The Neuron and the Synapse: The Physical Unit of Change

The junction smaller than you can picture

Everything you will ever learn — every fact, every skill, every reflex you've built — is stored in the strength of junctions so small you can't really picture them. Not in your neurons, exactly. In the gaps between them.

You have on the order of 86 billion neurons. That number gets quoted a lot, as if the cells were the point. They aren't. A neuron on its own does very little. What matters is that each one connects to thousands of others, and the whole system runs into the trillions of connections. Those connections are called synapses, and the synapse — not the neuron — is the part that changes when you learn.

NP-01 told you the brain rewires itself. This lesson opens up the hardware and shows you the one component that actually does the rewiring. Meet the part that changes.

Why nobody explains this properly

People love to say "the brain rewires." It sounds precise. But almost nobody who says it can tell you what the wire is — where it runs, what a connection physically consists of, or what "strengthening" one actually means. Without that, "rewiring" is just a nicer poster than "you can do it."

You can't reason about change while the thing that changes is a black box. If you don't know what a synapse is, then spaced repetition, sleep, practice, and habit are all just advice you either follow or don't. Once you can see the junction — and see that it's adjustable — every one of those becomes a lever you understand, not a rule you obey.

So this lesson is deliberately mechanical. By the end you'll be able to trace a signal from one neuron to the next, name each part it passes through, and put your finger on the exact spot where a new memory physically lives.

Mechanism 1 — The anatomy of one neuron

Start with a single cell. A neuron has four working parts, and each does one job.

• Dendrites — the branching, tree-like fibres that receive. Signals from other neurons arrive here. A single neuron can have thousands of them.
• Soma (cell body) — the hub that integrates. It gathers everything the dendrites are receiving and adds it up.
• Axon — the long cable that transmits. If the soma decides to fire, the signal shoots down the axon.
• Axon terminals — the endings that hand off. At the far end, the axon splits into terminals that pass the signal on to the next neuron.

The signal travelling down the axon is electrical. It's called an action potential, and it has one crucial property: it is all-or-nothing. The neuron doesn't send a stronger or weaker spike depending on the input. It either fires a full spike or it fires nothing at all — like a trigger, not a dimmer. This matters more than it sounds, and we'll come back to it.

Read that flow once more. The first four steps happen inside one neuron. The fifth — the jump to the next cell — is where the interesting part is, and it's not electrical at all.

Mechanism 2 — The synapse: the gap that has to be crossed

Here's the fact that surprises people. The axon terminal of one neuron does not touch the next neuron. There's a gap. A tiny one — but a real, physical space between the sending cell and the receiving cell. That space is the synaptic cleft, and the whole junction around it is the synapse.

So how does an electrical spike cross a gap it can't jump? It doesn't. It converts. When the action potential reaches the axon terminal, it triggers the release of chemical messengers called neurotransmitters into the cleft. Those molecules drift across the gap and bind to receptors on the next neuron's dendrite. That binding starts a fresh electrical signal in the receiving cell.

So the relay is: electrical → chemical → electrical. Electrical down the axon, chemical across the gap, electrical again in the next neuron.

That conversion looks like a design flaw — why break a fast electrical signal to squirt chemicals across a gap? Because it's exactly what makes the connection adjustable. A soldered wire is either connected or not. A chemical junction can send more messenger or less, have more receptors or fewer, be strengthened or weakened. This junction is the unit of change. Hold that thought — Mechanism 4 is built on it.

Mechanism 3 — Fire or don't: excitation vs inhibition

Not every input tells the next neuron the same thing. Some push it toward firing; some push it away.

• Excitatory inputs nudge the receiving neuron closer to firing. The most common excitatory neurotransmitter is glutamate.
• Inhibitory inputs nudge it away from firing. The main inhibitory one is GABA.

Now remember the all-or-nothing rule from Mechanism 1. A neuron is receiving thousands of these pushes and pulls at once, some saying "fire," some saying "don't." The soma sums them all up, moment by moment. If the excitation minus the inhibition crosses a certain threshold, the neuron fires a full spike. If it falls short, nothing happens.

A neuron is a voting machine that fires only when the balance tips past its threshold. Scale that up: what you experience as "thinking" or "remembering" is not one cell doing something clever. It's a particular pattern of firing spreading across a network of these summing devices. Change which junctions are strong, and you change which patterns fire easily — which is where we're going next.

