced methodologies offer new avenues and compelling data that offer the potential for defining the role of LTP in hippocampal function, learning, and memory.

Many features of LTP as a phenomenon make it a compelling candidate for the synaptic processes underlying neural information storage. First, LTP is induced rapidly. Soon after its induction (which usually involves postsynaptic depolarization induced by high frequency stimulation of a sufficient number of afferents), LTP appears within minutes. In some situations, the actual development of LTP is obscured by short-term potentiation (STP), an exponentially decaying increase in synaptic strength that appears to involve NMDA receptors as well,³ but decays within 5 to 20 min (Figure 2.1). Prior evaluations of STP and LTP determined the onset of LTP by subtracting STP. Hanse and Gustafsson⁴ suggest that it develops incrementally, reaching asymptotic levels by approximately 5 to 20 sec, depending upon the synapse studied.

FIGURE 2.1

LTP and heterosynaptic LTD in the dentate gyrus. Point plot representing magnitude of extracellular synaptic responses (slope of field EPSP, mV/ms) in awake behaving rats evoked once every minute. High frequency stimulation of the lateral perforant pathway (more...)

LTP is not always rapidly expressed and can show incremental growth over a period of 10 to 20 min. The precise reasons why such incremental LTP is observed in some cases and not others is unknown.

Our experience is that the methodology used to induce LTP can determine initial LTP magnitude. For example, rapid LTP induction is seen with direct stimulation of afferents in both commissural and perforant path inputs to the CA3 region. However, a slowly developing, incremental LTP often is observed when LTP is induced in an associative manner by pairing weak commissural or perforant path trains with a strong tetanus to a convergent CA3 afferent system.⁵ Thus, while LTP develops relatively rapidly, it can take some time to develop fully. In line with these observations, the emergence of both place cells in the CA1 region and the appearance of place field expansion, a gradual increase in the place field that reflects earlier firing of place cells with experience, also occur over a duration of 5 to 10 min. This is a crucial point because these processes are thought to involve associative synaptic changes and possibly LTP.^6,7

It should be noted that STP is not necessary for LTP to develop, and LTP is often observed even in the absence of STP. Our own experience is that STP is a phenomenon that, like rapidly induced LTP, is seen in vivo only following rather extended intense trains. For example, STP is observed following 0.5- to 1-sec bursts of 400 Hz, whereas STP is rarely observed in perforant path projections with brief bursts of stimulation given at 200 msec intervals that mimic theta rhythm (or so-called theta burst stimulation.⁸). Thus it remains to be determined whether STP reflects a process that can normally occur with endogenous patterns of synaptic activity, and, if so, whether it has any role in learning or memory.

Another feature is that LTP is associative. If high frequency stimulation of one set of afferents induces LTP, individual active synapses can also be recruited to express LTP — provided that the synapse is coactive within a delimited window. Associativity can be derived from the requirements for activation of the NMDA receptor (specifically, both glutamate and postsynaptic depolarization essential for relieving the magnesium block of the NMDA channel).

The property of associativity can be derived directly from and is essentially identical to the property of cooperativity,⁹ indicating that LTP has a threshold and a threshold number of afferents must be active to induce LTP. That LTP can be induced in “cooperating” synapses is indicative that associativity is an inherent property of this phenomenon. While it has been suggested that the term associativity refers to cooperative LTP involving distinct sets of afferents, it should be noted that this first demonstration of cooperativity used two distinct afferent systems to the dentate gyrus: the medial and lateral perforant pathways.⁹ However, because interactions among a critical number of afferent fibers may underlie this effect, it is probable that the cooperativity term was chosen out of prudence since the associativity term as envisaged by Hebb tacitly implies a post-synaptic integration of presynaptic activity at the postsynaptic element — something that cannot be ruled out in the cooperativity effect. Perhaps a more realistic and convincing demonstration of associativity is provided by studies showing that the pairing of low frequency afferent activity with induced postsynaptic depolarization also induces LTP.¹⁰

Input specificity is a crucial property of virtually all forms of LTP and refers to the fact that LTP is synapse-specific and restricted only to synapses of activated afferents (Figure 2.1). This is in contrast to the nonassociative phenomena such as sensitization in which specificity of stimuli (or afferent input) is absent and a more general facilitation of responses is observed in other afferent inputs. Obviously, input specificity is an important feature given that information storage capacity is increased when plasticity can be regulated at individual synapses.

Another feature of LTP is that it is remarkably persistent. Prior to its discovery, the only activity-dependent electrophysiological change that came close to LTP in duration was post-tetanic potentiation (PTP) a primarily presynaptic phenomenon lasting from seconds to minutes. By contrast, LTP in the hippocampal formation can persist from hours to weeks or months, depending upon the stimulation parameters.

In intact animals, LTP is decremental and usually decays within 1 to 2 weeks.¹¹ While this is certainly too brief a period for the storage of long-term memory, several points should be made with respect to LTP longevity. First, LTP in the hippocampus need not be permanent. Current findings support the view that, as is suspected in humans, the hippocampus has a time-delimited role in memory; persistent long-term memory is gradually consolidated in neocortical areas.¹² In this view, memories formed by the hippocampus are transferred to and consolidated in the neocortex, possibly during slow wave or REM sleep states.^13,14 This usually occurs within 2 weeks in rats, as indicated by both lesion and imaging data.¹⁵ Thus, if the hippocampus indeed serves as a temporary repository of information, LTP may not last long simply because it may not need to. In fact, any long-term retention of information within the hippocampus beyond the usual several weeks may even be detrimental to hippocampal-based memory, resulting in interference with previously stored patterns of synaptic activity.

How long then can LTP last? It is reported that LTP decay is an active process mediated by NMDA receptors, and blocking these receptors can prevent LTP decay.¹⁶ Although it is not known for certain, many forms of long-term depression (LTD) — which, like LTP is a long-term change in synaptic strength — reduces synaptic strength. Like LTP, LTD usually requires the activation of NMDA receptors. Thus LTD may mediate LTP decay. However, because LTP decay is an active process that requires NMDA receptor activation (and possibly the induction of LTD), LTP is theoretically, permanent and may last as long as the synapse itself, provided that it is not erased by subsequent synaptic activity and NMDAR activation.

2.3. LTP AND LTD: TRIGGERING, EXPRESSION, AND MAINTENANCE MECHANISMS

While LTP is often equated with memory, it is nothing but a model. In this case, LTP is thought to reflect an artificially induced manifestation of the cellular processes that occur during normal synaptic transmission and that normally underlie the synaptic changes that mediate memory. As such, it is often criticized on methodological grounds, but often for the wrong reasons. For example, it is noted that LTP is often induced using massive stimulation of a single afferent system, an event that is decidedly nonphysiological and unlikely to occur in anything but pathological states. However, LTP need not be induced this way; simply pairing low frequency afferent activity with postsynaptic depolarization or cell firing is sufficient to induce LTP.³⁵ Massive afferent stimulation usually is employed as a method to induce LTP or LTD, merely to achieve levels of postsynaptic depolarization necessary to activate NMDA receptors. As noted with the studies of cooperativity, a critical number of fibers must be activated to induce LTP.⁹ It has been estimated that in order to reach the LTP experimentally, a sufficient number of afferent inputs must be activated. Experimentally, it appears that this threshold is close to the stimulation intensity necessary to bring granule cells near their thresholds for firing, as indicated by the evocation of a “population spike” in field recordings.⁹ Extrapolating from this, it is likely that the postsynaptic depolarization necessary for eliciting LTP is near the cell threshold for eliciting an action potential. The fact that actively firing principal cells are frequently observed in behaving animals shows that the levels of post-synaptic depolarization necessary for NMDA receptor activation and LTP induction likely occur quite frequently during normal hippocampal operation. Thus it is likely that LTP is not a rare event and is likely to occur during normal hippocampal functioning.

An astonishing amount of effort has been put forth in an attempt to delineate the molecular mechanisms of LTP induction and expression. However, any delineation of processes must be tempered by the knowledge that LTP has a variety of forms, and any single molecular cascade is unlikely to reflect the diversity of processes mediating LTP, even among synaptic populations in the hippocampus. Roughly, a general scheme for LTP induction can be described for the factors that are important for LTP induction. These include an increase in postsynaptic depolarization³⁶ and an increase in postsynaptic calcium.³⁷ This is usually provided by the NMDA glutamate receptor, although postsynaptic calcium appears to be critical for LTP induction even at synapses that display LTP not mediated by NMDAR activation.^38,39 The subsequent activation of a number of kinases including calcium–calmodulin kinase II (CamKII), MAP kinase and ERK, PKA and PKC are also implicated in LTP development, although it should be noted that over 100 molecules have been implicated in the processes collectively described as LTP.⁴⁰

The roles of these kinases are twofold, with the first possibility remaining that rapid effects are due to direct actions of kinases and phosphorylation of AMPA subunits⁴¹ that may affect channel properties directly or may be essential for AMPA subunit trafficking.⁴²

The second role is phosphorylation of transcription factors, such as CREB.⁴³ Subsequent protein synthesis initiated by these transcription factors is thought to result in the synthesis of mRNA and proteins at the soma and the eventual targeting of these molecules to the potentiated synapse. It should be noted that an extensive amount of protein synthesis occurs locally within dendritic regions shortly after LTP induction. Although it is possible that these locally translated proteins participate directly in enhancing synaptic strength, they may serve as synaptic “tags”⁴⁴ necessary for targeting somatically synthesized proteins to potentiated synapses.

Is LTP a singular phenomenon? Although the LTP term often is used collectively to describe any increase in synaptic strength, it has become clear that LTP can be expressed at many synapses in the nervous system following high frequency synaptic activity. While many forms of LTP involve NMDA receptors, this is not always the case.⁴⁵ Furthermore, the cellular mechanisms implicated in LTP induction differ even among synapses within the hippocampal formation. For example, the expression of the immediate early genes (IEG) Arc, Homer, and Zif268 (egr-1) is observed with LTP induction in the dentate, but is not observed following LTP induction at Schaffer-CA1 synapses.⁴⁶ It thus appears that LTP may actually be a collection of synaptic phenomena that, while appearing similar in phenomenology, utilize distinct mechanisms of induction and possibly expression in different synaptic populations.

