Chapter 13 The biophysics of the neuron
12 minute read
Inside the human skull, the folded neural tube of vertebrate development blossoms into roughly 86 billion neurons and an equal number of glial cells. More than 100 trillion microscopic junctions connect these cells into a parallel electrochemical loom running on just 20 watts of power. With less energy than a dim porch bulb uses, this wet tissue sees light, remembers childhood, and composes music.
To understand how soft biological tissue carries rapid thoughts, consider the physical medium it must work with. Copper wires carry messages through a computer because free electrons race through dry metal at nearly the speed of light. A living cell has no metal wires and no dry insulation. It is a warm bag of salty water suspended in a salty bath, built from soft lipids and folded proteins. How can a fragile droplet of brine carry crisp electrical signals from your spinal cord to your toe?
13.1Ion channels and the resting membrane potential
Every neuron begins by building a wall between its inner cytoplasm and the surrounding extracellular fluid. That wall is the cell membrane, a double layer of fatty phospholipid molecules only five nanometers thick. Because oil and water do not mix, dissolved salt ions cannot easily cross this greasy film. In electrical terms, the lipid bilayer acts as a tiny capacitor with a specific capacitance near one microfarad per square centimeter. It holds positive and negative charges apart across an atomic gap, storing electrostatic energy at the boundary of the cell.
If the membrane were only an unbroken film of fat, no current would ever flow. Life turns this insulator into an active battery by studding the lipid wall with two families of protein machines. The first family consists of active ion pumps, most notably the sodium-potassium pump discovered by Jens Christian Skou. This enzyme burns roughly half of all the energy molecules that your brain produces every day. Each time a pump splits one molecule of ATP, it pushes three sodium ions out and pulls two potassium ions in.
Through billions of these strokes every second, the neuron turns itself into a charged chemical reservoir. Potassium becomes nearly 30 times more concentrated inside the cell than outside, reaching 140 millimolar within against 5 millimolar outside. Sodium shows the opposite pattern, staying scarce inside at 12 millimolar while crowding the outside bath at 145 millimolar. Chloride ions also stay mostly outside, while intracellular calcium is swept down to a ten-thousandth of a millimolar. Even a tiny influx of calcium can therefore act as a sharp chemical trigger inside the cell.
The second family of membrane proteins consists of selective ion channels. These narrow protein tunnels allow specific ions to coast passively down the steep chemical gradients built by the pumps. For decades, biophysicists wondered how a potassium channel could let large potassium ions pass while blocking smaller sodium ions. In 1998, Roderick MacKinnon solved this riddle by mapping the atomic crystal structure of the potassium pore.
In water, every ion wears a tight cloak of water molecules held by electrostatic attraction. To enter the narrowest neck of the channel, an ion must shed this watery coat. Inside the selectivity filter of the channel, rings of carbonyl oxygen atoms stand at a precise spacing. They embrace a bare potassium ion and mimic its lost water shell at zero net energy cost. A smaller sodium ion fits too loosely inside this rigid oxygen cage. It therefore stays outside in the water rather than pay the energy price of shedding its hydration shell.
Once selective channels open, two competing physical forces act on each species of ion. Thermal motion pushes ions to diffuse from high concentration toward low concentration. At the same time, any voltage difference across the membrane pulls positive ions toward the negative side. In the late 19th century, Walther Nernst and Max Planck combined these two tendencies into a single law of electrodiffusion. Suppose a membrane is permeable only to potassium. As potassium ions diffuse out down their concentration gradient, they leave unbalanced negative charges behind inside the cell. Very quickly, the interior becomes electrically negative relative to the outside bath. This growing negative voltage pulls positive potassium ions back in, opposing their outward diffusive escape.
Eventually, the inward electrical pull balances the outward diffusive push, and the net molar flux drops to zero. At that point of equilibrium, the voltage across the membrane equals the Nernst potential for that ion.
At human body temperature, the thermal voltage factor equals about 26.7 millivolts. Plugging in the physiological ion concentrations reveals four competing tug-of-war targets across the membrane. Potassium pulls the voltage toward minus 89 millivolts, and chloride pulls toward minus 75 millivolts. Sodium pulls strongly in the opposite direction toward plus 67 millivolts, while calcium pulls toward plus 128 millivolts.
A real resting neuron is never permeable to just one ion at a time. Instead, potassium, sodium, and chloride all leak across the membrane simultaneously according to their respective permeabilities. David Goldman, Alan Hodgkin, and Bernard Katz solved for the steady voltage where the total electric current across all three ions vanishes. Their formula is the Goldman-Hodgkin-Katz voltage equation.