Mechanism 4 — Where learning actually lives: synaptic strength

You now have everything you need for the payoff. Learning is not neurons appearing or vanishing. It is a change in how much a given synapse influences the next neuron — how loudly one cell's vote counts in another's tally.

A synapse can be made stronger in a few concrete ways:

• The sending terminal releases more neurotransmitter per spike.
• The receiving dendrite grows more receptors, so it responds more to the same amount.
• Entirely new synapses form between the two cells — a physical, structural change.

That capacity for a synapse to change its strength has a name: synaptic plasticity. And it runs in both directions. A synapse strengthened by use undergoes long-term potentiation (LTP). A synapse weakened by disuse undergoes long-term depression (LTD).

LTP was first demonstrated in the 1970s — Bliss and Lømo showed in 1973 that stimulating a pathway in the hippocampus left its synapses lastingly strengthened afterwards. That was the physical footprint of NP-01's "neurons that fire together wire together." When you drilled that guitar chord or that vocabulary list, you weren't storing it "in" a neuron. You were potentiating specific synapses — turning up the volume on exact junctions so that the same pattern fires more easily next time.

A memory is not a thing kept somewhere. It's a set of synapses whose strengths have changed.

Mechanism 5 — Myelin: the speed upgrade

One more part, kept brief. Many axons are wrapped in a fatty insulation called myelin, laid down in segments along the cable. Myelin does for an axon what insulation does for a wire: it makes the signal travel faster and more reliably.

Here's the plasticity twist. Myelin isn't fixed either. Practising a skill can thicken the myelin along the pathway you're using. A better-insulated, faster-firing circuit is part of why a well-practised skill stops feeling effortful and starts feeling automatic — the signal is quite literally moving faster and more cleanly than it did when you were a beginner.

That's the headline; the deeper story (which cells build the myelin, how it's timed) comes in a later lesson. For now, just add it to the model: synapses change how strongly a signal passes, and myelin changes how fast it travels.

What this means for you

Not advice — just what follows mechanically from the anatomy.

Try it: build a labelled model (15 min)

The model isn't yours until you've drawn it. This is the output of the lesson.

Step 1 — Draw two neurons. On paper or in a note app, sketch two neurons side by side, with the axon of the first one reaching toward the dendrites of the second. Rough is fine — boxes and lines count.

Step 2 — Label the parts. On the first neuron, label: dendrite, soma (cell body), axon, and axon terminal. In the gap between the two neurons, label the synapse and draw a few dots for neurotransmitter crossing the cleft to the second neuron's dendrite.

Step 3 — Draw the signal direction. Add an arrow showing the signal's path: dendrite → soma → axon → terminal → across the synapse → next dendrite. Mark the stretch along the axon "electrical" and the gap "chemical."

Step 4 — Write the one annotation that matters. Next to the synapse, answer this in one line: "Where, exactly, does a new memory physically live?" Your answer: in the changed strength of the synapse — more neurotransmitter, more receptors, or a new junction.

✅ Finish check: you have a labelled two-neuron diagram showing the signal path and the electrical→chemical→electrical relay, with the synapse marked as the site of change. You can now point to one spot on your own drawing and say in a sentence where learning physically happens — in the tuned strength of the synapse, not anywhere inside the neuron. Keep this diagram — NP-03 builds directly on it.

Summary card

• You have ~86 billion neurons, but the action of learning happens at the trillions of synapses between them. The synapse, not the neuron, is the unit of change.
• One neuron has four parts: dendrites receive, the soma integrates, the axon transmits, and the axon terminals hand off.
• The signal down the axon is electrical and all-or-nothing (the action potential). Across the synaptic gap it converts to chemical messengers (neurotransmitters), then back to electrical — electrical → chemical → electrical.
• A neuron sums excitatory (e.g. glutamate) and inhibitory (e.g. GABA) inputs and fires only when the balance crosses its threshold. Thought and memory are patterns of firing across networks.
• Learning = changed synaptic strength (synaptic plasticity): more transmitter, more receptors, or new synapses. Strengthened by use = LTP; weakened by disuse = LTD. LTP was first demonstrated in the 1970s.
• Myelin insulates axons and thickens with practice, making practised circuits faster — part of why skills come to feel automatic.

Next lesson: NP-03 — "Neurons That Fire Together Wire Together": Hebbian Learning (L1) Related: NP-01 What Neuroplasticity Actually Is · NP-03 Hebbian Learning · NP-06 Sleep and Memory Consolidation