Info

FIGURE 7 The NMDA receptor antagonist D-AP5 (i.e., APV) blocks learning in the Morris water maze. These data are from the landmark paper by Richard Morris and his colleagues (141), demonstrating that infusion into the CNS of DL-AP5 (an active mixture) but not the control, inactive enantiomer L-AP5 blocks learning in the water maze task. (A) This panel illustrates that pre-training infusion of DL-AP5 blocks learning of a spatially selective search strategy for locating the hidden platform. (B) This panel illustrates that post-training blockade of NMDA receptors does not affect memory recall. (C) This panel illustrates that the same infusion protocol leads to effective blockade of NMDA receptor-dependent LTP in the dentate gyrus. See text for additional discussion.

in the water maze (see reference 36 and Figure 7). A few years later Richard shocked us by revealing that this effect was not the result of a loss of the capacity of the animal to learn the spatial relationship of the hidden platform to the visual cues (37). Rather, NMDA receptor blockade appears to block the capacity of the animal to learn the task, that is to learn that there is a consistent relationship between the spatial cues and the hidden platform, and that they can use spatial cues to predict where the platform will be located. The NMDA receptor (and by inference LTP) is not necessary for spatial learning per se—it is necessary for learning more complex relationships about spatial information. These pioneering data are in nice agreement with the more recent work indicating that NMDA receptor-dependent processes are necessary for an animal to reconstitute a special representation from partial visual stimuli (35).

How might NMDA receptor-dependent LTP contribute to multimodal information processing in the hippocampus? This is certainly not clear at present. However, one can imagine a couple of ways in which the cellular and molecular properties of LTP induction might contribute to the processing of complex associations and the establishment of a lasting representation of the association. Please keep in mind that these examples are pronouncedly oversimplified. They are not based on realistic circuits nor are they nearly sufficient to account for the entirety of the observations. They simply serve to provide a frame of reference for how the molecular and cellular processes we have been discussing might enter into our thinking about unique roles for hippocampal LTP in forming complex associations.

Two possibilities are shown in Figure 8. In the first example imagine that initially a pyramidal neuron receives one strong input (Input 1) and two weak inputs (Inputs 2 and 3). Either Input 1 by itself or Inputs 2 and 3 firing simultaneously can trigger an action potential in their follower neuron. Now imagine that you have paired activity in Input 1 plus Input 2, causing LTP at Input 2. Similarly, you get paired activity of Input 1 and Input 3 and get LTP at Input 3. Now Inputs 1, 2, and 3 are all "strong" inputs and capable of firing an action potential. Consequently, Input 2 alone or Input 3 alone can reconstitute the response that previously required both Inputs 2 and 3.

A second example is conceptually similar. Imagine that a pyramidal neuron receives a strong input (Input 1), a modu-latory input such as ACh (Input 2) and a weak input (Input 3). Input 1 triggers a back-propagating action potential, which ACh modulation of dendritic K channels allows to propagate into the distal den-drites. This back-propagating action potential is paired with Input 3, causing LTP at this site. Input 3 is now sufficient to cause an action potential on its own and give a readout equivalent to Input 1. Thus, Input 3, by virtue of its association with a salience signal (ACh in this example), is now uniquely able to trigger the same response as Input 1.

As I emphasized earlier, these examples are not realistic models to try to account for the complex behavioral changes described in the earlier parts of this section. They are merely simple examples to give an idea of how multiple coincidence detection mechanisms, which we know exist in hippocam-pal pyramidal neurons, might play a part in representing complex associations. A much greater understanding of the particulars of the relevant neuronal circuits will be necessary to generate realistic models. Nevertheless, I think these examples give a taste of a few of the ways that LTP might be involved in the complex sensory information processing by the hippocampus.

B. Short-Term Information Storage in the Hippocampus

A second and distinct role for LTP that we will touch on briefly is its participation as a mechanism for short-term information storage in the hippocampus. At first blush, this may seem oxymoronic—long-term potentiation as a mechanism for short-term information storage. However, it's important to remember that what we call E-LTP in vitro may not last long in the behaving animal. E-LTP or other decremental forms of LTP may only last 15 minutes in vivo owing to the increased rates of reaction of the underlying biochemistry at 37°C. Moreover, there is no reason to think that LTP might not be established and then specifically erased after its role was finished. These two considerations bring to mind the findings by Moser et al. (16, 17), where they observed learning-associated synaptic potentiation in vivo that lasted for about 15 minutes or so.

One experimental observation that specifically prompts our concluding that LTP is involved in short-term information storage in the hippocampus is the finding that NMDA receptor activation is necessary for trace fear conditioning. Huerta et al. (38) used their sophisticated engineered mouse lacking NMDA receptors in area CA1 in order to probe the role of the hippocampus and NMDA receptor-dependent processes in time-dependent learning. The specific paradigm that they used was trace fear conditioning. They found that the

FIGURE 8 LTP in multimodal information processing. (A) Example 1 illustrates how pairing synaptic activity of strong plus weak activity can allow a single input to achieve the same effect that formerly required its activity plus another input. Associative activity allows a single input to represent subsequently either an entirely different input from its original meaning or a partial re-presentation of an original stimulus to reconstitute the entire original effect. (B) Example 2 illustrates how an excitatory input (Input 1) coupled with a modulatory input (ACh in this example) allows a different input (Input 3) to trigger a new response. The new response to Input 3 is now functionally equivalent to the original response to Input 1. See text for additional discussion.