Many now refer to distinct forms of LTP. Likewise, LTD displays distinct forms that can be expressed at many different synapses by a variety of receptor mechanisms.³⁸ The assignment of forms of LTP must be distinguished from temporal phases of LTP. Following NMDAR activation and LTP induction, the initial or early phase of LTP (E-LTP) requires the activation of NMDARs, but decays relatively rapidly. The later (> 3 hr) phases of LTP (L-LTP) are dependent upon protein synthesis.^47–49 Thus NMDAR-dependent LTP appears to be expressed via a concatenation of processes that mediate LTP maintenance.

The assignment of forms of LTP has been rooted in the different mechanisms of induction and the particular synapse under study. Thus it is not too speculative to suggest that these distinct forms of LTP induction also involve distinct molecular mechanisms of expression and maintenance. Conversely, it is also possible that the various forms of LTP induction seen in the hippocampal formation may simply reflect differences only among induction mechanisms and yet share common downstream maintenance mechanisms. In this case, the distinct mechanisms of LTP induction may serve distinct functions by defining the precise synaptic conditions necessary to induce LTP in a particular synaptic system. For example, sustained, protein synthesis-dependent LTP observed at mossy fiber synapses appears to require high frequency mossy fiber activity in order to provide postsynaptic depolarization,⁵⁰ postsynaptic calcium,⁵¹ and the release of opioid peptide co-transmitters.⁵²

As peptide co-transmitters often require repetitive presynaptic activity for their release from dense core vesicles, LTP at this synapse appears to display a strict requirement for high frequency mossy fiber activity.⁵² As bursting in granule cells is also a rare and highly regulated event,⁵³ this novel induction requirement may serve to impose tight constraints on plasticity in this sparse synaptic system. It is easy to imagine that such constraints may be necessary to maintain sparse activity among mossy fiber input lines, given that one principal function of the mossy fibers is thought to be the encoding of sparse orthogonal representations in the CA3 region.^54,55

That LTP involves multiple effector systems, kinases, and genes leads to important questions: are each of these kinases essential for establishing or maintaining the increase in synaptic strength? Or do some of these factors regulate other aspects of LTP such as its persistence? Or do some processes mediate metaplastic effects that regulate LTP or LTD thresholds following prior activity?⁵⁶ All these questions along with the view that LTP may involve a number of distinct induction and expression mechanisms indicate that the mechanisms of LTP are not only complex, but are unlikely to be delineated into a single sequential molecular process. Methodological issues also complicate analysis of LTP because the techniques used to induce LTP (namely high frequency afferent activation) are also likely to induce other processes and other effector systems unrelated to LTP expression. For example, LTP can be induced in the mossy fibers following high frequency stimulation. However, this same stimulation also induces mossy fiber synaptogenesis⁵⁷ and increases granule cell neurogenesis.⁵⁸ In such cases, analysis of the kinases and genes expressed presents a number of difficulties. Which genes generated by stimulation are essential for LTP as opposed to neurogenesis or synaptogenesis? Is mossy fiber synaptogenesis a distinct process, or is it perhaps the end point of mossy fiber LTP? It appears that the concomitant characterization of other synaptic processes altered with high frequency activity will be essential to dissect the myriad cellular processes induced by activity necessary for the induction and expression of LTP.

Further work will no doubt elaborate on the differences in the molecular machinery underlying LTP induction and expression, and make us reconsider classification of LTP based only upon only induction mechanisms. It should be noted that the majority of studies addressing LTP expression and maintenance have looked at time points under 3 hours. Because this is the time period prior to protein synthesis that is essential for sustained LTP,^47–49 it is entirely possible that the mechanisms underlying LTP expression and maintenance after 3 hours and even after several days, may be completely different. Thus, not only may the mechanisms of LTP maintenance differ among the various forms of LTP induction, but LTP at a given synapse may involve a concatenation of numerous maintenance processes over time that may change as LTP persists from hours to days.

It is also premature to assign LTP forms based on the specific synaptic population under study. Distinct forms of LTP expression can be observed at a given synapse, depending on induction variables such as the high frequency stimulation used or the behavioral state of the animal during LTP induction. In the former case, it is reported that high frequency stimulation that mimics the natural theta rhythm produces a “nondecremental” form of LTP at Schaffer-CA1 synapses,⁸ possibly by inducing multiple forms of LTP.^59–61 In the latter case, the temporal characteristic of LTP can depend on behavioral state of the animal when LTP is induced. While LTP in perforant path–dentate synapses typically persists for 5 to 7 days following a single session of stimulation,¹¹ the duration of LTP is extended to about 2 weeks if dentate LTP is induced while the animal is actively engaged in learning (such as during the initial exploration of a novel environment.⁶²)

The mechanisms underlying this effect remain to be elucidated, although it appears that monoamines likely play a role in this facilitating LTP longevity.^63,64 In both cases, when LTP is induced with theta bursts or induced in novel environments, it displays a distinct extended time course. This tacitly implies that distinct molecular mechanisms (or the modulation of a common molecular mechanism) mediate LTP maintenance when induced in these conditions. Not only do different synapses display different forms of LTP, depending on behavioral state or means of induction, but a given synapse may display distinct forms of LTP, depending upon the behavioral conditions during induction. The present challenge is to determine necessary mechanisms common to the different forms of LTP and the possible functional roles of distinct forms of LTP induction or maintenance normally occurring in the normal operation of the hippocampal formation in learning and memory.

2.4. PERSISTENT SYNAPTIC PLASTICITY: METAPLASTIC POINT OF VIEW

The most remarkable property of synapses lies in their capacity to modify the efficiency with which they transmit information from one neuron to another. This property, known as synaptic plasticity, is the basis of information storage in the brain. It enables us to store and use vast amounts of information in the form of learned behaviors and conscious memories.

Synaptic plasticity can be modulated, sometimes dramatically, by prior synaptic activity; this property is named metaplasticity.⁵⁶ It is induced by synaptic or cellular activity, but it is not necessarily expressed as a change in the efficacy of normal synaptic transmission. Instead, it is manifest as a change in the ability to induce subsequent synaptic plasticity, such as long-term potentiation or depression. Thus, metaplasticity is a higher order form of synaptic plasticity.

Another mechanism that could help maintain relatively constant activity levels is if neurons increased the strength of all excitatory connections in response to a prolonged drop in firing rates and vice versa. Such bidirectional plasticity of synaptic currents has recently been demonstrated in cultured cortical and spinal networks and occurs through a scaling up or down of the strength of all of a neuron’s excitatory inputs. This form of plasticity has been termed synaptic scaling.⁶⁵

Classical Hebbian plasticities (such as LTP and LTD) that are rapid and synapse-specific coexist with other long-lasting modifications of synapses (such as metaplasticity and synaptic scaling) that work over longer time scales and are crucial for maintaining and orchestrating neuronal network function. Metaplasticity and synaptic scaling are parts of the homeostatic plasticity mechanisms that stabilize neuronal activity. Bienenstock, Cooper, and Munro put forth one proposal (the BCM theory) to account for such homeostatic regulation in their model of visual cortical receptive field plasticity during development.⁶⁶ They suggested that the “modification threshold” orΘ m (level of post-synaptic response below which gives LTD and above which gives LTP) is dynamically regulated by the average level of post-synaptic activity (Figure 2.2).

FIGURE 2.2

Bidirectional modification of synaptic plasticity as predicted by BCM theory. NMDA-dependent concentration of Ca²⁺ or frequencies of conditioning stimulation represent experimental approaches through which the function can be reached. (Modified from Abraham, (more...)

For example, if visual cortical neurons suffer prolonged reductions in their activity because of visual deprivation, then the modification threshold would be correspondingly reduced. This adaptive response allows the preservation of a broad range of LTP and LTD responses despite treatments that restrict the firing repertoires of those neurons. A converse process involving an elevated modification threshold occurs if a neuron’s level of activity is increased over a prolonged period.

It is now widely accepted that the trafficking of AMPA-type glutamate receptors mediates rapid synaptic modification in the classic Hebbian forms of plasticity, LTP and LTD. Several other cellular and molecular changes have been implicated in synaptic homeostasis but one common feature of many forms of homeostatic plasticity is an alteration in the number or complement of NMDA-type glutamate receptors. A series of new studies has revealed that NMDA receptors cycle rapidly into and out of synapses and that regulated trafficking of NMDA receptors working cumulatively and over longer time scales can effectively modify the number and composition (NR2A/NR2B subunit ratio) of synaptic NMDA receptors as demonstrated in the visual cortex by Quinlan et al.⁶⁷ and Philpot et al.⁶⁸ Thus, an emerging concept is that activity-dependent alterations in NMDA receptor trafficking contribute to homeostatic plasticity at central glutamatergic synapses (Figure 2.3).

FIGURE 2.3

Experience-dependent changes in NMDA receptor subunit composition may underlie the sliding synaptic modification threshold. (a) Dynamic trafficking of NMDA receptor composition. (b) Model relating NMDA receptor subunit switch to changes in synaptic plasticity (more...)

Experience-dependent regulation of synaptic strength has long been hypothesized to be the physiological basis of learning and memory. Accordingly, an increase in synaptic strength accompanies learning in vivo as demonstrated by Rogan et al. in the amygdala⁶⁹ and persists in brain slices ex vivo as shown by the observations of McKernan and Shinnick-Gallagher in the same pathway⁷⁰ and by Rioult-Pedotti et al. in the motor cortex.⁷¹ Learning-induced potentiation of synaptic strength is also accompanied by an increase in the threshold for further synaptic enhancements.

These dual changes in synaptic function are thought to initiate and maintain the memory encoded by experience. In this regard, studies of the insular cortex (IC), a region of the temporal cortex implicated in acquisition and retention of conditioned taste aversion (CTA), demonstrated that induction of LTP in the basolateral amygdaloid nucleus (BLa)-IC projection previous to CTA training enhances the retention of this task.⁷² We recently showed that CTA training prevents the subsequent induction of LTP in the BLa-IC projection for at least 120 hours (unpublished data). These findings provide evidence that CTA training produces a change in the ability to induce subsequent synaptic plasticity on the BLa-IC pathway, supporting the notion that the mechanisms responsible for behaviorally induced synaptic changes are similar to those underlying electrically induced LTP. Accordingly, the activity history of a given neuron influences its future responses to synaptic input.

