Hodgkin-Huxley Neuron Action Potential

Squid-axon membrane and cable: stimulate, block channels, change temperature and ions, and watch the spike travel and the channels open.

Presets

Playback

Stimulus

Axon and membrane

Layers

Model check

The science

The Hodgkin-Huxley membrane

The membrane is a capacitor in parallel with three conductances: C dV/dt = -gNa m³h (V - ENa) - gK n⁴ (V - EK) - gL (V - EL) + I. Each gate variable x (m, h, n) obeys dx/dt = αx(V)(1-x) - βx(V)x with the original voltage-dependent rates of Hodgkin and Huxley (1952), measured on the squid giant axon: gNa = 120, gK = 36, gL = 0.3 mS/cm², C = 1 µF/cm², resting potential -65 mV. Sodium activates fast (m) and then inactivates (h); potassium activates slowly (n), which ends the spike and causes the undershoot.

All-or-none and refractoriness

A small stimulus decays back to rest; one that pushes V above the threshold turns on the regenerative sodium current and produces a spike of nearly fixed size and shape (about +40 mV). For a few milliseconds afterwards h is low and n is high, so the membrane is refractory: use the pulse-train preset to see the second spike fail or get smaller when the interval is short. Steady current above the rheobase makes the membrane fire repetitively.

Propagation along the axon

The axon is a cable of 200 compartments: C ∂V/∂t = (a / 2Ri) ∂²V/∂x² - Iion + Istim, with radius a = d/2 and axial resistivity Ri = 35.4 Ωcm. The ionic part is advanced with Rush-Larsen gate updates and the diffusion part with an implicit (backward Euler) tridiagonal solve. The speed of the travelling spike grows as sqrt(d); temperature multiplies all rates by Q10(T-6.3)/10 with Q10 = 3, so a warm axon fires faster and narrower. Reversal potentials come from the Nernst equation E = (RT/F) ln([out]/[in]): change the outside K+ or Na+ concentration and watch the resting potential and spike height shift.

Drugs and myelin

Tetrodotoxin (TTX) blocks Na channels and removes the spike; tetraethylammonium (TEA) blocks K channels and the spike becomes a long plateau. In the myelinated option the internodes have 50 times less capacitance and leak (many membrane layers in series) and no active channels, and active nodes are spaced every 0.5 cm; the spike jumps from node to node (saltatory conduction) and is faster and cheaper than in a bare axon of the same diameter.

Model checks (run in the page)

1. At -65 mV the steady-state gates must equal the textbook values (m 0.0529, h 0.5961, n 0.3177) and the resting potential must come out at -65 mV. 2. A weak pulse must give no spike while two very different strong pulses give spikes of nearly equal height (all-or-none). 3. The conduction speed of a 476 µm squid axon at 18.5 °C must be close to the 18.8 m/s Hodgkin and Huxley computed (21.2 m/s measured) and scale as the square root of the diameter.

Honest simplifications

This is the classic squid-axon model; mammalian neurons have many more channel types, geometry and synapses. The spatial grid (0.5 mm compartments) is far coarser than a real node of Ranvier, so the myelinated mode is illustrative of saltatory conduction, not a model of a real nerve fibre. The injected current is spread over an 8-compartment patch. Reversal potentials follow the Nernst equation but the intracellular concentrations are fixed, there is no ion pumping and no extracellular accumulation. The drawing is a schematic.

Membrane potential at three electrodes
Gating variables m, h, n
Membrane currents at electrode 1
Phase plane: V against n