Myelination and the Insulation of Axons

An axon is, electrically, a leaky cable: current injected at one point decays with distance because some of it leaks across the membrane before it can spread further. The distance over which membrane potential falls to 1/e of its initial value is the length constant, λ, and the time to reach 63% of a step change is the time constant, τ. Both depend on the axon's resistive and capacitive properties. Myelin is a substance — really a specialized region of glial plasma membrane, extraordinarily lipid-rich and compacted — that a supporting glial cell wraps concentrically around a segment of axon. This wrapping does two things simultaneously: it raises the effective resistance across the membrane (fewer paths for current to leak out) and it lowers the effective capacitance (multiple stacked membranes in series reduce total capacitance, much as stacking capacitor plates in series reduces net capacitance). Increasing λ and decreasing τ both make passive current spread faster and farther — exactly the changes that speed conduction. The sheath is not continuous along the whole axon. It is deposited in discrete stretches called internodes, each produced by one glial cell (one oligodendrocyte process in the central nervous system, one Schwann cell in the peripheral nervous system), separated by short, unmyelinated gaps called nodes of Ranvier. Nodes are where the axon membrane is exposed to extracellular fluid and where voltage-gated Na⁺ channels are densely clustered — on the order of 1,000–2,000 channels per μm², roughly 25–100 times the density found in the internodal membrane or in unmyelinated axons. Because the internode's high resistance and low capacitance let depolarization spread passively and nearly loss-free to the next node, and because that node is stocked with enough Na⁺ channels to regenerate a full action potential, the net effect is that action potentials appear to jump from node to node. This node-to-node hopping is called saltatory conduction (from Latin saltare, to leap), treated in full detail as the direct consequence of the architecture described here. The key terms to fix precisely: compact myelin (the tightly wrapped, cytoplasm-poor bilayers that give the sheath its insulating properties), the internode (myelinated segment between two nodes), the paranode (the tapering region at each end of an internode where myelin loops attach to the axolemma), the juxtaparanode (just internal to the paranode, rich in K⁺ channels), and the node of Ranvier itself (the bare, Na⁺-channel-dense gap). Each domain has a distinct molecular composition, and disrupting any one of them changes conduction — evidence, discussed below, that the domains are functionally as well as structurally distinct.

Step 1 — Glial wrapping and compaction. In the peripheral nervous system, a single Schwann cell first ensheathes a segment of one axon (myelinating Schwann cells associate with only one axon each, unlike non-myelinating Schwann cells which bundle several small unmyelinated axons together in Remak bundles). The Schwann cell's plasma membrane then spirals around the axon many times — commonly 10 to 160 wraps depending on final axon diameter — while cytoplasm is progressively squeezed out from between the wraps. In the central nervous system, one oligodendrocyte extends multiple processes, each of which wraps a different axon segment (a single oligodendrocyte can myelinate 20–60 internodes on different axons simultaneously, unlike the one-cell-one-internode rule of Schwann cells). As wrapping proceeds, the cytoplasmic faces of the membrane fuse to exclude cytoplasm, forming the major dense line seen in electron micrographs, while the extracellular faces appose to form the thinner intraperiod line. The result is compact myelin: a stack of fused bilayers with almost no aqueous or cytoplasmic space, ideal for blocking transverse ionic current.

Step 2 — Molecular architecture of compact myelin. Two protein families dominate and stabilize the compacted state. Myelin basic protein (MBP), a peripheral membrane protein on the cytoplasmic face, is required for major dense line formation; mice lacking functional MBP (the classic shiverer mutant) fail to compact CNS myelin properly and shiver continuously due to conduction defects. Proteolipid protein (PLP), an integral membrane protein in the CNS (its PNS structural counterpart role is filled largely by P0, or myelin protein zero), stabilizes the intraperiod line. Myelin is unusually lipid-rich — roughly 70–80% lipid by dry weight versus ~50% for a typical plasma membrane — with cholesterol and glycosphingolipids like galactocerebroside especially enriched, contributing to the tight, low-capacitance packing.