2.5. ROLE OF ACTIVITY-DEPENDENT SYNAPTIC PLASTICITY IN BRAIN FUNCTION

While NMDAR-dependent LTP and LTD in the CA1 region of the hippocampus remain the most extensively studied and therefore prototypic forms of synaptic plasticity, it is now clear that additional forms of LTP and LTD may share some, but certainly not all, of the properties and mechanisms of NMDAR-dependent LTP and LTD. Therefore, when discussing LTP and LTD, it is necessary to define at which specific synapses these phenomena are being studied, at what time point during development, and how they are triggered. Indeed, it may be most useful to conceptualize LTP and LTD as a general class of cellular/synaptic phenomena.⁷³

LTP and LTD are experimental phenomena that can be used to demonstrate the repertoire of long-lasting modifications of which individual synapses are capable. Given the ubiquity of various forms of LTP and LTD at excitatory synapses throughout the brain, it seems virtually certain that the brain takes advantage of neuronal capability to express long-lasting activity-dependent synaptic modifications as at least one of the key mechanisms by which experiences modify neural circuit behavior. Largely through correlational studies involving genetic and pharmacological manipulations, it is possible to begin to establish that in vivo experiences generate synaptic modifications analogous to LTP and LTD and these modifications are required for the behavioral or cognitive plasticity generated by the experience.

As mentioned above, the BCM theory suggests that the depression of deprived-eye inputs in visual cortex is specifically triggered by presynaptic activity when it fails to consistently correlate with a strong evoked postsynaptic response. Indeed the BCM theory was the motivation behind the ultimately successful search for homosynaptic LTD in the hippocampus and visual cortex. The idea that depression of responses in visual cortex is actually caused by activity in the deprived eye was tested experimentally by Rittenhouse et al.⁷⁴ Additional evidence that the mechanisms of LTD underlie sensory deprivation has come from studies of the somatosensory^75,76 and visual cortices.⁷⁷

Naturally occurring response potentiation can also be observed in the sensory cortex. For example, chronic recordings from adult mouse visual cortex have shown that closing one eye enables a gradual experience-dependent enhancement of the responses to stimulation of the other eye. The effect persists for many days after opening the deprived eye and fails to occur in mice with reduced expression of NMDA receptors in the superficial layers of the visual cortex.⁷⁸

It has become apparent that the neural mechanisms underlying adaptive forms of learning and memory likely also play a critical role in the pathophysiology of addiction.^79,80 In this regard, it has been found that administration of single doses of several classes of drugs of abuse caused significant increases in synaptic strength at excitatory synapses on mesolimbic dopaminergic cells that are critical for mediating several forms of long-lasting, drug-induced behavioral plasticity. These increases share mechanisms with LTP, involving for example the up-regulation of AMPA receptors.^81,82 Additionally, drugs of abuse can also modify the triggering of LTP and LTD of the mesolimbic system. Thus, long-lasting synaptic plasticity is thought to be a principal mechanism by which functional properties of the nervous system are expressed.

2.6. SUBSTRATES OF LTP AND LTD: STRUCTURAL PLASTICITY

Since LTP and LTD appear to play a role in plasticity in a variety of behaviors and structures, what are the consequences of this in the induction of LTP and LTD and the cascade of numerous kinases thought to be involved? Studies over the past 10 years have attempted to address the locus of LTP and focused on the locus of change (pre- or postsynaptic). It is generally accepted that postsynaptic mechanisms are essential for LTP induction, and the primary putative mechanisms that serve to maintain increased synaptic responses also involve the postsynaptic element.

Evidence currently implicates postsynaptic mechanisms in LTP induction and expression. In addition, even though many of the effects revealed by quantal analysis suggest presynaptic changes, considerable evidence indicates that these apparent presynaptic changes reflect a postsynaptic recruitment of “silent” synapses. Silent synapses possess only NMDA receptors, possibly the NR2B heteromers that have sustained calcium conductance. It is thought that the activation of NMDA receptors by activity and ambient glutamate, presumably provided heterosynaptically by active neighboring synapses, may initiate the translocation of calmodulin kinases to post-synaptic density, allowing for the assembly of functional receptors and the conversion of silent synapses to functional synapses that express both NMDA and AMPA glutamate receptors.⁸³ Currently, the trafficking of subunits and alterations of AMPA and NMDAR receptor stoichiometry at existing functional synapses will turn out to be one of the most compelling and comprehensive of the current models of LTP expression.^84,85

While silent synapses may explain the apparent presynaptic effects seen with quantal analysis, the abundance of more recent data suggesting presynaptic changes^86–90 cannot be ignored. The most parsimonious view is that changes take place both pre- and postsynaptically. Such a view is prudent since most studies of the mechanisms of LTP maintenance and expression have been restricted to the earliest time points in LTP expression (usually within 1 hour after its induction). As noted above, it is quite possible that the later protein synthesis-dependent phases that maintain LTP for days and weeks may use entirely different mechanism for later LTP expression, including presynaptic mechanisms.

One mechanism that appears crucial for both LTP and LTD expression involves AMPA receptors — multimeric ionotropic receptors that serve as the principal receptors mediating fast excitatory synaptic transmission at glutamatergic responses. Currently, the primary candidate mechanism for the increases and decreases is seen with LTP and LTD is thought to involve trafficking of AMPA subunits and the formation and alteration of AMPA glutamate receptors. These multimeric ionotropic receptors are the principal receptors that mediate fast glutamatergic responses. Of particular interest are the GluR1 and GluR2 subunits of this receptor that appear to be regulated in several ways. Also crucial is the activation of the type II calcium–calmodulin kinase (CamKII), a kinase that requires both calcium and calmodulin. A process thought important in both these mechanisms is a translocation of CaMKII to postsynaptic density^91,92 that can dramatically alter both its activity and its regulation by both calcium and calmodulin.

The mechanism of LTP expression that has received the most interest is AMPA receptor trafficking, an idea that likely originated with the discovery of subunit trafficking following the activation of silent synapses. It has been shown that after LTP, AMPA subunits are inserted or move laterally into the PSD immediately following LTP induction.⁹³ Conversely, these receptors are internalized via endocytosis with LTD.⁹⁴ In this process, it appears that both GluR1 and GluR2 subunits are transported with LTP induction, whereas GluR2 trafficking may occur alone during normal processes of synaptic homeostasis.⁸⁵ It is of historical note that in 1983 Gary Lynch and Michel Baudry presented a “new and specific” hypothesis of LTP, suggesting that LTP was a postsynaptic phenomenon involving the expression of new receptors into the postsynaptic density. This hypothesis was not only prescient, but notable in that Lynch and Baudry based their molecular hypothesis on established cellular mechanisms that had not been considered in terms of synaptic function.⁹⁵

A crucial component of this process appears to be CaMKII binding to the PSD via interaction with the NR2B NMDA receptor subunit that can “lock” CamKII in an active conformation.⁹¹ In this way, CamKII is in an ideal position to phosphorylate existing AMPA subunits. Alternatively, the CaMKII molecule may be part of a more extensive scaffolding device attached to the NR2B subunit that promotes translocation of AMPA subunits into synaptic regions and the PSD.⁹⁶ Although both subunit phosphorylation and receptor trafficking mechanisms are viable candidates for mechanisms that underlie LTP and LTD, perhaps these two mechanisms, AMPA phosphorylation and AMPA insertion, represent two mechanisms or forms of LTP within a heterogeneous synaptic population and reflect potentiation for extant and silent synapses, respectively. Another alternative is that such phosphorylation works in tandem with CaMKII to allow rapid insertion or removal of phosphorylated subunits.⁹⁷

In addition to AMPA subunits and receptors, another aspect of trafficking is the insertion of NMDA receptors following LTP induction. While such an effect may play a crucial role in increasing synaptic efficacy, it is possible that the dynamic modification of NMDA receptors may play other roles in inducing or maintaining plasticity. For instance, alterations of NMDARs may mediate metaplastic effects (see above) that allow the alteration of LTP or LTD thresholds in response to recent synaptic activity.⁹⁸

2.7. NEUROTROPHINS AND SYNAPTIC PLASTICITY

Neurotrophins constitute a family of structurally related proteins that includes nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), and neurotrophin 4/5 (NT-4/5) identified in the mammalian brain. The signaling and biological functions of these molecules are mediated primarily by the Trk receptor tyrosine kinases. NGF binds to TrkA; BDNF and NT-4/5 to TrkB; and NT-3 to TrkC (Figure 2.4).

FIGURE 2.4

Neurotrophins and their ligand binding preferences for each member of the TrK receptor family.

Neurotrophins play diverse roles in regulating neuronal structure, function, and survival during development and into adulthood. One unexpected facet of neurotrophin actions was elucidated in the early 1990s, when Lohof et al.⁹⁹ discovered that exogenously applied BDNF and NT-3 enhanced synaptic efficacy at the Xenopus neuromuscular junction. Since then, experiments from many laboratories have demonstrated that neurotrophins indeed play important roles in synaptic development and plasticity.

Recent studies suggest that one of the neurotrophins, BDNF, plays a critical role in long-term synaptic plasticity in the adult brain. BDNF acutely enhances glutamatergic synaptic transmission and increases phosphorylation of the NR2B subunit of the NMDA receptor in postsynaptic densities isolated from cortex and hippocampus. NMDA receptor-dependent long-term potentiation is associated with the up-regulation of BDNF and its TrkB receptor in the hippocampus of awake freely moving rats, as well as in hippocampal slices. In BDNF knockout mice, LTP is markedly impaired, but the changes are restored by either adenovirus-mediated transfection or by the bath application of BDNF.

Moreover, the application of exogenous BDNF to slices enhances electrically induced LTP in rodent hippocampus. In a series of studies using bath perfusion of BDNF onto hippocampal slices, Schuman and colleagues demonstrated long-lasting enhancement of transmission at Schaffer collateral-CA1 synapses.¹⁰⁰ Furthermore, using intrahippocampal infusions of BDNF in intact rats revealed a robust long-lasting potentiation at perforant path–granule cell synapses in the dentate gyrus.¹⁰¹ Thus, a combination of genetic and pharmacological approaches has revealed multiple and distinct contributions of BDNF signaling to LTP.

