Long Term Potentiation as a Physiological Phenomenon

I. Synapses in the Hippocampus—The Hippocampal Circuit II. A Breakthrough Discovery—LTP in the Hippocampus

A. The Hippocampal Slice Preparation

B. Measuring Synaptic Transmission in the Hippocampal Slice

III. NMDA Receptor-Dependence of LTP

A. Pairing LTP

B. Dendritic Action Potentials

IV. NMDA Receptor-Independent LTP


C. Mossy Fiber LTP in Area CA3

V. A Role for Calcium Influx in NMDA Receptor-Dependent LTP VI. Summary


In the last two chapters we discussed a number of different hippocampus-dependent forms of learning and memory and the idea that in several instances the hippocampus serves as a short- and long-term "memory buffer." In this chapter, we will try to understand how it is that lasting changes in function are achieved in the hippocampus and elsewhere in the brain, focusing our attention at the cellular level. This represents a landmark step forward: our first attempt at formulating a hypothesis concerning the precise memory-related events occurring in the CNS that are lasting changes.

First off, I must note that there is a huge gap in our knowledge between the issues of Chapter 2 and 3, behavior, and the types of persistent neuronal modifications we will be focusing on for most of the rest of the book. The particular circuits and neuronal connections that underlie most forms of mammalian learning and memory are mysterious at present, particularly for hippocampus-dependent forms of learning. There really is very little understanding of the means by which complex memories are stored and recalled at the neural circuit level—this will be a very important avenue of future research.

All is not lost, however. Despite our limited understanding of the particulars of the circuitry involved, a general hypothesis of memory storage is available and broadly accepted. This hypothesis is that:

Memories are stored as alterations in the strength of synaptic connections between neurons in the CNS.

The significance of this general hypothesis should be emphasized—this is one of the few areas of contemporary cognitive research for which there is a unifying hypothesis. This makes many of us optimistic that memory will be the first highorder cognitive process to be understood at the cellular and molecular level.

This general hypothesis has a solid underlying rationale. As described in the first two chapters, learning and memory manifest themselves as a change in an animal's behavior, and scientists capitalize upon this in order to study these phenomena by observing and measuring changes in an animal's behavior in the wild or in experimental situations. However, all of the behavior exhibited by an animal is a result of activity in the animal's nervous system. The nervous system comprises many kinds of cells, but the primary functional units of the nervous system are neurons. Because neurons are cells, all of an animal's behavioral repertoire is a manifestation of an underlying cellular phenomenon. By extension, changes in an animal's behavior such as occurs with learning must also be subserved by an underlying cellular change.

By and large, the vast majority of the communication between neurons in the nervous system occurs at synapses, and a generic synapse can be thought of as consisting of a presynaptic component, a postsynaptic component, and a synaptic cleft. Communication between the presy-naptic component and the postsynaptic component occurs across the synaptic cleft and is mediated by a chemical species, a neurotransmitter. Neurotransmitter is synthesized in the presynaptic cell and released in response to excitation of the presynaptic neuron. The neurotransmitter then diffuses across the synaptic cleft, where it binds to specific receptors on the postsynaptic cell.

As synapses mediate the neuron-neuron communication that underlies an animal's behavior, changes in behavior are ultimately subserved by alterations in the nature, strength, or number of interneuronal synap-tic contacts in the animal's nervous system. The capacity for alterations of synaptic connections between neurons is referred to as synaptic plasticity, and as described earlier one of the great unifying theories to emerge out of neuroscience research in the last century was that synaptic plasticity subserves learning and memory.

One of the pioneers in advancing this line of thinking was the Canadian psychologist Donald Hebb, who published his seminal formulation as what is now generally known as Hebb's Postulate:

When an axon of cell A . . . excites cell B and repeatedly or persistently takes part in firing it, some growth process or metabolic change takes place in one or both cells so that A's efficiency as one of the cells firing B is increased.

D. O. Hebb, The Organization of Behavior, 1949 (28).

Note the important contrast between Hebb's Postulate and its popular contemporary formulation—one (Hebb's) specifies cell firing and the other (the modern formulation) specifies synaptic change. These two phenomena are clearly different, and the current, exclusively synaptic, variant is incomplete. Changes in synapses are certainly important in information storage in the CNS, but we need to consider that the postsynaptic receptors sit in a membrane whose biophysical properties are carefully controlled. Regulation of membrane sodium channels, chloride channels, and potassium channels also contribute significantly to the net effect in the cell that any neurotransmitter-operated process can achieve.

Thus, limitations arise from ignoring potential long-term regulation of membrane biophysical properties. We need to consider that local changes in dendritic membrane excitability may be involved in cellular information processing, and also that global changes in cellular excitability that alter the likelihood of the cell firing an action potential may be a mechanism for information storage.

