The Resting and Action Potential: Membrane Charge, Threshold and Firing

01Resting membrane potential

A resting neuron holds a charge it pays for every second

A living nerve cell holds its interior at roughly minus seventy millivolts relative to the fluid outside it. That difference is the resting membrane potential. Sodium and potassium sit unevenly across the membrane, concentrated in opposite proportions in the extracellular fluid and the cytosol. Nothing about that arrangement holds itself. Both ions leak down their gradients continuously.

The sodium-potassium pump restores them. It is an electrogenic pump, exchanging three sodium ions out for two potassium ions in, so every cycle exports one net positive charge. Measured directly in inside-out membrane vesicles prepared from human red blood cells, that stoichiometry generated an ATP-dependent electrical potential of about 1 mV, inside positive Polvani 1989. The coupling ratio is not fixed. In the same vesicles the potential reversed to inside-negative once cytoplasmic sodium fell to 0.4 mM or below. That reversal fits a coupling ratio dropping toward one sodium for two potassium at 0.2 mM, so the pump's electrical contribution depends on the ionic state it works in.

What the charge costs

Holding the resting potential is a standing item on the neuron's energy budget. The first quantitative budget for gray matter was computed from anatomic and physiologic data on rodent brain. It assigned 13 percent of signaling energy to the resting potential and 47 percent to action potentials, with postsynaptic effects of glutamate at 34 percent and glutamate recycling at 3 percent Attwell 2001. In the same calculation, one extra action potential per cortical neuron per second raises oxygen consumption by 145 mL per 100 g of gray matter per hour. Charge is a purchase, and firing rate sets the bill.

The revised budget moved the numbers and made them regional. Recalculated in a simple model of cerebral cortex, 50 percent of cortical signaling energy goes to postsynaptic glutamate receptors, 21 percent to action potentials and 20 percent to resting potentials. In a detailed model of cerebellar cortex, 54 percent goes to maintaining resting potentials and 17 percent to action potentials Howarth 2012. One nervous system pays two different prices to stay charged, and local circuit architecture sets which price applies.

The Unified Model of Tone takes that as the first meaning of readiness. A cell holding charge is spending fuel to keep the freedom to act. Readiness is neither free nor passive, and the amount a region spends on it is a property of that region rather than a constant of biology.

What happens when the fuel stops

The pump falters first. In rat hippocampal slices, hypoxia alone reduced ATP to about 50 percent, and most neurons depolarized slowly and lost their membrane potential within 10 minutes Berg-Johnsen 1995. Hypoxia combined with glucose deprivation depleted ATP outright and produced fast anoxic depolarization followed by glutamate release. Ten minutes without oxygen and the charge is gone in that preparation. A neuron's electrical readiness and its energy supply are one system measured two ways.

02Findings

What the research shows

Measured values for membrane charge, threshold and conduction.

03Ion equilibrium potentials

The resting potential sits suspended between the equilibrium points of several ions

The resting potential is a weighted compromise among ions rather than the property of any one of them. Each ion has a voltage it would settle at on its own. The Nernst equation gives that equilibrium potential for a single ion. The Goldman equation extends it to the several ions crossing the membrane at once, weighting each by how easily it crosses. Together they explain why the resting potential rests where it does, suspended between the equilibrium points of potassium, sodium and chloride.

Hodgkin and Katz built that framework by changing the fluid around an isolated squid giant axon. Lowering external sodium reduced the height of the spike, which established that sodium entry sets the amplitude of the action potential Hodgkin 1949. The membrane voltage is a readout of the fluid on both sides of it.

Ion channels set the terms of that balance. Passive non-gated channels establish the resting potential and pumps maintain it. Voltage-gated channels wait, closed, for the event to come. The resting state is a balance of leak and pump held just beneath the threshold of firing.

The brain defends its own ionic environment

Brain extracellular fluid is not a passive filtrate of blood. Resting brain tissue holds extracellular potassium lower than the rest of the body and extracellular H+ higher Somjen 2002. High extracellular potassium and low extracellular H+ both raise neuronal excitability, so defending those values is how the tissue holds cerebral function steady. Choroid plexus, endothelium and astrocytes do the defending.

Firing disturbs the very values that make firing possible. Intense activity drives extracellular potassium up while extracellular sodium and calcium fall. When the regulation fails, positive feedback tips the tissue into seizure, spreading depression or anoxic depolarization Somjen 2002. The membrane's charge is never a private local fact.