These actions may be classified as permissive or instructive.¹⁰² Permissive refers to effects of BDNF that make synapses capable of LTP in the first place, but which are not causally involved in generating LTP. In contrast, instructive refers to BDNF signaling that is initiated in response to HFS and causally involved in the development of LTP. Recent findings suggest that BDNF is a key protein synthesis product needed to carry on the necessary functions for long-lasting modification of hippocampal synapses¹⁰³ and drives the formation of stable protein synthesis-dependent LTP, a process referred to as synaptic consolidation.¹⁰²

BDNF and its TrkB high affinity receptor are also abundantly expressed in neurons of the neocortex of the mammalian brain. BDNF also influences the development of patterned connections and the growth and complexity of dendrites in the cerebral cortex. Electrophysiological recordings made in slices of visual cortex have implicated BDNF in synaptic plasticity during the critical period for the formation of ocular dominance columns.¹⁰⁴ BDNF participates in the activity-dependent scaling of cortical synaptic strengths and adjusts the relative balance of cortical excitation and inhibition. BDNF enhances the magnitude of LTP induced by tetanic stimulation in visual cortex and also induces long-lasting potentiation of synaptic transmission in the visual cortices of young rats.¹⁰⁵

It is considered that memory process involves short- and long-term changes in synaptic communication, including alterations in electrical properties and structural modifications. Since BDNF significantly modulates both forms of synaptic changes, it may play an important role in learning and memory. In this regard, it has been demonstrated that spatial learning is associated with an increase in BDNF mRNA levels in the hippocampus. Furthermore, the expression of its TrkB receptor in the hippocampus was selectively increased in response to spatial learning. Tokuyama et al.¹⁰⁶ recently reported that BDNF mRNA is up-regulated in the inferior temporal cortex during the formation of visual pair association memory in monkeys. Previous studies of one of the authors demonstrated that acute intra-cortical microinfusion of BDNF in anesthetized adult rats induced a lasting potentiation of synaptic efficacy in the insular cortex, an area that has been implicated in the acquisition and storage of different aversive learning tasks.¹⁰⁷ Moreover, as mentioned above, we showed that induction of LTP in the Bla-IC projection previous to CTA training enhanced the retention of this task.⁷² In a similar manner, recently we showed that acute intracortical microinfusion of BDNF (in a concentration capable of inducing a lasting potentiation of synaptic efficacy in the insular cortex) enhanced the retention of CTA.¹⁰⁸

Although the signaling pathways of BDNF/TrkB activation in learning and memory formation remain to be determined, modulation of NMDA and non-NMDA receptor functions and the expression of synaptic proteins required for exocytosis may be implicated. The mitogen-activated protein kinase (MAPK) signaling pathway appears to be involved in pre- and postsynaptic BDNF actions.¹⁰⁹ Studies of Minichielo et al.¹¹⁰ provide genetic evidence that TrkB mediates hippocampal plasticity via recruitment of phospholipase C (PLC) and by subsequent phosphorylation of CAMKIV (calcium–calmodulin-dependent kinase IV) and CREB (cAMP response element-binding protein). Recent experimental evidence indicates that BDNF/TrkB signaling converges on MAPK pathways through the activation of extracellular-regulated kinase (ERK) to enhance excitatory synaptic transmission in vivo as well as hippocampal-dependent learning in behaving animals¹¹¹ (Figure 2.5).

FIGURE 2.5

Signaling pathways of BDNF/TrkB activation in learning and memory. Activity increases BDNF expression and release from pre- and postsynaptic sites. BDNF binds to TrkB receptors located on presynaptic axons, leading to its auto-phosphorylation and activation (more...)

Taken together, this evidence suggests that BDNF and other neurotrophins may represent a class of neuromodulators that regulate activity-dependent synaptic efficacy.

2.8. EXPERIENCE-DEPENDENT MODIFICATIONS: IS LTP INVOLVED IN LEARNING AND MEMORY?

Although most research into LTP and LTD focused on the cellular and molecular mechanisms surrounding its expression, these questions tacitly assume LTP is involved with learning. However, the cellular mechanisms underlying the induction and expression of LTP and LTD may be entirely moot if these phenomena are not involved in information storage in any appreciable way. Thus, principal among the many issues that surround LTP is whether LTP contributes to learning and memory. A definitive demonstration is lacking, and no one has yet seen a synapse display LTP with learning or a loss of synaptic LTP with forgetting. Clearly, such a demonstration presents technical difficulties we have yet to overcome. However, the convergence of confirmatory studies undertaken in the last 20 years supplies compelling evidence for the view that LTP is involved in learning and memory.

Morris and McNaughton⁵⁵ proposed three crucial experimental findings that are necessary to establish LTP as a memory mechanism. First, blocking LTP induction should impair acquisition without impairing retention of previously encoded information. Second, if LTP serves as a mechanism of encoding, saturating LTP should impair acquisition of new memories. Third, the selective erasure of LTP should disrupt the retention of established memories, but not impair acquisition of new information.

The first postulate suggests that blocking LTP induction should impair acquisition without impairing retention of previously encoded information. The pioneering work by Morris and colleagues^112,113 using NMDA antagonists has remained one of the most solid collections of studies implicating LTP in learning. In sum, it appears that both LTP induction and spatial learning are impaired by NMDAR antagonists (here AP-5). Importantly, dose response analysis of antagonist effects shows that AP-5 is effective in attenuating memory, but only at concentrations that also effectively block LTP.¹¹³

While numerous studies offer support for the dual effects of NMDAR antagonists in LTP and learning, it is appropriately noted in these studies that the simple fact that a drug blocks both LTP and learning does not denote causality; thus one cannot yet conclude that blocking LTP blocks learning. There always remains the possibility that treatments may exert effects on a third unknown variable crucial for learning, but separate from LTP, that also may be altered by NMDAR antagonists.

The disruption of other synaptic processes by the drug may also impair learning. The possibility that emerges is that theta rhythm, a 5- to 12-Hz frequency oscillation considered crucial for mnemonic hippocampal functions, may be a confound because a number of studies indicate that theta rhythm (specifically, type 2 theta, which is thought to reflect activity within the CA3-CA3 system)¹¹⁴ is attenuated by systemic administration of NMDAR antagonists.^115,116 Because blocking theta rhythm (for instance, with scopolamine, a muscarine antagonist) impairs learning, there remains the possibility that some of the effects of NMDAR antagonists on behavioral measures of learning may arise from alterations in theta rhythm rather than LTP.

Transgene knockouts have provided useful techniques for assessing the contributions of various proteins to LTP. Deletion of a number of genes including genes for NR1 NMDA receptors,¹¹⁷ the alpha isoforms of CamKII¹¹⁸ and CREB,⁴³ support the view that these proteins are necessary for the full expression of LTP.

While a number of reviews dealt with the various genetic manipulations and their effects on LTP and learning, several studies of note are based on methodologies using several levels of analyses. In one study, Silva and colleagues demonstrated an attenuation of both LTP and learning following a point mutation of the auto-phosphorylation site of CamKII.¹¹⁹ Crucially, the normal phosphorylating ability of CamKII appears unaltered. As previous accounts have suggested important roles for CamKII auto-phosphorylation in LTP, a sustained CamKII functioning could serve to maintain LTP by maintaining CamKII in an active state long after synaptic activation. In addition, it is proposed that the formation of CamKII assemblies and “hyperphosphorylation” of this assembly at their auto-phosphorylation sites may be a mechanism that serves in AMPA subunit trafficking to insert AMPA receptor subunits to the post-synaptic density.⁹⁶

While the deletion of the NR1 subunit would certainly be expected to impair LTP, a novel methodology employing these same knockouts is presented by Nakazawa¹²⁰ and remains an excellent example of combining genetic, electrophysiological, and behavioral methodologies to address LTP and its contribution to learning from an information–theoretic view.

The NR1 subunit was “conditionally deleted” in the CA3 region, a region with extensive recurrent connections that is thought to operate as an autoassociative device. Such devices are of interest in that they can perform pattern completion, that is, an entire stored pattern is recalled from only a subset of the original pattern. LTP among CA3 connections was impaired, as would be expected. However, the investigators went one step further and trained the animals on a spatial maze task that relied on extra-maze cues. Tests of recall involved removing some of the extra-maze cues. While animals with the normal complements of NMDA NR1 subunits showed only minor deficits, animals with NR1 knockouts in the CA3 region showed significant impairment in recalling the maze when these cues were removed. Together these data indicate that NR1 subunit deletions impaired LTP and learning and also that plasticity among the recurrent CA3 synapses prevented recall with a subset of patterns — evidence that firmly supports the view that the CA3 recurrent system performs its hypothesized autoassociative function.

The second postulate is that saturating LTP should both disrupt previously encoded memories and also impair acquisition of new memories. The first study to address this postulate¹²¹ reported that repeated stimulation of the perforant path saturated dentate LTP. Over the time LTP was saturated, deficits were observed in spatial memory tasks. Importantly, following the decay of LTP, these same animals were then able to acquire the task. However, initial attempts to replicate this study using the same methodology met with little success. This is likely due to procedural details.

While LTP was saturated by stimulation of perforant path fibers at one site in the study of Castro et al., stimulation with a single electrode in the angular bundle, a substantial tract containing virtually all of the perforant path fibers that project to the hippocampal formation, is unlikely to activate all perforant path inputs to the dentate.¹²² As the dentate and hippocampal formation are particularly resilient to damage,¹²³ a characteristic of distributed memory systems, the remaining unstimulated fibers and synapses within the angular bundle likely were sufficient to sustain spatial learning. Subsequent studies modified the methodology and used multiple stimulating electrodes that afforded a near-complete stimulation of the angular bundle. This, combined with lesions of the contralateral hippocampus, provided positive results confirming the prediction that saturation of LTP impairs the acquisition of a spatial task.¹²⁴

The third crucial experiment is that the selective erasure of LTP should disrupt the retention of established memories, but not impair acquisition of new memories. However, the technical knowledge of selectively erasing LTP is not yet available. Nevertheless, the inverse of this prediction is that enhancing LTP persistence should also enhance the persistence of memory. One study¹⁶ first showed that CPP, an NMDA antagonist, blocks the decay of perforant path–dentate LTP induction in behaving animals when administered 1 hour after LTP induction and daily thereafter for 6 days. Thus, sustained blockade of NMDARs prevented the decay of LTP over a 1-week period.

This is significant in that it indicates that LTP decay is an active process mediated by NMDA receptors. This finding tacitly suggests that, at least theoretically, not only is LTP permanent, but that the decay of LTP in the dentate is an active process, suggesting it is important for the normal operation of the dentate gyrus, possibly operating to prevent interference of newly acquired information with older accrued synaptic changes in the dentate. Because NMDAR antagonists block LTP decay and LTP is thought to manifest physical changes that constitute memory, it would be predicted that sustaining LTP also would sustain memory. This appears to be the case; animals trained to criteria on an eight-arm radial maze were then returned to their home cages for 1 week, during which they received either CPP daily at a dose effective in blocking LTP decay or the water vehicle. Following this 1-week period, the animals were then returned to the radial maze. During the first session, animals that received the vehicle only showed significantly more errors than animals that received CPP. Importantly, the performance of the CPP-treated animals in terms of maze errors was identical to scores observed on the last day of training to criterion 1 week earlier. Thus sustaining LTP also sustains memory, although it remains to be determined which synaptic populations affected by CPP were crucial for preserving the spatial memory, as systemic administration of CPP was used.