In a quiet neuron, leak potassium channels stand open while almost all sodium channels remain tightly shut. This gives a resting permeability ratio of potassium to sodium of roughly 25 to 1. Because potassium dominates this tug-of-war, the interior of a quiet neuron rests near minus 70 millivolts. Across a five-nanometer membrane, that modest voltage creates a fierce electric field of 14 million volts per meter. Every resting nerve cell is a cocked electrostatic spring waiting for a trigger.
13.2The Hodgkin and Huxley action potential
Maintaining a resting voltage inside a cell body is only half of the communication problem. A motor neuron in your spine must send commands down a threadlike axon that stretches a full meter to your foot. What happens if a neuron simply injects a passive puff of voltage into one end of this biological cable?
Lord Kelvin worked out the physics of leaky underwater telegraph cables in the 1850s, and the exact same equation governs a passive nerve fiber. The salty core resists longitudinal current flow, while current leaks out across the membrane resistance and charges the membrane capacitance. In an unmyelinated biological axon, the core is so resistive and the membrane so leaky that its characteristic space constant is only one millimeter. A passive voltage pulse decays exponentially with distance and vanishes within a few millimeters. If nerves relied on passive conduction, a command from your brain would fade out before leaving your skull.
Evolution solved this attenuation crisis by turning the axon from a passive wire into a line of dominoes. Instead of letting a signal fade over distance, the membrane regenerates the electrical pulse at every point along the fiber. This self-regenerating solitary wave is the action potential, or neural spike.
How the membrane produces this spike remained a mystery until the early 1950s. Working in Plymouth, England, Alan Hodgkin and Andrew Huxley studied the giant axon of the squid Loligo. Nearly a millimeter thick, this giant nerve fiber let them thread fine silver wires right down the inside of the cell. Using an electronic feedback circuit called the voltage clamp, they held the membrane voltage fixed and measured the ionic currents directly.
Hodgkin and Huxley discovered voltage-gated ion channels whose permeability shifts dramatically whenever the membrane voltage changes. By separating the inward sodium current from the outward potassium current, they deduced how charged protein gates control each pore. The potassium channel behaves as if four identical activation gates must open together before ions can cross. If the variable represents the probability that a single potassium gate is open, the potassium conductance scales as the fourth power of that probability.
The sodium channel is even cleverer because it combines two opposing gates in one molecule. Three fast activation gates swing open when the cell depolarizes, while one slower inactivation gate plugs the inner pore shortly after it opens. Coupling these gating probabilities to spatial current flow along the axon radius and intracellular resistivity yields the Hodgkin-Huxley membrane equation.
Each of the three gating probabilities relaxes toward its voltage-dependent steady state with its own voltage-dependent gate time constant, distinct from the passive membrane time constant. Because Cambridge University's electronic computer was down for repairs, Huxley integrated this four-dimensional nonlinear system for weeks on a hand-cranked calculator. The equations reproduced the exact shape, threshold, and propagation speed of the living squid spike.
The secret of the action potential lies in a clean separation of timescales between the gates. At any voltage, the sodium activation gate responds in about a tenth of a millisecond. That is roughly 10 times faster than sodium inactivation or potassium activation. When an incoming stimulus nudges the voltage past a threshold near minus 55 millivolts, the fast sodium activation gates snap open first.
Once a few sodium channels open, positive sodium ions rush into the cell. That inward rush of positive charge raises the internal voltage further, which opens still more sodium channels. This runaway positive feedback loop ignites an explosive upstroke, shooting the voltage from minus 70 millivolts to plus 40 millivolts in a millisecond.
Just as the spike reaches its peak, the two slower negative feedback gates catch up. Over one to two milliseconds, the sodium inactivation gates swing shut and choke off the inward sodium flood. At the same time, the potassium gates swing wide open, unleashing an outward surge of potassium ions. This outward current drags the voltage rapidly back down past its resting level to around minus 80 millivolts.
This delayed reset explains why nerve impulses travel strictly forward and never echo backward. In the wake of a passing spike, the sodium inactivation gates remain plugged and the potassium channels stay open for several milliseconds. During this refractory period, the membrane behind the wavefront cannot fire again, forcing the pulse to march strictly forward.
In vertebrates, glial cells wrap axons in dozens of tight insulating lipid layers called the myelin sheath. This wrapping cuts the transverse capacitance and leak conductance of the cable a hundredfold, leaving bare micrometer gaps called nodes of Ranvier every millimeter. Clustered at these bare nodes, sodium channels let the electrical pulse leap from node to node at 120 meters per second.
13.3Synapses and rhythms of the cortex
When a racing action potential reaches the end of an axon, its electrical journey hits a sudden chasm. Separating the sending axon terminal from the next dendrite is a water-filled gap 20 nanometers wide called the synaptic cleft. Because the tiny current of an axon terminal cannot jump this watery moat directly, the neuron converts electricity into chemistry.