This last point has been criticized as too limiting because with global changes in excitability one loses the computational power of selectively altering the response at a single synaptic input (i.e., synapse specificity). However, we don't know how the neuron or the CNS compute a memory output. The fundamental unit of information storage may not be the synapse but the neuron. Future experiments will be necessary to resolve this issue; nevertheless, it is worthwhile to keep in mind the possibility that regulation of excitability as well as alterations in synaptic connections may play a role in memory storage.


As Hebb had postulated, most contemporary theories regarding the cellular basis of learning suggest that information storage is subserved by activity-dependent alterations at the synapse. Because we are focusing on hippocampus-dependent forms of memory for the most part, we should therefore ask the question: what is the synaptic structure of the hippocampus?

The main excitatory (i.e., glutamatergic) synaptic circuitry in the hippocampus, in overview, consists of three modules (see Figure 1 and references 1, 2, and 3). As we

FIGURE 1 The entorhinal/hippocampal system. (A) This panel diagrams the principal inputs, outputs, and intrinsic connections. (B) In this panel, the central components of the circuit are delineated in a more anatomically correct fashion, illustrating the principal intrinsic connections of the dentate gyrus and hippocampus proper. (C) This is an expansion of area CA1 showing some of the synaptic inputs onto a single pyramidal neuron in area CA1. See text for additional details. Hippocampal diagram reproduced with permission from Johnston and Wu (25).

FIGURE 1 The entorhinal/hippocampal system. (A) This panel diagrams the principal inputs, outputs, and intrinsic connections. (B) In this panel, the central components of the circuit are delineated in a more anatomically correct fashion, illustrating the principal intrinsic connections of the dentate gyrus and hippocampus proper. (C) This is an expansion of area CA1 showing some of the synaptic inputs onto a single pyramidal neuron in area CA1. See text for additional details. Hippocampal diagram reproduced with permission from Johnston and Wu (25).

discussed in the last chapter, information enters the dentate gyrus of the hippocampal formation from cortical and subcortical structures via the perforant path inputs from the entorhinal cortex (Figure 1). These inputs make synaptic connections with the dentate granule cells of the dentate gyrus. After synapsing in the dentate gyrus, information is moved to area CA3 via the mossy fiber pathway, which consists of the axonal outputs of the dentate granule cells and their connections with pyramidal neurons in area CA3. After synapsing in area CA3, information is moved to area CA1 via the Schaffer-collateral path, which consists largely of the axons of area CA3 pyramidal neurons along with other projections from area CA3 of the contralateral hippocampus as well. After synapsing in CA1, information exits the hippocampus via projections from CA1 pyramidal neurons and returns to subcortical and cortical structures.

The connections in this synaptic circuit are retained in a fairly impressive manner if one makes transverse slices of the hippocampus, because the inputs, "trisynaptic circuit," and outputs are laid out in a generally laminar fashion along the long axis of the hippocampal formation. This is a great advantage for in vitro electrophysiological experiments, which I will return to shortly.

I also should emphasize that the trisy-naptic circuit just outlined is a great oversimplification; there are a great many additional synaptic components of the hippocampus! I will highlight a few illustrative examples here, most of which we will return to later in the book (see also Figure 1). There are inhibitory gamma-amino-butyric acid containing (GABAergic) interneurons that make synaptic connections with all the principal excitatory neurons outlined earlier. These GABAergic inputs serve in both a feedforward and feedback fashion to control excitability. There are also many recurrent and collateral excitatory connections between the excitatory pyramidal neurons, particularly in the area CA3 region. There is a direct projection from the entorhinal cortex to the distal regions of CA1 pyramidal neuron dendrites, a pathway known as the stratum lacunosum moleculare.

Finally, there are many modulatory projections into the hippocampus that make synaptic connections with the principal neurons (see Figure 1 and Box 1). These inputs are via long projection fibers from various anatomical nuclei in the brain stem region, and they are, by and large, not directly excitatory or inhibitory but rather serve to modulate synaptic connectivity in a fairly subtle way. There are four predominant extrinsic modulatory projections into the hippocampus. First, there are inputs of norepinephrine (NE)-containing fibers that project from the locus ceruleus. Second, there are dopamine (DA)-containing fibers that arise from the substantia nigra. There also are inputs using acetylcholine (ACh) from the medial septal nucleus and 5-hydroxytrypramine (5HT, serotonin) from the raphe nuclei.

Brain Battalion

Brain Battalion

Get All The Support And Guidance You Need To Be A Success At Beefing Up Your Brain. This Book Is One Of The Most Valuable Resources In The World When It Comes To Understanding Your Brain to Come Up With A Battle Plan For Intelligence.

Get My Free Ebook

Post a comment