As more sophisticated techniques of analyzing hippocampal operation evolve, new opportunities to address the role of LTP in learning will arise as well. This is the case with hippocampal place cells. Simultaneous with the discovery of LTP was the discovery of hippocampal place cells, principal cells of the hippocampus (pyramidal cells, granule cells) that fire in complex bursts in particular regions of an environment.¹²⁵ While place cells were in the process of being characterized about the same time as LTP was first touted as a possible mnemonic device, only recently have there been any successful attempts to link place cells with LTP. This may be due in part to the absence of data indicating the dynamic aspects of place cells. Early in their discovery, it was believed that place fields were “hard wired” due to their rapid emergence. However, subsequent studies using high density recording revealed in the mid 1990s that place cells show dynamic changes with experience, suggesting they can be plastic in their firing features.¹²⁶ Subsequently, a number of studies attempting to link LTP with place cells emerged. Some studies attempted to determine whether treatments that impair LTP also impair the characteristics of place cells and their place fields. For instance, knockouts and transgenic point mutations of a number of proteins thought crucial for LTP (such as NMDA receptor NR1 subunits, CamKII, CREB, and GluR2) were consistent in their effects in that they did not prevent place field formation, although they appeared to disrupt selectivity of place cells, producing irregular and diffuse fields, even though these same treatments reduced LTP expression.¹²⁷

If place fields depend upon information embedded within the synapses of hippocampal networks due to LTP, then it would be expected that a disruption of established plasticity using LTP would similarly disrupt established place fields. In studies in the spirit of the LTP saturation experiments, Buzsaki and colleagues¹²⁸ induced LTP following the establishment of place fields in an environment. LTP was induced by stimulation of the ventral hippocampal commissure (which contains primarily commissural fibers from CA3 that project to both CA3 and CA1). Subsequent analysis of place fields revealed that after tetanization, re-exposure to the environment led to remapping. Importantly, LTP was measured at various electrode sites and revealed various degrees of LTP at different sites following stimulation. It was observed that place fields were more likely to be altered in regions showing larger degrees of LTP. This experiment is a variation of one of the hypotheses of Morris and McNaughton⁵⁵ indicating that saturating LTP should disrupt previously encoded memories. These results offer compelling evidence that LTP is involved with the dynamic behaviors of place cells.

A curious finding in this study was that once LTP decayed, the alterations in place cell responses disappeared as well. This is in stark contrast to the endogenously formed place fields that, once established, remain stable indefinitely, even with subsequent learning. The reasons for this discrepancy are unknown, but several possible explanations exist. First, LTP in this study was induced using standard bursts of 200 Hz. While that level is sufficient for LTP induction, a number of studies noted that induction by theta bursts produced a more sustained LTP,⁸ possibly by inducing multiple forms of LTP.^59,60

Second, certain behavioral states and neuromodulators are crucial for the formation of both stable LTP and stable place fields. For instance, the presence of essential neuromodulators (such as dopamine, norepinephrine, and serotonin) associated with attention or arousal may be crucial for the normal stabilization of place field firing. Such factors may not only serve a permissive role in plasticity; their absence may preserve stability such that stable synaptic strengths (and place fields) predominate unless subsequent synaptic plasticity occurs in the presence of specific neuromodulatory signals. Such an effect would suggest plasticity and stability states regulated by neuromodulators and behavioral states and may reflect a partial solution to the stability–plasticity dilemma. In support of this view, we have reported that when LTP is induced while animals explore novel cages, both LTP magnitude and longevity are increased dramatically.⁶² Thus, had LTP been induced in an aroused condition or in a novel environment, sustained LTP and sustained place fields might have been observed.

Of particular relevance to learning are the studies investigating the phenomenon of place cell remapping. Normally, place cells fire within circumscribed regions; cell activity remains stable and the cells retain their place specificity with repeated exposures to the same environment. This selective activity will persist, even with removal of a subset of cues. However, if the environment is changed beyond a threshold or an animal is placed in an unfamiliar environment, the most common result is a complete remapping of cells within that environment, frequently in a pattern that bears no resemblance to the original distribution of place fields.

Obviously, the remapping effect suggests rapid plastic interactions among principal cells that map features of the altered or new environment and as such would be expected to involve LTP. However, the results here are somewhat confounding. In a study by Kentros et al.,⁷ the effects of CPP, an NMDA antagonist, on the remapping of place fields was addressed. Normally, remapping occurs when animals are then placed in a novel environment. These remapped place fields in both environments are stable among the distinct environments for days, and do not interfere with each other.

Would blocking LTP also block remapping? This was addressed by administering CPP and exposing an animal to a novel environment. As noted above, CPP had no effect on established place fields. However, the surprising result is that CPP did not alter initial remapping; in fact, this new remapped state persisted for 90 min after exposure. Clearly memory occurred in the presence of an NMDAR antagonist. However CPP did eliminate this stability of the newly formed map 24 hours later, such that reintroduction to the same novel environment 1 day later resulted in another distinct remapping of the environment. Thus NMDARs appear necessary for the long-term stability of place fields. A similar effect is seen with protein synthesis inhibitors that specifically block establishment of the late phase of LTP.^47–49 Here, as in the studies with NMDAR antagonists, anisomycin, a protein synthesis inhibitor, similarly blocked the long-term stability of place fields.

These studies suggest NMDARs and protein synthesis are necessary for the long-term stability of remapped place fields, a finding that fits with the view that L-LTP, which is mediated by NMDARs and protein synthesis, is involved in sustaining newly generated place fields. However, the most interesting aspect of this study is that NMDARs, and, presumably, LTP, apparently are not involved in the rapid experience-dependent formation of such fields. This is a peculiar result, given that remapping involves conditions in which LTP would be an ideal mechanism (specifically, rapid, associative and activity-dependent induction that usually precedes the later protein synthesis-dependent L-LTP). While these studies suggest the importance of LTP, particularly late or L-LTP, in long-term stability of place fields, the more important finding is that LTP, or more accurately, NMDARs, appear not to be involved in the rapid formation of place cell representations.

What plastic process, then, may underlie initial rapid remapping? If CPP blocks long-term stability (and presumably, L-LTP), how could L-LTP be involved in this later stability in the absence of the initial E-LTP? One possible explanation is that E-LTP is indeed involved in this process, although NMDAR blockade may not alter its induction. As an example, mossy fiber CA3 LTP and its presumed role in encoding offer a plausible framework to explain these results. A common theme in models of hippocampal function is that the CA3 region serves an autoassociative function by virtue of its extensive recurrent (CA3-CA3) excitatory connections. The view first formalized by Marr⁵⁴ that autoassociative processes are performed by the CA3 region remains a feature of most models of hippocampal function.^55,129

These recurrent connections constitute the majority of excitatory inputs to CA3 pyramidal cells and can display Hebbian LTP.⁵² Representations within the CA3 autoassociative system is thought to involve the formation of attractor states (or ensembles of active recurrently connected CA3 cells) as a result of plastic changes among CA3-CA3 synapses. Importantly, the formation of CA3 attractors is thought to occur via LTP of CA3-CA3 synapses by the coincident activity of the dentate mossy fiber projection, whose synapses are thought to act as powerful “detonator synapses.”^54,55,130 The concurrent activation of CA3 pyramidal cells by detonator synapses is thought to mediate associative LTP induction of only those CA3-CA3 synapses that immediately preceded mossy fiber activity. Thus mossy fibers serve to establish a sparse ensemble of co-active CA3-CA3 cells thought to reflect a hippocampal “representation” within the CA3 region.

In the above context, how can the lack of an effect of NMDAR antagonists on remapping be explained? First, it should be noted that mossy fiber LTP was one of the first to be deemed a distinct form of LTP due to its independence from NMDARs and its unusual time course in vivo.^45,131 Second, our studies demonstrate an unusual property of associative LTP. If LTP is induced in a synaptic population in an associative manner, in this case by pairing weak synaptic activation of one pathway with a more intense activation of a separate set of converging afferents, LTP is observed in both pathways. However, the induction of LTP in the weak or “associated” pathway appears to be sensitive only to manipulations that block LTP induction in the other more intense, associating pathway used to induce LTP.

For example, NMDA receptor antagonists such as CPP normally block LTP induction in medial perforant path projections to the CA3 region. By contrast, projections arising from the lateral perforant path are unaffected by these antagonists.^132,133 However, if LTP is induced in lateral perforant path-CA3 synapses by the strong coactivation of the NMDAR-dependent medial perforant path, NMDAR antagonists effectively block LTP in both pathways. Conversely, if LTP is induced in this NMDAR-dependent medial pathway by strong activation of the NMDAR-independent lateral perforant pathway, CPP has no effect on associative LTP at medial perforant path synapses and LTP is observed in both lateral and medial perforant path-CA3 synaptic populations.

Extrapolation of this feature of associative LTP with current models assigning a detonator (or associating) function to mossy fiber synapses suggests the possibility that rapid encoding does indeed involve NMDAR-dependent LTP. However, during encoding, as NMDAR-dependent LTP among CA3-CA3 recurrents is thought to be induced in an associative manner via activation of the NMDAR-independent mossy fiber detonator synapses, it is possible that rapid remapping is unaffected by the NMDAR antagonist simply because associative LTP may have been induced by the coactivation of the mossy fiber detonators. Because the mossy fiber synapses display LTP that is insensitive to NMDAR antagonists, NMDAR antagonists may have had no effect on CA3-CA3 LTP or initial remapping due to the associative induction of NMDAR-dependent LTP by the stronger NMDAR-independent mossy fiber pathway.

Taken together, evidence for the role of LTP in establishing place fields is supportive; however, it is likely that a more definitive role for LTP in establishing place cells fields will emerge as details of place cell dynamics are revealed. Other than remapping, plastic processes associated with place cells and their fields are apparently quite subtle. Studies addressing such subtle changes in place cell dynamics have emerged; in particular, place cells can display dynamic changes with experience. In particular, if a rat runs along a linear trajectory that sequentially fires cells A, B, and C, repeated experiences with this trajectory result in an asymmetric “expansion” of the place field in the opposite direction of the animal’s trajectory. Thus, over time, the activity of cell B slowly shifts in the direction of A, such that place cell B fires closer to the A place field earlier in the trajectory.⁷ Because this phenomenon experience-dependent, it is not unreasonable to assume associative plastic processes may underlie its development. In addition, the closer association of sequences of cells would be expected with associative experience-dependent plastic changes, making LTP a likely candidate.