Serum potassium moves the membrane in living people

Kidney function is nerve function. In nine patients with chronic renal failure, axonal excitability parameters correlated with serum potassium across a range of 4.3 to 6.1 mM Kiernan 2002. They did not correlate with urea or creatinine. Patients whose axonal resting potentials were normal had normal serum potassium despite elevated markers of renal dysfunction, and hemodialysis reduced the abnormalities.

A swing of under 2 mM in the blood therefore shifts the resting potential of peripheral axons by a measurable amount. In this model, membrane voltage is a body-wide value read at one site. Blood chemistry, kidney clearance and breathing all reach it.

04The axon initial segment

The action potential starts at one short stretch of axon carrying fifty times the sodium channels

Graded potentials sum at the trigger zone in the initial segment of the axon. When their depolarization reaches threshold voltage, voltage-gated sodium channels spring open, sodium floods inward, and an action potential is generated. That segment holds the job because of density. Antibody staining, whole-cell voltage clamp, sodium imaging and modeling put sodium channel density at the axon initial segment of rat cortical pyramidal neurons at roughly 50 times the density in proximal dendrites Kole 2008. Computational models of the segment required about 2,500 pS per square micrometer there to reproduce what the recordings showed.

That measurement settled a live dispute. Earlier patch-clamp work had suggested the initial segment carried no more sodium channel density than the cell body. It carries fifty times the dendritic density, anchored in place by the actin cytoskeleton. A trigger zone is a trigger zone because of what is packed into it.

All or none, and what threshold actually means

The spike obeys the all-or-none law. Once threshold is crossed, the response is full and uniform, never partial. Threshold itself is a trade between stimulus strength and stimulus duration rather than a single voltage. Single-fiber recordings from the median giant fiber of the common earthworm give a mean rheobase of 1.01 V and a mean chronaxie of 0.06 ms, with conduction at 30.2 m/s Bähring 2014. A weak stimulus held long enough fires the fiber. A stronger one delivered briefly may fail.

Blood chemistry tunes the action potential

Extracellular potassium sets the resting level, sodium sets the amplitude of the spike, and calcium governs sodium conductance. Frankenhaeuser and Hodgkin established the calcium effect in isolated squid axons, where external calcium shifts the voltage dependence of sodium activation Frankenhaeuser 1957. Raise external calcium and the gate opens later. Lower it and the membrane fires to less.

A patient carrying the Nav1.4 I693T mutation showed what that means in a body. Low calcium and magnesium shifted channel activation in the hyperpolarizing direction, producing myotonia that progressed to outright membrane inexcitability, and magnesium reversed it Mankodi 2015. This is one case report supported by computer simulation. It illustrates the mechanism precisely: a mineral level in the blood decided whether that membrane fired too easily or stopped firing at all.

Hodgkin and Huxley put numbers on the whole sequence

Alan Hodgkin and Andrew Huxley separated the sodium and potassium currents that build the spike in the giant axon of the squid Loligo Hodgkin 1952. They then wrote the quantitative description of membrane current that is still taught as the conductance model of the nerve impulse Hodgkin 1952. The equations arrived in 1952 and have not needed replacing.

A spike costs less than the early budgets assumed

The two currents overlap less than the textbook picture implies. Direct measurement in nonmyelinated mossy fibers of the rat hippocampus put total sodium influx per action potential at 1.3 times the theoretical minimum Alle 2009. Earlier energy budget calculations had used a factor of 4. Sodium and potassium channel kinetics are matched so that the two currents barely overlap. The field revised its own number downward.

Efficiency is not uniform across cell types. In mouse central neurons recorded at 37 degrees C, sodium entry in cortical pyramidal cells is nearly confined to the rising phase of the spike. Only about 25 percent more sodium enters than the theoretical minimum. Fast-spiking GABAergic neurons, including cerebellar Purkinje cells and cortical interneurons, admit twice the minimum Carter 2009. The difference comes from the shape of the action potential rather than from cell-specific channel kinetics.

The all-or-none law governs the amplitude of the spike and says nothing about its cost. In this model the cost is the limit. A cell that pays twice as much per spike cannot hold a high rate for as long, so metabolic supply decides which parts of a circuit can stay loud.