As the voltage spike invades the presynaptic terminal, it opens voltage-gated calcium channels in the active zone. Extracellular calcium ions pour rapidly into the terminal. Waiting just inside the membrane are tiny lipid bubbles called synaptic vesicles, each packed with neurotransmitter molecules. Each vesicle is tethered to the cell membrane by coiled protein strands known as the SNARE complex.
James Rothman, Randy Schekman, and Thomas Südhof revealed how this molecular release machinery works. When four or five calcium ions bind to a sensor protein called synaptotagmin, the SNARE strands zipper tightly together. Within 200 microseconds, this molecular winch fuses the vesicle with the outer membrane and spills neurotransmitter across the cleft.
Across the narrow gap, neurotransmitter molecules bind to receptor proteins waiting on the receiving dendrite. The main excitatory messenger in the vertebrate brain is glutamate. When glutamate binds to AMPA receptors, sodium ions rush in and nudge the receiving cell toward its firing threshold. This brief depolarizing wave is an excitatory postsynaptic potential.
The main inhibitory messenger in the brain is gamma-aminobutyric acid, or GABA. When GABA binds to its receptors, it opens chloride channels that clamp the voltage near minus 75 millivolts and dampen excitation. Alongside these fast messengers, slower neuromodulators such as dopamine, serotonin, norepinephrine, and acetylcholine bathe wide brain territories. They tune the learning rate of millions of synapses at once to signal reward, vigilance, or surprise.
How do these chemical junctions record our memories as we learn? In 1949, Donald Hebb proposed a simple rule for how experience rewires the brain. When cell A repeatedly takes part in firing cell B, some metabolic change strengthens the synapse from A to B. In short, neurons that fire together wire together.
In 1973, Tim Bliss and Terje Lømo discovered this synaptic strengthening in the rabbit hippocampus, calling it long-term potentiation. Neurobiologists soon traced this effect to a remarkable protein sitting on the receiving dendritic spine, the NMDA glutamate receptor. Unlike an AMPA receptor, which opens whenever glutamate touches it, the NMDA receptor is a molecular coincidence detector. It opens only when the sending cell and the receiving cell fire in the right causal sequence.
At the resting voltage of minus 70 millivolts, even when glutamate binds to an NMDA receptor, no ions pass through its tunnel. The pore is physically plugged from the outside by a magnesium ion held in place by the cell's negative interior. If the sending neuron fires alone while the receiving neuron stays silent, the magnesium plug stays wedged in place.
Suppose instead that just after the sending cell releases glutamate, the receiving cell fires an action potential of its own. That spike sweeps backward into the dendrite, raising the voltage inside the spine toward minus 20 millivolts. This positive shift electrostatically repels the positive magnesium ion and pops it out of the pore. With glutamate bound and the magnesium plug expelled, calcium rushes through the open NMDA channel into the dendritic spine.
This sharp calcium surge activates an enzyme called CaMKII, which inserts extra AMPA receptors into the synapse and enlarges the spine. The junction becomes permanently more sensitive to future messages. Conversely, if the receiving cell fires slightly before the sending cell arrives, only a weak trickle of calcium leaks in. That weak trickle activates phosphatase enzymes that strip AMPA receptors away and weaken the synapse.
In the late 1990s, Henry Markram, Guo-qiang Bi, and Mu-ming Poo measured this timing window to the millisecond. For a spike time difference , the change in synaptic weight obeys spike-timing-dependent plasticity.
When the presynaptic cell fires just before the postsynaptic cell, so that the time difference is positive, the synapse strengthens. When the postsynaptic cell fires first, so that the time difference is negative, the synapse weakens. Thanks to a single magnesium ion inside the NMDA pore, synapses do not merely record correlation. They wire themselves to capture directed temporal causality, learning which events in the world cause which outcomes.
When billions of these causal synapses link together across the mammalian neocortex, collective rhythms emerge. About 80% of cortical neurons are excitatory pyramidal cells, while 20% are fast-spiking inhibitory interneurons. In 1972, Hugh Wilson and Jack Cowan modeled the coupled dynamics of these excitatory and inhibitory neural populations. Whenever excitatory pyramidal cells fire a volley of spikes, they excite local inhibitory interneurons. Those interneurons immediately fire back a blanket of GABA inhibition that silences the pyramidal cells for about 20 milliseconds. As the inhibition wears off, the pyramidal cells fire again. Every cortical column thus acts as a pacemaker that hums in the gamma band between 30 and 80 hertz. Nested inside slower theta and beta rhythms, these oscillations coordinate cortical computation across the brain.