The phenomenon of asymmetric place field expansion also fits with the features of LTP seen experimentally. One aspect of LTP seen in addition to an increase in the amplitude of synaptic responses is a reduction in the latency of the population spike. Cells fire earlier at potentiated synapses after LTP induction, presumably due to the increases in synaptic strength. Thus asymmetric place field expansion and the earlier firing of recently activated cells within a sequence may be consequences of the reduced latency to cell discharge that normally follows LTP induction.

The property of asymmetric expansion fits well with the view that an associative process underlies this phenomenon. Subsequent studies revealed that in contrast to remapping, systemic administration of NMDAR antagonists blocks the asymmetric expansion of place fields.¹³⁴ Again, as noted for LTP, caution must be used when interpreting the effects of NMDAR antagonists and possible confounds because they appear to alter theta rhythm.^115,116 As place cell firing is tightly coupled to theta rhythm even within a place field, it is entirely possible that the temporal aspects place cell activity and their plastic coupling may depend on the integrity of theta during learning.

This progress is illustrative of an important heuristic approach to linking LTP with memory; many of the negative findings observed in studies linking LTP with behavioral learning arise primarily as a result of our rudimentary understanding of subtleties in the processes of LTP (such as the finding that the induction of associative LTP is determined exclusively by the receptor mechanisms of the stronger coactive pathway). Likewise, studies attempting to link LTP and place field formation are hindered by our lack of understanding of both the mechanics of LTP and the role of place cells in memory and their dynamic changes with experience. Future studies that integrate newly discovered features of place cell dynamics (Guzowski et al.¹³⁵) with behavioral manipulations clearly serve as avenues that will contribute to our knowledge of LTP and our understanding of hippocampal function in general.

OUTSTANDING QUESTIONS AND NEW DIRECTIONS

By elucidating the underlying mechanisms of LTP and LTD, it seems possible to reconstruct some of the subcellular events triggered by experience and deprivation to alter neuronal function. It is also likely that more information about LTP and hippocampal function in general, will follow the studies of place cell dynamics and their plastic properties, particularly during remapping. The coming years will bring further understanding of the factors regulating the induction, maintenance, and distribution of long-lasting synaptic modifications and their contributions to normal adaptive behavior.

ACKNOWLEDGMENTS

The authors thank D. Castillo and A. Gómez-Palacio Shjetnan for their help during the preparation of the figures. Part of the work was supported by PAPIIT IN213503 (M.L.E.), NIH/NINDS and NIH/NIGMS (B.E.D.).

REFERENCES

1.

Bliss TVP, Lomo T. Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J. Physiol. 1973;232:331. [PMC free article: PMC1350458] [PubMed: 4727084]

2.

Bliss T, Gardner-Medwin A. Long-lasting potentiation of synaptic transmission in the dentate area of the unanesthetized rabbit following stimulation of the perforant path. J. Physiol. 1973;232:357. [PMC free article: PMC1350459] [PubMed: 4727085]

3.

Xie X, Barrionuevo G, Berger TW. Differential expression of short-term potentiation by AMPA and NMDA receptors in dentate gyrus. Learn. Mem. 1996;3:115. [PubMed: 10456082]

4.

Hanse E, Gustafsson B. Onset and stabilization of NMDA receptor-dependent hippocampal long-term potentiation. Neurosci. Res. 1994;20:15. [PubMed: 7984337]

5.

Martinez CO, Do VH, Derrick BE. Associative LTP among afferents to the CA3 region in vivo. Brain Res. 2002;94:86. [PubMed: 12020879]

6.

Mehta MR, Barnes CA, McNaughton BL. Experience-dependent asymmetric expansion of hippocampal place fields. Proc. Natl. Acad. Sci USA. 1997;94:8918. [PMC free article: PMC23195] [PubMed: 9238078]

7.

Kentros C, Hargreaves E, Hawkins RD, Kandel ER, Shapiro M, Muller RV. Abolition of long-term stability of new hippocampal place cell maps by NMDA receptor blockade. Science. 1998;280:2121. [PubMed: 9641919]

8.

Staubli U, Lynch G. Stable hippocampal long-term potentiation elicited by “theta” pattern stimulation. Brain Res. 1987;435:227. [PubMed: 3427453]

9.

McNaughton BL, Douglas RM, Goddard DV. Synaptic enhancement in fascia dentata: cooperativity among coactive afferents. Brain Res. 1978;157:277. [PubMed: 719524]

10.

Gustafsson B, Wigstrom H, Abraham WC, Huang YY. Long-term potentiation in the hippocampus using depolarizing current pulses as the conditioning stimulus to single volley synaptic potentials. J. Neurosci. 1987;7:774. [PMC free article: PMC6569059] [PubMed: 2881989]

11.

Barnes CA. Memory deficits associated with senescence: a neurophysiological and behavioral study in the rat. J. Comp. Physiol. Psychol. 1979;93:74. [PubMed: 221551]

12.

Squire LR. Memory and the Brain. Oxford University Press; New York: 1987.

13.

Wilson MA, McNaughton BL. Reactivation of hippocampal ensemble memories during sleep. Science. 1994;265:676. [PubMed: 8036517]

14.

Cartwright RD. The role of sleep in changing our minds: a psychologist’s discussion of papers on memory reactivation and consolidation in sleep. Learn. Mem. 2004;11:660. [PMC free article: PMC534693] [PubMed: 15576882]

15.

Bontempi B, Laurent-Demir C, Destrade C, Jaffard R. Time-dependent reorganization of brain circuitry underlying long-term memory storage. Nature. 1999;400:671. [PubMed: 10458162]

16.

Villarreal D, Do V, Haddad E, Derrick BE. NMDA antagonists sustain LTP and spatial memory: evidence for active processes underlying LTP decay. Nat. Neurosci. 2002;5:48. [PubMed: 11740500]

17.

Hebb DO. The Organization of Behavior. John Wiley & Sons; New York: 1949.

18.

Barrionuevo G, Schottler F, Lynch G. The effects of repetitive low frequency stimulation on control and “potentiated” synaptic responses in the hippocampus. Life Sci. 1980;27:2385. [PubMed: 7207026]

19.

Bear MF. Homosynaptic long-term depression: a mechanism for memory? Proc. Natl. Acad. Sci. USA. 1999;96:9457. [PMC free article: PMC33710] [PubMed: 10449713]

20.

Mulkey RM, Herron CE, Malenka RC. An essential role for protein phosphatases in hippocampal long-term depression. Science. 1993;261:1051. [PubMed: 8394601]

21.

Christie BR, Abraham WC. Priming of associative long-term depression in the dentate gyrus by theta frequency synaptic activity. Neuron. 1992;9:79. [PubMed: 1321647]

22.

Debanne D, Thompson SM. Associative long-term depression in the hippocampus in vitro. Hippocampus. 1996;6:9. [PubMed: 8878736]

23.

Wickens JR, Abraham WC. Involvement of L-type calcium channels in heterosynaptic long-term depression in the hippocampus. Neurosci. Lett. 1991;130:128. [PubMed: 1721110]

24.

Normann C, Peckys D, Schulze CH, Walden J, Jonas P, Bischofberger J. Associative long-term depression in the hippocampus is dependent on postsynaptic N-type Ca²⁺ channels. J. Neurosci. 2000;15:8290. [PMC free article: PMC6773198] [PubMed: 11069935]

25.

Massey PV, et al. Differential roles of NR2A and NR2B-containing NMDA receptors in cortical long-term potentiation and long-term depression. J. Neurosci. 2004;24:7821. [PMC free article: PMC6729941] [PubMed: 15356193]

26.

Moult PR, et al. Tyrosine phosphatases regulate AMPA receptor trafficking during metabotropic glutamate receptor-mediated long-term depression. J. Neurosci. 2006;26:2544. [PMC free article: PMC6793648] [PubMed: 16510732]

27.

Wang J, Yeckel MF, Johnston D, Zucker RS. Photolysis of postsynaptic caged Ca²⁺ can potentiate and depress mossy fiber synaptic responses in rat hippocampal CA3 pyramidal neurons. J. Neurophysiol. 2004;91:1596. [PMC free article: PMC2867667] [PubMed: 14645386]

28.

Lei S, et al. Depolarization-induced long-term depression at hippocampal mossy fiber-CA3 pyramidal neuron synapses. J. Neurosci. 2003;23:9786. [PMC free article: PMC6740888] [PubMed: 14586006]

29.

Domenici MR, Berretta N, Cherubini E. Two distinct forms of long-term depression coexist at the mossy fiber-CA3 synapse in the hippocampus during development. Proc. Natl. Acad. Sci. USA. 1998;95:8310. [PMC free article: PMC20972] [PubMed: 9653183]

30.

Derrick BE, Martinez JL Jr. Associative bidirectional modifications at the hippocampal mossy fibre CA3 synapse. Nature. 1996;381:429. [PubMed: 8632800]

31.

Kobayashi K, Manabe T, Takahashi T. Presynaptic long-term depression at the hippocampal mossy fiber CA3 synapse. Science. 1996;273:648. [PubMed: 8662556]

32.

Brown GD, Dalloz P, Hulme C. Mathematical and connectionist models of human memory: a comparison. Memory. 1995;3:113. [PubMed: 7796301]

33.

Skaggs WE, McNaughton BL. Computational approaches to hippocampal function. Curr. Opin. Neurobiol. 1992;2:209. [PubMed: 1638156]

34.

Morris RGM. Computational neuroscience: modeling the brain. In: Morris RGM, editor. Parallel Distributed Processing Implications for Psychology and Neurobiology. Oxford University Press; New York: 1989. p. 203.

35.

Gustafsson B, Wigstrom H. Long-term potentiation in the hippocampal CA1 region: its induction and early temporal development. Progr. Brain Res. 1990;83:223. [PubMed: 2203099]

36.

Malinow R, Miller JP. Postsynaptic hyperpolarization during conditioning reversibly blocks long-term potentiation. Nature. 1986;320:529. [PubMed: 3008000]

37.

Lynch G, Larson J, Kelso S, Barrionuevo G, Schottler F. Intracellular injections of EGTA block induction of hippocampal long-term potentiation. Nature. 1983;305:719. [PubMed: 6415483]

38.

