Why are your nerves wrapped in fat?

Picture 1

The race

Each nerve is 2 centimetres long, fired at its left end. Everything is slowed down five thousand times: the real race is over in about a thousandth of a second.

The wrapped nerve wins easily. In this simulation it runs at 34 metres per second: twelve times as fast as the same nerve bare (2.7 m/s), and nearly twice as fast as the squid's (19 m/s), which has over 2,000 times its cross-section. (Over these 2 cm it wins by more, 0.48 ms to 1.26, partly from a quicker start.) And watch its gaps: only they fire, flashing one after another as the signal passes.

Why does wrapping win? A nerve's signal moves by charging up the membrane ahead of it, and charging takes time. Wrapping means the long stretches between the gaps need only a sliver of the charge, so the signal races across them. To make a bare nerve twice as fast, you'd have to make it four times as wide.

Real nerves differ in the details. Your fastest ones are warmer and more thickly wrapped and reach about 120 m/s; your thinnest bare ones manage about 1 m/s. The race comes out the same way.

Picture 2

The wave

A nerve signal isn't like current in a copper wire, pushed through from one end. It's renewed at every patch: each patch of membrane fires by itself, when the patch behind it has charged it up past a threshold. The membrane is a thin insulator with salt water on both sides, so it's a capacitor, and charging a capacitor takes time. In a bare nerve, every bit of membrane has to be charged before it fires, so the signal crawls.

Myelin is a stack of membranes wrapped tightly round the nerve, 200 here. Stacked up, they make a far thicker insulator, which takes far less charge to bring up to the same voltage: each wrapped stretch needs only a sliver of what bare membrane would. (Myelin also seals the nerve, so little current leaks out, but what speeds it up is the charge.) And it has no channels, so only the gaps between the wrapping fire. When one gap fires, its current runs along the inside, charging the wrapped stretch on the way, and brings the next gap up to its threshold.

The bare nerve's spike is a short bump creeping along. The wrapped one's is so fast that it's longer than the whole nerve: about seven gaps are rising at any moment, and each wrapped stretch just follows the gaps at its ends. Thicker wrapping charges faster still: slide the myelin up to 800 layers and it runs at about 70 m/s. (Here the wrapping takes no room. Real myelin does, so for a nerve of a given overall width there's a best balance, with the axon about 60 to 70% of the width: wrap more and the axon inside must be thinner.)

Picture 3

Who carries it

So what actually travels down a nerve? Look inside the squid's, the widest. At the front of the spike, sodium channels open and Na⁺ rushes in from outside, which is what fires the membrane. Behind it, the sodium channels shut themselves, and potassium channels open and K⁺ flows out, resetting it. The shut sodium channels stay shut for a moment, which is why the signal can't turn round and run backwards.

But the current that runs along the inside of the nerve, charging the next patch, is mostly potassium: more than half of it. Inside a nerve, K⁺ is the most plentiful ion, so it carries most of whatever current flows; the large negative ions that can't leave the nerve carry much of the rest, moving the other way. Sodium starts the signal; along the inside, potassium carries most of its current.

And nothing goes far. The signal shifts each ion inside by less than a nanometre. In the same millisecond, each ion's own random jiggling moves it thousands of times further. What travels is the wave, the way a wave goes round a stadium while everyone stays in their seat.

So

Why fat?

A nerve signal is a chain of patches, each charging the next one up. There are two ways to make it faster. One is to make the nerve thicker, so the current along the inside meets less resistance; but a bare nerve's speed grows only as the square root of its width, so twice as fast costs four times as wide. Squid took that road, for the nerve that fires their escape jet. The other is to wrap the nerve, so that only the gaps fire and the long stretches between them need only a sliver of the charge. Most animals with backbones took that road (a few others, prawns among them, found it on their own), and it's why a nerve a tenth the width of a hair can beat the squid's. Wrapping saves energy too: only the gaps let sodium in, so the pumps that bail it back out afterwards have far less to do.

And while we're here

Other questions the pictures answer

Why does a stubbed toe hurt twice?

First a sharp pain, then a second or so later, a dull ache. Your pain nerves come in both kinds. Thin wrapped ones, at about 5 to 30 m/s, bring the sharp first pain; thin bare ones, at about 1 m/s, bring the ache, and over a metre from toe to spine, the difference is most of a second.

What happens when myelin is damaged?

Slide the myelin down to its fewest layers. Too thin, and each wrapped stretch soaks up so much charge that the next gap never reaches its threshold: the signal stops. In multiple sclerosis, the body's immune system damages myelin in the brain and spinal cord, and signals through the damaged stretches slow down or fail.

Under the hood

How a nerve is simulated

The race alone is a classic. In 1952 Hodgkin and Huxley described the squid nerve's channels with a set of equations and worked out by hand how fast its signal travels: 18.8 m/s (they measured 21.2). Their method, channels joined to the cable equation, which treats a nerve as a leaky electrical wire, is still what neuroscience's simulators run. This page checks against them: its squid nerve runs at 18.7 m/s.

What this simulation adds is the ions. It follows each species inside the nerve (K⁺, Na⁺, Cl⁻, and the large negative ions that can't cross the membrane) as each one diffuses and drifts down its own electrochemical potential: the same drift–diffusion rule that runs our lithium-ion battery. That's how it can say which ions carry the current in picture 3. A cable model only knows that a current flows.

It leaves things out, to run live: the nerve has one dimension, its length; the fluid outside is held fixed; every nerve uses Hodgkin and Huxley's squid channels, at their 18.5 °C; and myelin is modelled as a perfect seal that divides the membrane's charge by its number of layers, with gaps 1 µm wide every millimetre. Real mammal nerves are warmer, with channels of their own, and faster.