Step 3 — Nodal domain organization. At each node of Ranvier, the axon membrane is uncovered and exposed to the extracellular space (in the PNS, interdigitating Schwann cell microvilli partially cover it; in the CNS, nodes are typically bare or contacted by astrocytic processes). Voltage-gated Na⁺ channels (chiefly Naᵥ1.6 in mature nodes) cluster at extremely high density, anchored by a cytoskeletal scaffold built around ankyrin-G and βIV-spectrin, which link the channels to the actin/spectrin cytoskeleton and to cell-adhesion molecules such as neurofascin-186. Flanking each node are the paranodes, where myelin loops form axo-glial junctions with the axolemma via a complex of Caspr (contactin-associated protein), contactin, and glial neurofascin-155. These junctions act like molecular septate-junction seals, physically restricting the lateral diffusion of nodal Na⁺ channels into the internode and electrically sealing the periaxonal space so that current generated at the node cannot simply leak sideways under the myelin. Just internal to the paranode, in the juxtaparanode, voltage-gated K⁺ channels (Kᵥ1.1/Kᵥ1.2) cluster, partly to dampen re-excitation and stabilize resting potential, held in place by Caspr2 and TAG-1.

Step 4 — Passive spread through the internode. Under the compact myelin of the internode, the axon membrane itself still has its normal, relatively low density of channels, but it barely matters electrically: transverse current cannot easily cross the many stacked bilayers (resistance is high) and the multiple membranes in series drastically cut effective capacitance. Local circuit current generated by an action potential at one node therefore flows longitudinally down the axoplasm, across the low-capacitance, high-resistance internodal membrane region, and depolarizes the next node with very little decrement or delay. This is passive, "electrotonic" spread exactly as described by cable theory, but occurring over internodal distances of roughly 100 times the axon diameter (commonly 0.2–2 mm, scaling with fiber caliber) instead of the sub-millimeter decay length typical of a bare unmyelinated membrane.

Step 5 — Regeneration at the next node. When the passively spread depolarization reaches the next node, its Na⁺ channel density is high enough that even the attenuated signal reliably crosses threshold, triggering a full new action potential. The safety factor — the ratio of available depolarizing current to the current needed to reach threshold — is normally 5-fold or greater in healthy myelinated fibers, providing a wide margin so that even at the biophysical limits of internodal length, conduction does not fail. The action potential therefore appears to jump: node, internode (silent, passive), node, internode, and so on — the saltatory pattern whose full biophysical consequences (velocity scaling, energy cost, and comparison to unmyelinated fibers) are developed as their own topic.

Consequences for cable properties, quantified. Two cable-theory quantities capture why this matters:

• Length constant, λ = √(rₘ / rᵢ), where rₘ is membrane resistance per unit length and rᵢ is internal (axoplasmic) resistance per unit length. Myelin raises rₘ roughly in proportion to the number of wrapped membranes (each layer adding resistance in series), so λ increases — often several-fold — letting depolarization reach farther before decaying.
• Time constant, τ = rₘ · cₘ. Myelin lowers cₘ (membrane capacitance per unit length) because capacitors in series (each myelin lamella is itself a capacitor) combine as 1/C_total = ΣΣ(1/Cᵢ), shrinking total capacitance roughly in inverse proportion to the number of wraps. A smaller τ means the membrane charges and discharges faster, so the local circuit current reaches threshold at the next node sooner.

The joint effect — larger λ, and faster charging via smaller τ — is why conduction velocity in myelinated axons scales roughly linearly with axon diameter (v ≈ 6 × diameter in μm, in m/s, a widely used clinical rule of thumb for large myelinated fibers), whereas unmyelinated axon velocity scales only with the square root of diameter, requiring impractically large diameters to reach comparable speeds.

Electron microscopy established the spiral-wrap structure. Betty Ben Geren (Betty Geren Uzman), working in the 1950s, used electron microscopy of developing peripheral nerves to show that myelin forms by a Schwann cell membrane spiraling repeatedly around the axon, with the mesaxon (the double membrane connecting the Schwann cell surface to the wrap) visibly tracing the path of wrapping at successive developmental stages. This directly overturned earlier, vaguer notions of myelin as a secreted or deposited substance and established it as a specialized, topologically continuous extension of the glial plasma membrane. Subsequent high-resolution EM by Jean-Pierre Rossignol, Cedric Raine, and others resolved the major dense line and intraperiod line, confirming that compaction proceeds by exclusion of cytoplasm from between successive wraps.