Johnston D, Williams S, Jaffe D, Gray R. NMDA receptor-independent long-term potentiation. Annu. Rev. Physiol. 1992;54:489. [PubMed: 1314043]

39.

Yeckel MF, Kapur A, Johnston D. Multiple forms of LTP in hippocampal CA3 neurons use a common postsynaptic mechanism. Nat. Neurosci. 1999;2:625. [PMC free article: PMC2951317] [PubMed: 10404192]

40.

Sanes JR, Lichtman JW. Can molecules explain long-term potentiation? Nat. Neurosci. 1999;2:597. [PubMed: 10404178]

41.

Derkach V, Barria A, Soderling TR. Ca²⁺/calmodulin kinase II enhances channel conductance of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionate type glutamate receptors. Proc. Natl. Acad. Sci USA. 1999;96:3269. [PMC free article: PMC15931] [PubMed: 10077673]

42.

Oh MC, et al. Extrasynaptic membrane trafficking regulated by GluR1 serine 845 phosphorylation primes AMPA receptors for long-term potentiation. J. Biol. Chem. 2006;281:752. [PubMed: 16272153]

43.

Bourtchuladze R, et al. Deficient long-term memory in mice with a targeted mutation of the cAMP-responsive element-binding protein. Cell. 1994;79:59. [PubMed: 7923378]

44.

Frey U, Morris RG. Synaptic tagging: implications for late maintenance of hippocampal long-term potentiation. Trends Neurosci. 1998;385:533. [PubMed: 9610879]

45.

Harris EW, Cotman CW. Long-term potentiation of guinea pig mossy fiber responses is not blocked by N-methyl D-aspartate antagonists. Neurosci. Lett. 1986;70:132. [PubMed: 3022192]

46.

French PJ, et al. Subfield-specific immediate early gene expression associated with hippocampal long-term potentiation in vivo. Eur. J. Neurosci. 2001;13:968. [PubMed: 11264669]

47.

Stanton PK, Sarvey JM. Rapid spine delivery and redistribution of AMPA receptors after synaptic NMDA receptor activation: blockade of long-term potentiation in rat hippocampal CA1 region by inhibitors of protein synthesis. J. Neurosci. 1984;4:3080. [PMC free article: PMC6564864] [PubMed: 6502226]

48.

Krug M, Lossner B, Ott T. Anisomycin blocks the late phase of long-term potentiation in the dentate gyrus of freely moving rats. Brain Res. Bull. 1984;13:39. [PubMed: 6089972]

49.

Otani S, Marshall CJ, Tate WP, Goddard GV, Abraham WC. Maintenance of long-term potentiation in rat dentate gyrus requires protein synthesis but not messenger RNA synthesis immediately post-tetanization. Neuroscience. 1989;28:519. [PubMed: 2710327]

50.

Jaffe D, Johnston D. Induction of long-term potentiation at hippocampal mossy fibers follow a Hebbian rule. J. Neurophys. 1990;64:948. [PubMed: 2230936]

51.

Williams S, Johnston D. Long-term potentiation of hippocampal mossy fiber synapses is blocked by postsynaptic injection of calcium chelators. Neuron. 1989;3:583. [PubMed: 2577129]

52.

Derrick BE, Martinez JL Jr. Opioid receptor activation is one factor underlying the frequency dependence of mossy fiber LTP induction. J. Neurosci. 1994;14:4359. [PMC free article: PMC6577017] [PubMed: 7913121]

53.

Jung MW, McNaughton BL. Spatial selectivity of unit activity in the hippocampal granular layer. Hippocampus. 1993;3:165. [PubMed: 8353604]

54.

Marr D. Simple memory: a theory for archicorticortex. Proc. R. Soc. London B. 1971;262:23. [PubMed: 4399412]

55.

Morris RGM, McNaughton BL. Memory storage in a distributed model of hippocampal formation. Trends Neurosci. 1987;10:408.

56.

Abraham W, Bear M. Metaplasticity: plasticity of synaptic plasticity. Trends Neurosci. 1996;19:126. [PubMed: 8658594]

57.

Escobar M, Barea-Rodriguez EJ, Derrick BE, Martinez JL Jr. Mossy fiber synaptogenesis is induced by high frequency stimulation of mossy fibers. Brain Res. 1997;751:330. [PubMed: 9099823]

58.

Derrick BE, York A, Martinez JL Jr. Increases in granule cell neurogenesis following stimulation that induces mossy fiber LTP. Brain Res. 2000;857:800. [PubMed: 10700582]

59.

Morgan SL, Teyler TJ. Electrical stimuli patterned after the theta-rhythm induce multiple forms of LTP. J. Neurophysiol. 2001;86:1289. [PubMed: 11535677]

60.

Raymond CR, Redman SJ. Different calcium sources are narrowly tuned to the induction of different forms of LTP. J. Neurophysiol. 2002;88:249. [PubMed: 12091550]

61.

Hoffman DA, Sprengel R, Sakmann B. Molecular dissection of hippocampal theta-burst pairing potentiation. Proc. Natl. Acad. Sci. USA. 2002;99:7740. [PMC free article: PMC124338] [PubMed: 12032353]

62.

Davis CD, Jones FL, Derrick BE. Novel environments enhance the induction and maintenance of long-term potentiation in the dentate gyrus. J. Neurosci. 2004;24:6497. [PMC free article: PMC6729872] [PubMed: 15269260]

63.

Li S, Cullen WK, Anwyl R, Rowan MJ. Dopamine-dependent facilitation of LTP induction in hippocampal CA1 by exposure to spatial novelty. Nat. Neurosci. 2003;6:526. [PubMed: 12704392]

64.

Straube T, Korz V, Balschun D, Frey JU. Requirement of beta-adrenergic receptor activation and protein synthesis for LTP-reinforcement by novelty in rat dentate gyrus. J. Physiol. 2003;552:953. [PMC free article: PMC2343450] [PubMed: 12937286]

65.

Turrigiano GG, Leslie KR, Desai NS, Ruterford LC, Nelson SB. Activity dependent scaling of quantal amplitude in neocortical neurons. Nature. 1998;391:892. [PubMed: 9495341]

66.

Bienenstock EL, Cooper LN, Munro PW. Theory for the development of neuron selectivity: orientation specificity and binocular interaction in visual cortex. J. Neurosci. 1982;2:32. [PMC free article: PMC6564292] [PubMed: 7054394]

67.

Quinlan EM, Olstein DH, Bear MF. Bidirectional, experience-dependent regulation of N-methyl-D-aspartate receptor subunit composition in the rat visual cortex during postnatal development. Proc. Natl. Acad. Sci. USA. 1999;96:876. [PMC free article: PMC23143] [PubMed: 10536016]

68.

Philpot BD, Sekhar AK, Shouval HZ, Bear MF. Visual experience and deprivation bidirectionally modify the composition function of NMDA receptors in visual cortex. Neuron. 2001;29:157. [PubMed: 11182088]

69.

Rogan MT, Staubli UV, LeDoux JE. Fear conditioning induces associative long-term potentiation in the amygdala. Nature. 1997;390:604. [PubMed: 9403688]

70.

McKernan MG, Shinnick-Gallagher P. Fear conditioning induces a lasting potentiation of synaptic currents in vitro. Nature. 1997;390:607. [PubMed: 9403689]

71.

Rioult-Pedotti MS, Friedman D, Hess G, Donoghue JP. Strengthening of horizontal cortical connections following skill learning. Nat. Neurosci. 1998;1:230. [PubMed: 10195148]

72.

Escobar ML, Bermúdez-Rattoni F. Long-term potentiation in the insular cortex enhances conditioned taste aversion retention. Brain Res. 2000;852:208. [PubMed: 10661514]

73.

Malenka RC, Bear M. LTP and LTD: an embarrassment of riches. Neuron. 2004;44:5. [PubMed: 15450156]

74.

Rittenhouse CD, Shouval HZ, Paradiso MA, Bear MF. Monocular deprivation induces homosynaptic long-term depression in visual cortex. Nature. 1999;397:347. [PubMed: 9950426]

75.

Allen CB, Celikel T, Feldman DE. Long-term depression induced by sensory deprivation during cortical map plasticity in vivo. Nat. Neurosci. 2003;6:291. [PubMed: 12577061]

76.

Celikel T, Szostak VA, Feldman DE. Modulation of spike timing by sensory deprivation during induction of cortical map plasticity. Nat. Neurosci. 2004;7:534. [PMC free article: PMC3082358] [PubMed: 15064767]

77.

Heynen AJ, et al. Molecular mechanisms for loss of visual cortical responsiveness following brief monocular deprivation. Nat. Neurosci. 2003;6:854. [PubMed: 12886226]

78.

Sawtell NB, Huber KM, Roder JC, Bear MF. Induction of NMDA receptor-dependent long-term depression in visual cortex does not require metabotropic glutamate receptors. J. Neurophysiol. 1999;82:3594. [PubMed: 10601487]

79.

Hyman SE, Malenka RC. Addiction and the brain: the neurobiology of compulsion and its persistence. Nat. Rev. Neurosci. 2001;2:695. [PubMed: 11584307]

80.

Kelley AE. Memory and addiction: Shared neural circuitry and molecular mechanisms. Neuron. 2004;44:161. [PubMed: 15450168]

81.

Ungless MA, Whistier JL, Malenka RC, Bonci A. Single cocaine exposure in vivo induces long-term potentiation in dopamine neurons. Nature. 2001;411:583. [PubMed: 11385572]

82.

Faleiro LJ, Jones S, Kauer JA. Rapad synaptic plasticity of glutamatergic synapses on dopamine neurons in the ventral tegmental area in response to acute amphetamine injection. Neuropsychopharmacology. 2004;29:2115. [PubMed: 15150533]

83.

Liao D, Hessler NA, Malinow R. Activation of post-synaptically silent synapses during pairing-induced LTP in CA1 region of hippocampal slice. Nature. 1995;375:400. [PubMed: 7760933]

84.

Malinow R, Malenka RC. AMPA receptor trafficking and synaptic plasticity. Annu. Rev. Neurosci. 2002;25:103. [PubMed: 12052905]

85.

Collingridge GL, Isaac JTR, Wang YT. Receptor trafficking and synaptic plasticity. Nat. Rev. Neurosci. 2004;5:952. [PubMed: 15550950]

86.

Emptage NJ, Reid CA, Fine A, Bliss TV. Optical quantal analysis reveals a presynaptic component of LTP at hippocampal Schaffer-associational synapses. Neuron. 2003;38:797. [PubMed: 12797963]

87.