05Refractory and conduction

After firing a neuron cannot fire again, and myelin decides how fast the spike travels

Immediately after an action potential, sodium channel inactivation reduces axon excitability. The absolute refractory period is a window of total inexcitability, during which no input, however strong, raises a second spike. A relative refractory period follows, during which a stronger than usual stimulus will Boërio 2004. These windows cap how fast a neuron can signal and keep traffic moving in one direction down the axon.

Refractoriness is measurable in living people using paired-pulse and collision techniques. It varies with nerve type and with the subject, and it lengthens in alcoholic, diabetic and toxic neuropathies as well as in Guillain-Barre syndrome, carpal tunnel syndrome and multiple sclerosis Boërio 2004. The recovery window is a clinical readout of the state of the axon rather than an abstract property of membranes.

Saltatory conduction was demonstrated in single myelinated fibers

Where oligodendrocytes in the central nervous system and Schwann cells in the periphery wrap the axon, the action potential leaps between gaps in the sheath. Those gaps are the nodes of Ranvier, and the leap is saltatory conduction. Huxley and Stampfli demonstrated that leap directly in isolated single myelinated peripheral fibers instead of inferring it from conduction speed Huxley 1949.

The speed difference is more than an order of magnitude. Adult myelinated tibial nerve in the rabbit conducts at about 43 m/s Simpson 2013. Human skin sympathetic fibers, unmyelinated and recorded directly by microneurography in the tibial nerve, conduct at 0.93 plus or minus 0.09 m/s Kondo 2004. Insulation is what separates those two numbers, and the traffic carrying background autonomic tone runs slow.

Insulation alone does not set conduction speed

Lengthening the internode does not by itself speed the nerve. A rabbit limb-lengthening model raised internodal length by 33 percent, from 0.95 to 1.3 mm, and conduction velocity did not change, measured in vivo and in isolated tibial nerves Simpson 2013. Axon diameter, myelin thickness and g-ratio held constant, and theoretical analysis puts the plateau above a nodal separation near 0.6 mm.

The speed of an action potential is therefore set by the whole geometry of the fiber. The model reads that null result as the ordinary case rather than an oddity. A third more internode bought no milliseconds because the fiber was already past the length at which internode is the limiting term. That is why the same manipulation changes one nerve and leaves another where it was.

Myelin buys speed, and something pays for it

Saltatory conduction covers distance at a fraction of the metabolic cost the bare axon would pay. That saving is real at the axon and it does not extend to the tissue. In a signaling budget derived for rodent optic nerve, white matter synapses consume 0.5 percent or less of the energy of gray matter synapses. The cost of building myelin is repaid within months by the reduced ATP cost of action potentials. The cost of maintaining the oligodendrocyte's own resting potential usually outweighs the saving Harris 2012. A glial cell holds a charge so the axon can fire cheaply.

06Where threshold is set

Threshold is a moving value set by body chemistry and by channel biophysics together

Breathing changes nerve threshold. Twenty minutes of hyperventilation in six human subjects lowered the threshold current required to produce sensory and motor potentials of constant size by 20 to 30 percent Mogyoros 1997. All six reported paresthesias in the hand and face. Four developed muscle twitching and cramps. Nothing was applied to the nerve. The subjects changed how they breathed.

Ischemia lowered threshold in the same six subjects as well, by 15 to 20 percent, and it did so through different machinery. Hyperventilation produced none of the refractoriness, supernormality or latency changes that ischemia caused, acting instead on the persistent sodium channels that operate near resting potential Mogyoros 1997. Two inputs raise excitability by two routes. The model predicts that split, because output alone never identifies the input that produced it.

One amino acid moves the threshold

The erythromelalgia mutation L858H in human Nav1.7 amplifies net sodium influx during subthreshold depolarizations 27-fold. That figure comes from dynamic clamp, which imposes the mutant conductance at physiological levels on small dorsal root ganglion neurons in an isolated preparation Vasylyev 2014. Current threshold falls, and it varies linearly with Nav1.7 conductance. Pain can begin as a threshold problem in an intact nerve rather than as a report of damaged tissue.

The trigger zone is not a passive detector

The axon initial segment regulates the integration of synaptic inputs, intrinsic excitability and transmitter release, and it shifts in disease Kole 2012. It behaves as a dynamic signal processing unit rather than a fixed threshold detector. Where threshold sits at that segment is a state the cell holds, set by traffic the cell received earlier.

