Action Potential Propagation: Myelin and the Speed of a Nerve Signal

A spike, not a signal that just travels

An axon is a poor electrical cable in the ordinary sense — its cytoplasm is a weak conductor and its membrane leaks — so a passive electrical pulse injected at one end would decay to nothing within a couple of millimetres. What actually propagates along a real axon is not a fading pulse but a self-regenerating one: the action potential. At rest the membrane sits around −70 mV, held there by ion pumps and leak channels. When local depolarization pushes the voltage past a threshold near −55 mV, voltage-gated sodium channels in that patch of membrane snap open, sodium rushes in, and the local voltage rockets upward to roughly +40 mV before voltage-gated potassium channels open and sodium channels inactivate, driving the voltage back down and briefly below rest.

The Hodgkin-Huxley cable equation

Alan Hodgkin and Andrew Huxley's 1952 model, built from voltage-clamp experiments on the squid giant axon, describes each patch of membrane as a capacitor (the lipid bilayer) in parallel with three variable conductances — sodium, potassium, and a small leak — each governed by voltage- and time-dependent gating variables they called m, h and n. Coupling many such patches along the axon's length through the cytoplasm's axial resistance turns this into a cable equation: current from an active, depolarized patch spreads passively forward through the cytoplasm, and if it's strong enough to push the neighbouring patch's own sodium channels past threshold, that patch fires its own full-amplitude spike rather than simply relaying an attenuated echo. The signal that "travels" is really a chain of local regenerations, which is why the spike arrives downstream exactly as large as it started — an all-or-nothing digital pulse riding on an otherwise leaky analog cable.

// Hodgkin-Huxley membrane current, per unit patch of axon
C * dV/dt = -[ gNa*m^3*h*(V-ENa) + gK*n^4*(V-EK) + gL*(V-EL) ] + Iaxial
// Iaxial = current flowing in passively from the neighbouring active patch
// m, h, n each relax toward voltage-dependent steady states with their own time constants

// a patch fires only once its neighbour's Iaxial pushes V past threshold (~ -55 mV)
// just-fired patches are refractory (h ≈ 0) → wave cannot regenerate backward

One-way traffic: the refractory period

Passive current from an active patch spreads in both directions along the cable, so why doesn't the spike also regenerate backward into the patch it just came from? Because the sodium channels in that patch have entered an inactivated state — a separate gate (Hodgkin and Huxley's h variable) that closes shortly after opening and cannot reopen until the membrane has repolarized and recovered for roughly a millisecond, the absolute refractory period. Current spreading backward simply finds no channels able to respond. This one-way ratchet is what turns a symmetric physical cable into a directional signal, and it also sets an upper bound on firing rate — an axon cannot fire a second spike until the previous patch's channels have recovered.

Myelin: skip the slow parts

In unmyelinated axons the spike must regenerate at essentially every patch, and conduction speed is limited by how fast passive current can charge up the membrane capacitance of each successive patch — increasing speed then mostly means increasing axon diameter, which is metabolically expensive and space-hungry (the squid's giant axon is a famous example of that brute-force solution). Vertebrate axons instead wrap long stretches in myelin, a fatty sheath laid down by oligodendrocytes (in the central nervous system) or Schwann cells (in the periphery), leaving only small gaps — the nodes of Ranvier — exposed and densely packed with voltage-gated sodium channels. Myelin sharply reduces the membrane's capacitance and increases its resistance under the sheath, so passive current barely leaks out along the insulated stretch and reaches the next node almost undiminished. The spike then only has to actively regenerate at the nodes, appearing to jump from node to node — saltatory conduction (from the Latin saltare, to leap) — which is both dramatically faster and far cheaper in ion-pumping energy than firing continuously along the whole length.

Frequently asked questions

Try it live

The full cable model runs live in Action Potential Propagation. Toggle myelination on and off along the same axon and watch the spike switch from a slow, continuous crawl to a fast leap between nodes of Ranvier.