Microvolt to Nanovolt

By the time electrical activity from neurons reaches your scalp, it has been attenuated enormously. Each neuron fires at roughly 70 millivolts internally, but the skull and cerebrospinal fluid act like a lossy, low-pass filter. Billions of neurons fire asynchronously, and their fields mostly cancel. Only when large populations synchronise — as in alpha waves during relaxed wakefulness — does a coherent signal of 20–100 μV emerge at the scalp. Intracranial electrodes placed directly on the brain surface (electrocorticography) pick up signals 10–100 times larger, in the millivolt range.

The Seebeck effect: when two different metals are joined and the junction is heated, electrons in each metal diffuse at different rates, creating a net voltage. A type-K thermocouple (chromel–alumel) generates about 41 μV per degree Celsius. This means measuring a 0.01°C change requires resolving 0.41 μV — well within the microvolt regime. The effect works because the electron energy distribution in each metal responds differently to temperature, and the voltage is the integral of these differences along the wire.

Not directly, but a good moving-coil phono cartridge outputs about 3–5 mV at its hottest, and the quietest grooves on a vinyl record may produce only 5–20 μV. A phono preamp with 40–60 dB of gain boosts this to line level. The signal-to-noise challenge is real: the thermal noise of the cartridge's coil resistance at room temperature is itself in the microvolt range, which is why audiophiles obsess over low-noise preamp designs. Below about 1 μV, you are essentially trying to hear the random jiggling of electrons.

Electrooculography (EOG) picks up eye-movement potentials of 15–200 μV. Electroretinography (ERG) captures retinal responses as low as 5 μV. But the subtlest commonly measured biosignal is the auditory brainstem response (ABR), used in newborn hearing screening — it is about 0.1–0.5 μV, requiring hundreds of averaged recordings to pull the signal out of background EEG noise. Foetal ECG detected through the mother's abdomen sits at roughly 1–10 μV. Below that, you need implanted electrodes.

Because the noise you are trying to reject is millions of times larger than the signal. Mains hum from power lines induces about 1–10 mV of 50/60 Hz interference on the human body — up to 10,000 times bigger than a 1 μV biosignal. A differential amplifier subtracts the signal at two nearby electrodes, cancelling the common interference while preserving the local signal difference. Common-mode rejection ratios above 100 dB (100,000:1) are standard in medical instrumentation. Without this, every EEG recording would just be a picture of your wall socket's frequency.

A nanovoltmeter — yes, that is a real product category. Keithley (now Tektronix) makes bench instruments that resolve down to about 0.1 nV. They work by using chopper-stabilised amplifiers that mechanically or electronically reverse the input polarity hundreds of times per second, cancelling out the amplifier's own drift. Without this trick, the instrument's internal thermal voltages would swamp the signal. You also need low-thermal-EMF cables and connectors made from tellurium copper, because even touching a regular banana plug generates microvolts of thermoelectric noise.

Your skin is a warm, slightly salty, electrochemically active surface. When it contacts a metal, you get a galvanic potential from sweat ions reacting with the conductor, plus a thermoelectric voltage from the temperature difference between your finger and the ambient metal. These effects easily produce tens of millivolts — about ten million times larger than a nanovolt. This is why nanovolt-level experiments use robotic probe stations or at minimum latex gloves, clean-room protocols, and thermally stabilised enclosures.

SQUIDs (superconducting quantum interference devices) sidestep conventional amplifier noise entirely. They exploit quantum mechanical tunnelling of Cooper pairs across Josephson junctions, which makes them sensitive to magnetic flux changes of a single flux quantum (about 2 × 10⁻¹⁵ weber). The corresponding voltage signals are in the nanovolt range. The superconducting loop screens out thermal noise because it operates at 4 kelvin, where Johnson noise is negligible. Magnetic shielding rooms made of mu-metal block external interference, letting the SQUID resolve brain signals a billion times weaker than Earth's magnetic field.

Individual ion channel openings in cell membranes produce current pulses of a few picoamps, which across the channel's resistance create voltage blips of roughly 1–10 nV. Patch-clamp electrophysiology can detect these, but it measures current, not voltage, so the nanovolt figure is inferred. At the whole-organism level, magnetoencephalography (MEG) detects magnetic fields from brain currents whose equivalent electrical signals at the sensor are in the low nanovolt range. Single-neuron action potentials, by contrast, are millivolts — a million times larger.

Quantum mechanics, specifically Johnson–Nyquist noise. Any resistor at temperature T generates a random voltage noise of √(4kTRΔf), where k is Boltzmann's constant, R is resistance, and Δf is bandwidth. At room temperature with a 1 kΩ source and 1 Hz bandwidth, this is about 4 nV. You can beat this floor by cooling the source to cryogenic temperatures, narrowing the measurement bandwidth, or using quantum-limited amplifiers like SQUIDs or parametric amplifiers. At absolute zero the thermal noise vanishes, but quantum zero-point fluctuations remain — a truly fundamental limit.