Threshold sets how much of an arriving input survives to output, which is where gain has its cellular substrate. A human axon whose threshold current has fallen 20 to 30 percent answers stimuli it previously ignored, and nothing about those stimuli changed. The change is in the receiver. Measured across a whole reflex loop rather than one axon, the same relation is reported as a ratio of output to input, and gain is where the library keeps that reading.

07Spike summation

A spike is the smallest unit in which load on the spine reaches a membrane

A single action potential means little on its own. Summed across thousands of synapses, spikes are the entries in the running account that functional neurology calls the central integrative state. The Neuron and the Central Integrative State defines and measures that account. This page keeps the currency rather than the account.

Somatic input enters that account as spikes from named receptors. Single-unit recordings in ten anesthetized adult cats showed paraspinal muscle spindle discharge rising more to the impulse than to the preload in 10 of 16 manipulative loads Pickar 2001. After 7 of those 10, the spindles fell silent for 1.3 plus or minus 0.6 seconds, with a range of 0.1 to 4.3 seconds. Six of the 16 loads unloaded the spindles instead.

That mixed result is the prediction rather than a defect in the data. One load profile met afferents sitting at different resting discharge rates and drove their firing in opposite directions. The recording covered five muscle spindles, four Golgi tendon organs and one presumed Pacinian corpuscle, so it establishes the mechanism and not an effect size. Mechanical load applied to the spine reached individual afferents and changed what they sent.

A five millisecond input changes forty seconds of spindle firing

Nine mechanically assisted thrusts reached peak forces of 68 to 122 N, each lasting under 5 ms. They raised muscle spindle discharge during the thrust and, in some afferents, decreased it for more than 40 seconds afterward Reed 2015. Those spindle signals are action potentials, and they kept arriving long after the mechanical event ended. The work was a feasibility pilot in one cat with three isolated L6 afferents.

A brief mechanical event reset an ongoing discharge rate instead of merely evoking a burst. The force numbers show that size was not what did it, since greater peak force did not reliably buy more spikes. Between 68 and 122 N, what changed the afterdischarge was not the newtons.

Why the membrane is this practice's subject

We are the practice organized around the nervous system, and a charged membrane is where that organization begins. Mechanical load at a spinal segment arrives at that membrane as spikes from spindles, tendon organs and joint receptors, and those spikes enter the same summation that decides whether the trigger zone reaches threshold. In the Unified Model of Tone, the mechanical state a segment is holding and the threshold state of the neurons it feeds are one variable read at two depths. The tension a receptor sits under sets the rate at which it reports. What an input matched to that segment changes is the threshold current, not the anatomy.

The claim is testable in millivolts and in milliamps. Serum potassium reaches this voltage, arterial carbon dioxide reaches it, and mechanical load at a segment reaches it. If threshold current at a peripheral axon answered those three as unrelated inputs, each moving it without regard to the others, the coupling this page rests on would not survive the measurement. Tone carries the library's general form of that commitment. Here it is narrow: one voltage, reachable from the kidney, from the lungs and from the spine.

Membrane voltage inside its healthy range is health, because a cell holding that range keeps the freedom to answer input in either direction. A membrane held too far depolarized fires to noise and exhausts its own fuel. A membrane held too far hyperpolarized stops answering signal that matters. Both are illness, counted in the same currency.

08Tone

How this system expresses tone

All nine foundations are present wherever tone is. Three of them are read directly in millivolts.

Six more foundations are legible in the same millivolts. Oscillation: the refractory period caps how often a spike can repeat, so an axon reports in frequency rather than in amplitude. Prediction: subthreshold traffic moves the trigger zone toward or away from firing before the input that matters arrives. Constraint: during the absolute refractory period no stimulus of any size produces a second spike. Input quality: an earthworm giant fiber answers a 1.01 V stimulus or fails to, depending on how long it is held, which is why duration is part of the signal. Coupling: serum potassium, arterial oxygen and breathing rate all arrive at the same membrane voltage. Time course: a thrust lasting under 5 ms altered spindle discharge for more than 40 seconds.

09Across the library

How this page relates to the rest of the library

Where the charged membrane is read across the rest of the library.

10Frequently asked

Questions about this topic

11The sources

References

Sources: primary literature, linked inline.