Physiological recordings proved that conduction is faster with myelin, and that current concentrates at nodes. Ichiji Tasaki and Taiji Takeuchi, in a series of studies in the 1930s–1950s on frog and toad myelinated nerve fibers, used local extracellular recording electrodes placed at different points along single fibers. They found that the externally recorded action current was large specifically at the nodes of Ranvier and much smaller over the internodes — direct electrophysiological evidence that the internode is nearly electrically silent and that essentially all inward current generating the impulse occurs at nodal gaps. When they blocked or destroyed function at a single node (for instance with local anesthetic or by mechanical/chemical injury restricted to a nodal gap), conduction failed or was drastically slowed even though the internodes on either side were untouched — evidence that nodes, not internodes, are functionally essential trigger zones. Later, intracellular and patch-clamp recordings (from the 1970s onward) confirmed enormously higher Na⁺ channel density at nodes versus internodal or unmyelinated membrane, using both voltage-clamp current measurements and, eventually, saxitoxin/tetrodotoxin binding assays that quantify channel number directly by counting toxin-binding sites per unit membrane area.

Comparative conduction velocity measurements across species and fiber types demonstrated the functional payoff. Direct measurement of conduction velocity in nerve trunks (using paired stimulating and recording electrodes at known separations, timing the latency of the compound action potential) showed that large myelinated mammalian A-fibers conduct at 70–120 m/s, while unmyelinated C-fibers of similar or even greater length conduct at only 0.5–2 m/s. Critically, when conduction velocity is plotted against fiber diameter across many myelinated fibers, the relationship is closely linear (v ≈ 6 × diameter, in m/s per μm, for large mammalian fibers), whereas for unmyelinated fibers of the same species the relationship instead follows a square-root law. This difference in scaling — linear versus square-root — is the quantitative fingerprint of myelination's mechanism and was documented across many species by researchers including Bernard Katz, Alan Hodgkin, and later Stephen Waxman's comparative work on invertebrate giant axons (fast but unmyelinated, achieving speed only through enormous diameter) versus vertebrate myelinated fibers (fast at a fraction of the diameter).

Genetic and knockout evidence pinned down individual molecular components. The shiverer mouse, carrying a large deletion in the myelin basic protein gene, has CNS myelin that fails to compact properly (very few major dense lines form), and these mice show tremor, seizures, and premature death, along with markedly slowed or blocked central conduction — direct genetic proof that MBP is necessary for functional compact myelin. Mice engineered to lack Caspr or neurofascin-155 fail to form stable paranodal junctions; recordings from their axons show current leaking into the internodal space and dramatically slowed conduction velocity despite grossly normal-looking myelin wraps, establishing that the paranodal seal — not merely the presence of wrapped membrane — is required for the full speed benefit. Together, structural (EM), physiological (local recording, voltage clamp), comparative (velocity–diameter scaling), and genetic (knockout) evidence converge on the same mechanistic picture described above.

Example 1 — Estimating conduction velocity from fiber diameter and comparing to an unmyelinated axon of equal cost. A myelinated mammalian motor axon has an outer (axon plus myelin) diameter of 12 μm, with axon diameter roughly 8 μm (myelin thickness scales such that axon diameter is about 0.6–0.7 of total fiber diameter — the "g-ratio," typically 0.6–0.8 for optimal conduction). Using the clinical rule of thumb v ≈ 6 × (axon diameter in μm) m/s for large myelinated fibers: v ≈ 6 × 8 = 48 m/s (order-of-magnitude estimate; real values for 8 μm axons run roughly 40–50 m/s, consistent with measured motor nerve conduction velocities). Now compare an unmyelinated axon built to the same total cross-sectional space (12 μm diameter, no myelin wrap needed). Unmyelinated conduction velocity scales roughly as v ≈ k√d, with k on the order of 1.6–2 m/s per √μm for typical invertebrate axoplasm resistivity: v ≈ 1.8 × √12 ≈ 1.8 × 3.46 ≈ 6.2 m/s. The myelinated fiber is roughly 8-fold faster than an unmyelinated fiber of the same outer diameter. To match 48 m/s with an unmyelinated axon using the square-root law, you would need: 48 = 1.8 × √d → √d ≈ 26.7 → d ≈ 710 μm. That is a diameter roughly 60 times larger than the 12 μm myelinated fiber — a vivid, quantitative demonstration of why myelin, not brute diameter, is the vertebrate solution to fast conduction in a space-limited nervous system (this is exactly the scale of the classic squid giant axon, ~0.5–1 mm, needed for fast escape reflexes without any myelin at all).

