Introduction to Behavioral Neuroscience · Learning and Memory
Synaptic Mechanisms of Long-Term Memory
On this page 9 sections
In 30 seconds
Where is a memory physically stored? The leading answer, supported by decades of research, is: in the strength of synapses. When two neurons repeatedly fire together, the connection between them can become stronger — long-term potentiation (LTP) — or weaker — long-term depression (LTD) — and these changes are widely regarded as cellular mechanisms of learning and memory. This topic walks through the molecular machinery: how glutamate receptors (AMPA and NMDA) detect coincident activity, how calcium entry triggers signaling cascades, how Early LTP Fast phase; existing proteins modified; no gene expression Full entry → becomes Late LTP Slow phase; new gene expression and protein synthesis Full entry → requiring gene expression and new proteins, and how the neuron physically remodels — growing dendritic spines and strengthening connections. The Aplysia work from the implicit-learning topic and the LTP work in the rodent hippocampus converge on one principle: experience alters synaptic strength, and altered synaptic strength is the memory. As with all models, LTP is a candidate mechanism, not the whole story — no single experiment has yet proven that a specific LTP change is a specific memory, and current research continues to refine the picture.
Why this matters
This is the mechanism behind everything else in the chapter. Understanding LTP and LTD explains why spaced practice beats cramming (consolidation needs protein synthesis and time), why sleep matters, and why amnesia patients cannot form new explicit memories (their hippocampal circuits cannot produce lasting synaptic change). It also frames real-world phenomena: drugs that affect glutamate or calcium signaling can impair memory formation; Alzheimer's disease involves synapse loss and dysfunction before cell death; and "memory-enhancing" claims, from supplements to nootropics, should be evaluated against this biology. For exam purposes, the molecular steps of LTP — NMDA receptor Glutamate receptor blocked by Mg²⁺ at rest Full entry →, magnesium block, calcium, CaMKII Kinase activated by calcium; central to early LTP Full entry →, AMPA insertion — are classic test material, and the early/late LTP distinction is a favorite short-answer question.
The college version
Core Concepts
Hebb's rule: the cellular version of "practice makes perfect"
In 1949, psychologist Donald Hebb proposed that when neuron A repeatedly and persistently takes part in firing neuron B, the connection from A to B is strengthened — popularly summarized as "neurons that fire together, wire together." Hebb's rule Coincident firing strengthens connections Full entry → predicted exactly what researchers later found: synapses strengthen when the presynaptic cell's activity reliably coincides with postsynaptic depolarization. The rule is the conceptual bridge between behavior (repeated pairing, rehearsal, conditioning) and biology (changed connectivity).
Long-term potentiation (LTP): the synapse gets stronger
In 1973, Bliss and Lømo showed that a brief, high-frequency burst of stimulation to a pathway in the rabbit hippocampus produces a long-lasting increase in the strength of the stimulated synapses — LTP. It has the properties a memory mechanism should have: it is rapidly induced (seconds), persistent (hours to weeks in the lab, longer with repetition), associative (weak inputs are potentiated when they coincide with strong inputs — a cellular correlate of classical conditioning), and input-specific (only the stimulated synapses change, not all synapses on the neuron). LTP is most easily studied in the hippocampus, which is fitting — the hippocampus is the structure whose damage destroys new explicit memory formation.
The molecular cast: AMPA, NMDA, and the magnesium block
The key synapse in hippocampal LTP uses the neurotransmitter glutamate, which binds two receptor types on the postsynaptic cell. AMPA receptors are the workhorses: they open quickly, let sodium in, and produce the normal fast excitation. NMDA receptors are the coincidence detectors: they also let ions through, but at resting potential their pore is blocked by a magnesium ion, so nothing flows even when glutamate binds. The block is removed only when the postsynaptic cell is already depolarized — which happens when AMPA receptors have just fired strongly. So the NMDA receptor opens only when presynaptic glutamate release coincides with strong postsynaptic depolarization: the cellular correlate of "two neurons firing together." When it opens, it admits Calcium (Ca²⁺) Ion that enters via NMDA receptors and triggers plasticity cascades Full entry →, the trigger ion.
Early LTP: calcium, kinases, and more AMPA receptors
The calcium that enters through NMDA receptors activates enzymes, notably CaMKII, which phosphorylates target proteins. Within minutes, this produces early LTP: the cell inserts additional AMPA receptors into the postsynaptic membrane, so each presynaptic action potential produces a larger response. Early LTP lasts minutes to a few hours and does not require new protein synthesis — it is a fast, local modification of existing molecules. It is the "working" phase: strong enough to hold the change through the immediate aftermath of learning, but not yet permanent.