Choi S, Klingauf J, Tsien RW. Fusion pore modulation as a presynaptic mechanism contributing to expression of long-term potentiation. Philos. Trans. R. Soc. London B. 2003;358:695. [PMC free article: PMC1693158] [PubMed: 12740115]

88.

Zakharenko SS, et al. Presynaptic BDNF required for a presynaptic but not post-synaptic component of LTP at hippocampal CA1-CA3 synapses. Neuron. 2003;39:975. [PubMed: 12971897]

89.

Lauri SE, et al. Functional maturation of CA1 synapses involves activity-dependent loss of tonic kainate receptor-mediated inhibition of glutamate release. Neuron. 2006;50:415. [PubMed: 16675396]

90.

Schulz PE, Cook EP, Johnston D. Changes in paired-pulse facilitation suggest presynaptic involvement in long-term potentiation. J. Neurosci. 1994;14:5325. [PMC free article: PMC6577083] [PubMed: 7916043]

91.

Bayer KU, De Koninck P, Leonard AS, Hell JW, Schulman H. Interaction with the NMDA receptor locks CaMKII in an active conformation. Nature. 2001;411:801. [PubMed: 11459059]

92.

Leonard AS, et al. Regulation of calcium/calmodulin-dependent protein kinase II docking to N-methyl-D-aspartate receptors by calcium/calmodulin and alpha-actinin. J. Biol. Chem. 2002;277:48441. [PubMed: 12379661]

93.

Si SH, et al. Phosphorylation of the AMPA receptor GluR1 subunit is required for synaptic plasticity and retention of spatial memory. Science. 1999;284:1755. [PubMed: 12628184]

94.

Luthi A, et al. Hippocampal LTD expression involves a pool of AMPARs regulated by the NSF-GluR2 interaction. Neuron. 1999;24:389. [PubMed: 10571232]

95.

Lynch G, Baudry M. The biochemistry of memory: a new and specific hypothesis. Science. 1984;224:1057. [PubMed: 6144182]

96.

Lisman JE, McIntyre CC. Synaptic plasticity: a molecular memory switch. Curr. Biol. 2001;11:788. [PubMed: 11591339]

97.

Boehm J, Malinow R. AMPA receptor phosphorylation during synaptic plasticity. Biochem. Soc. Trans. 2005;33:1354. [PubMed: 16246117]

98.

Perez-Otaño I, Ehlers MD. Homeostatic plasticity and NMDA receptor trafficking. Trends Neurosci. 2002;28:229. [PubMed: 15866197]

99.

Lohof AM, Ip NY, Poo MM. Potentiation of developing neuromuscular synapses by the neurotrophins NT-3 and BDNF. Nature. 1993;363:350. [PubMed: 8497318]

100.

Kang H, Schuman EM. Long-lasting neurotrophin-induced enhancement of synaptic transmission in the adult hippocampus. Science. 1995;267:1658. [PubMed: 7886457]

101.

Ying SW, et al. Brain-derived neurotrophic factor induces long-term potentiation in intact adult hippocampus: requirement for ERK activation coupled to CREB and upregulation of Arc synthesis. J. Neurosci. 2002;22:1532. [PMC free article: PMC6758896] [PubMed: 11880483]

102.

Bramham CR, Messaoudi E. BDNF function in adult synaptic plasticity: the synaptic consolidation hypothesis. Progr. Neurobiol. 2005;76:99. [PubMed: 16099088]

103.

Pang PT, et al. Cleavage of proBDNF by tPA/plasmin is essential for long-term hippocampal plasticity. Science. 2004;306:487. [PubMed: 15486301]

104.

McAllister AK, Katz LC, Lo DC. Neurotrophins and synaptic plasticity. Annu. Rev. Neurosci. 1999;22:295. [PubMed: 10202541]

105.

Jiang B, et al. Brain-derived neurotrophic factor induces long-lasting potentiation of synaptic transmission in visual cortex in vivo in young rats, but not in the adult. Eur. J. Neurosci. 2001;14:1219. [PubMed: 11703451]

106.

Tokuyama W, Okuno H, Hashimoto T, Li YX, Miyashita Y. BDNF up-regulation during declarative memory formation in monkey inferior temporal cortex. Nat. Neurosci. 2000;3:1134. [PubMed: 11036271]

107.

Escobar ML, Figueroa-Guzmán Y, Gómez-Palacio Schjetnan A. In vivo insular cortex LTP induced by brain-derived neurotrophic factor. Brain Res. 2003;991:274. [PubMed: 14575905]

108.

Castillo V, Figueroa-Guzmán Y, Escobar ML. Brain-derived neurotrophic factor enhances conditioned taste aversion retention. Brain Res. 2006;1067:250. [PubMed: 16364259]

109.

Yamada K, Mizuno M, Nabeshima T. Role for brain-derived neurotrophic factor in learning and memory. Life Sci. 2002;70:735. [PubMed: 11833737]

110.

Minichiello L, et al. Mechanism of Trk-mediated hippocampal long-term potentiation. Neuron. 2002;36:121. [PubMed: 12367511]

111.

Tyler WJ, Alonso M, Bramham CR, Pozzo-Miller LD. From acquisition to consolidation: on the role of brain-derived neurotrophic factor signaling in hippocampal-dependent learning. Learn. Mem. 2002;9:224. [PMC free article: PMC2806479] [PubMed: 12359832]

112.

Morris RG, Anderson E, Lynch GS, Baudry M. Selective impairment of learning and blockade of long-term potentiation by an N-methyl-D-aspartate receptor antagonist AP5. Nature. 1986;319:774. [PubMed: 2869411]

113.

Davis S, Butcher SP, Morris RG. The NMDA receptor antagonist D-2-amino-5-phosphonopentanoate (D-AP5) impairs spatial learning and LTP in vivo at intracerebral concentrations comparable to those that block LTP in vitro. J. Neurosci. 1992;12:21. [PMC free article: PMC6575679] [PubMed: 1345945]

114.

Buzsaki G. Theta oscillations in the hippocampus. Neuron. 2002;33:325. [PubMed: 11832222]

115.

Pitkanen M, Sirvio J, Ylinen A, Koivisto E, Riekkinen P Sr. Effects of NMDA receptor modulation on hippocampal type 2 theta activity in rats. Gen. Pharmacol. 1995;26:1065. [PubMed: 7557252]

116.

Leung LW, Desborough KA. APV, an N-methyl-D-aspartate receptor antagonist, blocks the hippocampal theta rhythm in behaving rats. Brain Res. 1988;463:148. [PubMed: 2904294]

117.

Tsien JZ, Huerta PT, Tonegawa S. The essential role of hippocampal CA1 NMDA receptor-dependent synaptic plasticity in spatial memory. Cell. 1996;87:1327. [PubMed: 8980238]

118.

Hinds HL, Tonegawa S, Malinow R. CA1 long-term potentiation is diminished but present in hippocampal slices from alpha-CaMKII mutant mice. Learn. Mem. 1998;5:344. [PMC free article: PMC311262] [PubMed: 10454359]

119.

Giese KP, Fedorov NB, Filipkowski RK, Silva AJ. Auto-phosphorylation at Thr286 of the alpha calcium-calmodulin kinase II in LTP and learning. Science. 1998;279:870. [PubMed: 9452388]

120.

Nakazawa K, et al. Requirement for hippocampal CA3 NMDA receptors in associative memory recall. Science. 2002;297:211. [PMC free article: PMC2877140] [PubMed: 12040087]

121.

Castro CA, Silbert LH, McNaughton BL, Barnes CA. Recovery of spatial learning deficits after decay of electrically induced synaptic enhancement in the hippocampus. Nature. 1989;342:545. [PubMed: 2586626]

122.

Barnes CA, et al. LTP saturation and spatial learning disruption: effects of task variables and saturation levels. J. Neurosci. 1994;14:5793. [PMC free article: PMC6576980] [PubMed: 7931545]

123.

McNaughton BL, Barnes CA, Meltzer J, Sutherland RJ. Hippocampal granule cells are necessary for normal spatial learning but not for spatially selective pyramidal cell discharge. Exp. Brain Res. 1989;76:485. [PubMed: 2792242]

124.

Moser EI, Krobert KA, Moser MB, Morris RG. Impaired spatial learning after saturation of long-term potentiation. Science. 1998;281:2038. [PubMed: 9748165]

125.

O’Keefe J, Nadel L. The Hippocampus as a Cognitive Map. Oxford University Press; New York: 1978.

126.

Wilson MA, McNaughton BL. Dynamics of the hippocampal ensemble code for space. Science. 1993;261:1055. [PubMed: 8351520]

127.

Jeffery KJ, et al. A proposed architecture for the neural representation of spatial context. Neurosci. Biobehav. Rev. 2004;28:201. [PubMed: 15172764]

128.

Dragoi G, Harris KD, Buzsaki G. Place representation within hippocampal networks is modified by long-term potentiation. Neuron. 2003;39:843. [PubMed: 12948450]

129.

Treves A, Rolls ET. Computational analysis of the role of the hippocampus in memory. Hippocampus. 1994;4:374. [PubMed: 7842058]

130.

Henze DA, Wittner L, Buzsaki G. Single granule cells reliably discharge targets in the hippocampal CA3 network in vivo. Nat. Neurosci. 2002;5:790. [PubMed: 12118256]

131.

Derrick BE, Martinez JL Jr. A unique opioid peptide-dependent form of long-term potentiation is found in the CA3 region of the rat hippocampus. Adv. Biosci. 1989;75:213.

132.

Do V, Martinez CO, Martinez JLM, Derrick BE. Long-term potentiation in direct perforant path projections to hippocampal area CA3 in vivo. J. Neurophys. 2002;87:669. [PubMed: 11826036]

133.

Kosub KA, Do VH, Derrick BE. NMDA receptor antagonists block heterosynaptic long-term depression (LTD) but not long-term potentiation (UP) in the CA3 region following lateral perforant path stimulation. Neurosci. Lett. 2005;374:29. [PubMed: 15631891]

134.

Ekstrom AD, Meltzer J, McNaughton BL, Barnes CA. NMDA receptor antagonism blocks experience-dependent expansion of hippocampal “place fields” Neuron. 2001;31:631. [PubMed: 11545721]

135.

Guzowski JF, Knierim JJ, Moser EI. Ensemble dynamics of hippocampal regions CA3 and CA1. Neuron. 2005;44:581. [PubMed: 15541306]