Long-Term Potentiation: Part One: The Phenomenon, and Its Opposite

Abstract

Long-term potentiation is the most studied mechanism in neuroscience. Stimulate a pathway in the right pattern and the synapses along it become stronger, for hours or days. It was first described in 1973, and in 2024 the Royal Society held a discussion meeting under the title "Long-term potentiation: 50 years on".

This is the first of three parts, one per publication. It follows a review from that meeting, and its subject is the half of the story that popular accounts leave out: the mechanism that weakens synapses.

A system that can only strengthen connections saturates. What makes plasticity useful is that it runs in both directions.

1.What Is Not in Dispute

Hardy Hagena and Denise Manahan-Vaughan open from a position of consensus, and it is worth quoting because so much of what follows in this series is contested.

There is broad agreement that persistent, use-driven and experience-dependent modifications of synaptic strength form the cellular basis of long-term information storage and memory updating by the hippocampus. Associative learning drives long-lasting changes in synaptic efficacy; decades of evidence support this.

The forms that matter are those persisting over hours or days — long enough to retain information and to open time windows in which past and prospective experience can be associated.

So the phenomenon is real, reproducible and central. The arguments in Parts Two and Three are about what it is doing, not whether it happens.

2.The Missing Half

Long-term depression is the mirror process: patterns of activity that persistently weaken synaptic transmission rather than strengthening it.

It attracts a fraction of the attention, partly because weakening sounds like the absence of learning rather than a form of it. That intuition is wrong in a specific way.

A network in which synapses can only strengthen runs out of dynamic range. Every experience pushes weights up; eventually they saturate and nothing further can be encoded. Bidirectional plasticity is not a refinement on the model — it is what makes the model workable.

The review's framing is that the interplay of potentiation and depression is what enables memory representations. Not potentiation with depression as housekeeping, but the two together defining what a representation is.

3.The Processes That Sit On Top

Three additional mechanisms appear in the review, and each complicates the simple picture in a useful way.

Metaplasticity is plasticity of plasticity: prior activity changes how readily a synapse will subsequently potentiate or depress. The rules themselves are not fixed.

Synaptic tagging addresses a problem that would otherwise be fatal to the account. Proteins required to make a change persistent are synthesised in the cell body, but changes occur at individual synapses. Tagging proposes that an active synapse is marked, and captures those proteins when they arrive — which explains how a cell-wide resource reaches a synapse-specific destination.

Homeostatic plasticity keeps overall activity within a workable range, scaling synaptic weights so that a neuron neither falls silent nor fires continuously.

Together these mean that a synapse's strength at any moment reflects its own history, the cell's recent protein synthesis, and a network-level set point. "LTP encodes the memory" is a considerable simplification of that.

4.From Synapses to Ensembles

The reason this matters beyond synaptic physiology is the step the authors take next: persistent plasticity is likely to enable the generation of linked neuronal ensembles that support the creation and updating of memory representations.

That is the bridge between this literature and the engram work covered elsewhere in this section. A memory is not held at a synapse; it is held in a population, and plasticity is proposed as the means by which the population becomes linked.

It also sets up the tension the rest of this series examines. Our engram series described experiments in which a memory survived the loss of synaptic potentiation entirely — the trace persisted while the strengthening did not. That result does not fit comfortably with the account above, and Part Three takes up an alternative.

5.What the Evidence Is Made Of

This literature is rodent, and predominantly hippocampal. It rests on electrode recordings in slices and in freely behaving rats, on genetic manipulation of mice, and on pharmacological blockade.

Human evidence for LTP-like plasticity exists — from paired-associative stimulation and from sensory tetanisation protocols — but it measures an evoked response rather than a memory, and it is a small literature beside the animal work.

Fifty years of work has established the phenomenon thoroughly. What it has not established is a demonstration in a behaving animal that a specific memory is stored as a specific set of potentiated synapses, which remains the field's outstanding problem.

Editorial Comment

MindHeaven® makes no claim relating to long-term potentiation, synaptic plasticity or memory formation, and no ingredient in our formulations has been shown to influence any of them.

We publish this series partly as a corrective to how the term is used commercially. "Supports synaptic plasticity" appears on products across this category, and it works precisely because plasticity sounds like a capacity that could be topped up. On the evidence here it is a set of bidirectional, history-dependent, homeostatically constrained processes — and more of it is not obviously better.

Part Two follows the search for the molecule that makes potentiation persist, which is the most concrete question in the field and has taken forty years to approach an answer.

A note on sources. This topic was filed with three reviews; two of them — a history of the LTP controversy in Hippocampus and a molecular review in Neuroscience Research — are closed access with no open copy available. We substituted three open reviews, including two written for the Royal Society's fiftieth-anniversary meeting and a major review in Physiological Reviews. We consider the replacements stronger than the originals, and record the substitution here rather than leaving it invisible.