Late LTP: genes, proteins, and structural change
For the change to last beyond a few hours, signaling must reach the nucleus. Calcium and kinase cascades (including the transcription factor CREB Transcription factor that switches on memory-related genes Full entry →) trigger gene expression and protein synthesis, producing late LTP. New proteins are shipped to the synapse, where they stabilize the added AMPA receptors and drive structural change: the postsynaptic density grows, dendritic spines enlarge or new spines form, and in some cases the presynaptic terminal releases more transmitter. These physical changes are the reason late LTP lasts days or longer — and why protein synthesis inhibitors or sleep deprivation (which impairs the molecular machinery) block the persistence of memory. Repetition and spacing matter because each round of practice recruits this gene-expression program more effectively.
Long-term depression (LTD): the yin to LTP's yang
LTD is a long-lasting decrease in synaptic strength, typically produced by low-frequency stimulation that lets in a smaller, slower rise in calcium. Instead of CaMKII, LTD activates phosphatases that remove phosphate groups, causing AMPA receptors to be removed from the membrane. LTD is not just the opposite of LTP; it is biologically important in its own right. It may prune unused connections, refine circuits during development, and help the brain forget or update — and it is a candidate mechanism for the weakening of memories, including extinction learning. Together, LTP and LTD give the brain bidirectional control: up for learning, down for forgetting and fine-tuning.
Converging evidence from Aplysia
The Aplysia story from the implicit-learning topic provides the molecular echo. In sensitization, serotonin released by facilitatory interneurons activates a cascade — cAMP → protein kinase A (PKA) → CREB — that strengthens the sensory–motor synapse and, with repeated training, produces new synaptic connections. Notice the convergence: sensitization in the snail, LTP in the mammal, and late LTP in both use CREB-dependent gene expression for long-term change. Evolution conserved the molecular machinery of memory.
Limits of the model
LTP/LTD is a candidate mechanism, not a proven identity. Most LTP research uses artificial stimulation patterns in slices or anesthetized animals; whether the exact changes seen in the lab are the ones that encode a specific natural memory is still being worked out. Memory also involves network-level reorganization (systems consolidation), neuromodulators like dopamine and acetylcholine that gate plasticity, and multiple cell types (including inhibitory neurons and glia). The honest summary: Synaptic plasticity Change in the strength of a synapse with experience Full entry → is necessary for many forms of long-term memory, and LTP/LTD are its best-studied cellular candidates — a working model that has been refined for fifty years and continues to evolve.
Common Confusions
| Do Not Confuse | With | Difference |
|---|---|---|
| AMPA receptors | NMDA receptors | AMPA produce normal fast excitation and are inserted during LTP; NMDA are blocked by Mg²⁺ at rest and act as coincidence detectors admitting Ca²⁺. |
| Early LTP | Late LTP | Early LTP: minutes–hours, existing proteins, no gene expression. Late LTP: hours–days, CREB + protein synthesis + structural change. |
| LTP | LTD | LTP strengthens synapses (Ca²⁺ → CaMKII → AMPA insertion); LTD weakens them (low Ca²⁺ → phosphatases → AMPA removal). Different calcium dynamics. |
| Magnesium block removed | Receptor opens without glutamate | NMDA needs BOTH glutamate binding AND depolarization to unblock; neither alone suffices — that is what makes it a coincidence detector. |
| Synaptic consolidation | Systems consolidation | Synaptic: cellular changes at the synapse over minutes–hours. Systems: gradual hippocampal-to-neocortical reorganization over weeks–years. |
| LTP as memory itself | LTP as candidate mechanism | LTP has the right properties and is necessary for many memories, but no experiment has proven a specific LTP change is a specific memory. |
| Habituation (Aplysia) | Sensitization (Aplysia) | Habituation = weakened sensory–motor synapse; sensitization = strengthened synapse via serotonin → cAMP → PKA → CREB. |
Eli explains
The same idea, in plain words
Explain it like I’m 10
Imagine two friends passing a ball. The first time they play together, the pass is clumsy. Every time they practice, the pass gets smoother, and their muscles remember the move — that's like a synapse getting stronger. In your brain, when two neurons fire at the same time, a special doorway opens and lets in a tiny signal that says "strengthen this connection." At first the change is quick and small, like a sticky note. With practice and sleep, the brain builds new little branches and the connection becomes permanent, like writing it in a book. And connections you never use get weaker, so the brain can tidy up — that's the forgetting that keeps your brain from getting cluttered.
Worked example
You are learning a new chord. The first day, your motor cortex neurons that drive your fingers fire together with sensory feedback neurons, and in your hippocampus the neurons encoding "chord shape" and "sound" fire together repeatedly. Each co-firing opens NMDA receptors at those synapses, calcium floods in, and CaMKII adds AMPA receptors — early LTP forms within minutes, which is why the chord feels slightly easier by the end of the session. But if you never practice again, the added receptors are removed and the change fades — the memory is gone by next week. Instead you practice every other day and sleep between sessions: the spaced practice and sleep allow CREB-driven gene expression, protein synthesis, and new dendritic spines — late LTP. Now the chord is automatic: the skill has been consolidated at the synaptic level, and the hippocampal involvement fades as cortical circuits take over (systems consolidation). Meanwhile, the chord you abandoned years ago slowly undergoes LTD at its unused synapses — the brain's way of pruning what you stopped using.