FIGURE 8 LTP in multimodal information processing. (A) Example 1 illustrates how pairing synaptic activity of strong plus weak activity can allow a single input to achieve the same effect that formerly required its activity plus another input. Associative activity allows a single input to represent subsequently either an entirely different input from its original meaning or a partial re-presentation of an original stimulus to reconstitute the entire original effect. (B) Example 2 illustrates how an excitatory input (Input 1) coupled with a modulatory input (ACh in this example) allows a different input (Input 3) to trigger a new response. The new response to Input 3 is now functionally equivalent to the original response to Input 1. See text for additional discussion.

introduction of even a brief 30-second delay between CS and US presentation rendered the learning dependent upon hippocampal NMDA receptors. These data strongly suggest that one role for LTP in area CA1 is the temporary storage of information so that events can be associated over time.

In an earlier work, Richard Morris made a conceptually similar observation indicating a role for NMDA receptor-dependent processes in short-term information storage. Richard trained animals in delayed match-to-place task and found that the extent of NMDA receptor dependency varied based on the length of the delay period between stimulus and match (37).

Again, these data indicate a role for LTP or similar processes in temporary information storage.

The general idea coming out of studies of this sort is that LTP or a similar process in the hippocampus is involved in the storage of episodes of experience for brief periods of time (see references 38-41). This allows that the episode can be processed into the appropriate temporal context (i.e., what came before, what came after) so that associations can be made between one event or sequence of events and another. As with the other roles for LTP that we have been discussing, the precise function that LTP might play in contributing to this phenomenon is not clear (see reference 38). Nevertheless, it does not seem like such a great leap to think that an activity-dependent phenomenon that results in a persisting change (LTP) could contribute to short-term information buffering in a relatively straightforward fashion.

C. Consolidation of Long-Term Memory—Storage of Information Within the Hippocampus for Downloading to the Cortex

In this section, we will discuss the possible role of LTP in the classic function of the hippocampus—memory consolidation. As we have already discussed in several sections of the book, a wide variety of evidence has already shown that the hippocampus is involved in the consolidation of long-term memories. For example lesion studies including studies of human patients have shown this to be the case. Also numerous drug infusion studies, some of which will be described later, have shown a role for the hippocampus in memory consolidation. A key point with the drug infusion studies is that drugs can be infused into the hippocampus post-training and interfere with long-term memory formation. Appreciation of the significance of this was what led to the distinction of hippocampal memory consolidation as a distinct process from the initial events triggering memory formation.

A number of different experiments have shown that hippocampal protein synthesis and mRNA synthesis is necessary for the consolidation of long-term memories (42, 43). In fact, these same studies make it clear that multiple stages of protein-synthesis-dependent cellular processes are required for memory consolidation, as injection of protein synthesis inhibitors at different time points after training can lead to disruption of memory consolidation.

Moreover, specific molecular processes such as ERK activation, CREB activation, Arc induction, and C/EBP induction are necessary for hippocampus-dependent memory consolidation (19, 44-46). Our old friend the NMDA receptor is also necessary for hippocampus-dependent memory consolidation (47-50). The necessity of these specific molecular events for memory consolidation, known also to be necessary for LTP induction, strongly implicates LTP as a component of memory consolidation. Specifically, the triggering of LTP in the hippocampus is hypothesized to be necessary for consolidation of memories that are stored in the cerebral cortex.

Additional lines of evidence support this conclusion as well. Riedel et al. (51) showed that synaptic activity in the hippocampus (i.e., AMPA/kainate receptor function) is necessary for hippocampus-dependent memory consolidation. More direct evidence for a role for LTP in hippocampal memory consolidation was obtained in an additional study by Brun et al. (52). These investigators showed that delivering LTP-inducing stimulation to the dentate gyrus, after training, led to disruption of memory consolidation. This effect was blocked by NMDA receptor blockade, demonstrating that the triggering of LTP or a related phenomenon, as opposed to network firing, is what is disrupting the consolidation.

Thus, taken together there is a substantial body of direct and indirect findings that indicate that LTP is participating in the consolidation of hippocampus-dependent memory formation. The relevant LTP is occurring in the hippocampus, and it is triggering changes downstream in cortical targets of the hippocampus that store the memory. We will return to a broad-brush-stroke model of how this might happen a little later.

It is important to note that the relevant LTP is probably not triggered immediately, and may even be triggered multiple times as part of the consolidation of a single memory. NMDA receptor antagonist studies make it clear that an NMDA receptor-independent mechanisms exists

Was this article helpful?

0 0
Eliminating Stress and Anxiety From Your Life

Eliminating Stress and Anxiety From Your Life

It seems like you hear it all the time from nearly every one you know I'm SO stressed out!? Pressures abound in this world today. Those pressures cause stress and anxiety, and often we are ill-equipped to deal with those stressors that trigger anxiety and other feelings that can make us sick. Literally, sick.

Get My Free Ebook


Post a comment