Example 2 — Reasoning through a demyelinating lesion's effect on safety factor and clinical conduction block. A patient with an acute inflammatory demyelinating episode (a Guillain–Barré-spectrum process) has lost roughly half the myelin wraps over a set of internodes in a peripheral motor nerve, while the nodes themselves and their Na⁺ channel clusters remain structurally intact early in the disease. Reasoning: halving the number of wrapped bilayers roughly halves the effective membrane resistance contribution attributable to myelin (resistances in series roughly halve when half as many resistive layers are stacked) and roughly doubles the effective capacitance (fewer series capacitors sum to a larger net capacitance). Both changes shrink λ and lengthen τ. The depolarization arriving at the next (still Na⁺-channel-rich) node is therefore smaller in amplitude and slower to rise. Two outcomes are possible depending on severity:

• If the residual safety factor (available depolarizing current ÷ threshold current) stays above 1, the node still fires, but later than normal — this produces measurable slowing of conduction velocity on a nerve conduction study, the classic electrodiagnostic signature of demyelination.
• If demyelination is severe enough that the safety factor drops below 1 at one or more internodes, the depolarization reaching the next node never reaches threshold — this produces conduction block, in which the compound action potential amplitude drops sharply when stimulating proximal to, versus distal to, the lesion, even though the axon distal to the block is entirely intact and excitable to direct local stimulation.

This distinction — slowing versus block, and the fact that the axon remains intact and excitable when stimulated directly beyond the lesion — is precisely how clinical neurophysiologists distinguish demyelinating neuropathies from axonal (axon-loss) neuropathies on a nerve conduction study, a direct clinical application of the cable-theory mechanism above.

Ichiji Tasaki (1910–2009) was a Japanese-American biophysicist whose meticulous single-fiber electrophysiology on frog and toad myelinated nerves, conducted first in Japan in the 1930s and continued after he moved to the United States in the 1950s to the National Institutes of Health, provided the direct physiological proof that action currents concentrate at the nodes of Ranvier. Working with extracellular electrodes small enough to resolve activity at individual nodes along a single fiber — an extraordinary technical feat given the equipment of the era — Tasaki showed that blocking a single node abolished propagation across an otherwise intact internode, and that externally recorded currents were far larger at nodal than internodal sites. His collaborator Taiji Takeuchi extended these node-localized recordings and helped establish the quantitative relationship between nodal spacing and conduction velocity. Tasaki's work, alongside contemporaneous studies by Alan Hodgkin and Bernard Katz on ionic mechanisms of excitability, gave saltatory conduction its firm experimental foundation, transforming it from a plausible structural inference (based on Ranvier's 19th-century anatomical descriptions of the nodes) into a demonstrated physiological mechanism.

Understanding myelin's cable-property mechanism underlies the interpretation of every routine nerve conduction study and electromyogram (EMG) performed in clinical neurology: slowed conduction velocity with preserved amplitude points toward demyelination, while reduced amplitude with preserved velocity points toward axon loss, a distinction that guides diagnosis of conditions from Charcot-Marie-Tooth disease (many subtypes caused directly by mutations in myelin genes such as PMP22 or MPZ/P0) to acquired inflammatory neuropathies. In multiple sclerosis, autoimmune attack on CNS myelin (and, importantly, damage extending to paranodal proteins and even axons themselves) produces the conduction slowing and block that generate clinical symptoms, and the disease's fluctuating course partly reflects transient conduction failure at demyelinated but not-yet-axon-damaged sites — the biological basis for why some MS symptoms are reversible early in the disease. Guillain–Barré syndrome and its variants involve acute autoimmune attack on peripheral myelin or, in some subtypes, directly on nodal/paranodal adhesion molecules, again producing conduction block that can be life-threatening when it affects respiratory motor axons. On the therapeutic frontier, understanding the signals that trigger oligodendrocyte precursor cells to differentiate and remyelinate axons has driven a wave of remyelination-promoting drug development for MS, aiming to restore nodal architecture rather than merely suppress inflammation. In agriculture and comparative physiology, understanding myelin's absence in most invertebrates (with rare convergent exceptions, such as some myelinated shrimp and copepod axons) informs research on why vertebrate reflex arcs can be so much faster relative to body and nerve size — directly relevant to comparative neuroscience and evolutionary biology of nervous system design. Finally, myelin biology intersects with basic neurodevelopmental research, since human CNS myelination continues into the third decade of life and its trajectory is studied via diffusion MRI markers correlated with cognitive maturation.

Ask the tutor: How exactly does the "safety factor" at a node get calculated, and what happens electrophysiologically the moment it drops just below 1?