Key takeaways
- Hebb's rule: neurons that fire together, wire together — coincident pre- and postsynaptic activity strengthens the synapse.
- LTP: persistent, input-specific strengthening of a synapse after strong stimulation; first shown in the rabbit hippocampus (Bliss & Lømo, 1973).
- The NMDA receptor is the coincidence detector: magnesium blocks its pore at rest; depolarization (via AMPA) removes the block, so calcium enters only when release and depolarization coincide.
- Calcium is the trigger: Ca²⁺ → CaMKII → AMPA receptor insertion = early LTP (minutes–hours, no new proteins).
- Late LTP needs gene expression: CREB, protein synthesis, new dendritic spines — lasting days; blocked by protein-synthesis inhibitors and impaired by poor sleep.
- LTD is the weakening counterpart: low-frequency activity, phosphatases, AMPA removal — important for pruning, updating, and extinction.
- Aplysia convergence: serotonin → cAMP → PKA → CREB strengthens the sensory–motor synapse; CREB is the conserved gateway to long-term change.
- Model limits: LTP/LTD are the leading candidate cellular mechanisms of memory, not yet proven to be the exact storage of any specific memory.
Check yourself
6 review questions from the chapter. Try each one, then open the answer.
State Hebb's rule in your own words and explain why the NMDA receptor is called a coincidence detector.
Show answer
Hebb's rule: when neuron A repeatedly helps fire neuron B, the A→B connection is strengthened ("fire together, wire together"). The NMDA receptor is a coincidence detector because its pore is magnesium-blocked at rest and opens only when glutamate binds AND the postsynaptic cell is already depolarized — so it responds only to simultaneous pre- and postsynaptic activity.
Why does calcium entry through NMDA receptors require both glutamate release and postsynaptic depolarization?
Show answer
Glutamate binding alone leaves the NMDA pore blocked by Mg²⁺ at resting potential. Strong presynaptic release depolarizes the postsynaptic cell via AMPA receptors; that depolarization repels the magnesium ion, and only then can calcium flow in. Both conditions are required.
Contrast early and late LTP: duration, molecular requirements, and structural consequences.
Show answer
Early LTP lasts minutes to a few hours, requires calcium → CaMKII and insertion of existing AMPA receptors, and needs no new protein synthesis. Late LTP lasts days or longer, requires calcium/CREB-driven gene expression and protein synthesis, and produces structural changes such as enlarged or new dendritic spines.
What is LTD, what molecular pathway produces it, and why might weakening synapses be useful?
Show answer
LTD is a long-lasting decrease in synaptic strength produced by low-frequency stimulation and a smaller calcium rise; it activates phosphatases that remove AMPA receptors from the membrane. Weakening is useful for pruning unused connections, refining circuits, and supporting forgetting and extinction learning — bidirectional control keeps the brain flexible.
How does the Aplysia sensitization cascade (serotonin → cAMP → PKA → CREB) converge with hippocampal LTP?
Show answer
Both use calcium/cAMP-driven cascades that converge on CREB-dependent gene expression to produce long-lasting strengthening (new proteins, new or stronger connections). The conservation of the CREB gateway across snails and mammals indicates the molecular machinery of long-term memory is evolutionarily ancient.
Why does spacing practice and getting sleep improve long-term retention from the perspective of synaptic mechanisms?
Show answer
Spaced practice and sleep let the protein-synthesis-dependent late phase of LTP proceed: each practice session reactivates the calcium cascades, sleep supports the molecular machinery and replay of activity, and new proteins and spines stabilize the connection. Cramming without sleep leaves only fragile early LTP, which decays.
Study tools & related lessonsKey vocabulary · Related
Key vocabulary
- Synaptic plasticity
- Change in the strength of a synapse with experience
- LTP (long-term potentiation)
- Long-lasting increase in synaptic strength after strong activity
- LTD (long-term depression)
- Long-lasting decrease in synaptic strength after weak/low-frequency activity
- Hebb's rule
- Coincident firing strengthens connections
- AMPA receptor
- Fast glutamate receptor producing normal excitation
- NMDA receptor
- Glutamate receptor blocked by Mg²⁺ at rest
- Calcium (Ca²⁺)
- Ion that enters via NMDA receptors and triggers plasticity cascades
- CaMKII
- Kinase activated by calcium; central to early LTP
- CREB
- Transcription factor that switches on memory-related genes
- Early LTP
- Fast phase; existing proteins modified; no gene expression
- Late LTP
- Slow phase; new gene expression and protein synthesis
- Dendritic spine
- Small protrusion on dendrites where synapses form
- Reconsolidation window
- Period when a retrieved memory is labile and updatable
Sources & references
This lesson was adapted from the open educational references above; their licenses and attributions are preserved. See Copyright & Licensing.
Educational content only. It is not medical, legal or professional advice. Found an error? Tell us.