The Spark: From Galvani’s Frogs to the First Human EEG

Imagine a rain-lashed night in 1780s Bologna: Luigi Galvani, a middle-aged anatomy professor at the University of Bologna, hunches over a dissecting table in his dimly lit lab, a half-skinned frog splayed out in front of him. His original goal is mundane: to study the anatomy of the frog’s sciatic nerve and the muscle fibers of its thigh. He’s just touched the exposed nerve with a metal scalpel when, without warning, the frog’s leg twitches violently. He jumps, then leans in, perplexed. He touches it again. Another twitch. Over the next few weeks, he runs a series of experiments that will accidentally ignite a 240-year scientific revolution: when he connects the frog’s nerve to two different metals—say, a brass hook and an iron railing—the leg contracts every time the metals touch. He calls this mysterious force “animal electricity,” a unique life force that flows through nerves to trigger movement. He has no idea that he has just demonstrated the first direct link between electrical activity and biological motion, the foundational proof of concept that would eventually make brain-computer interfaces (BCIs) possible.

Galvani’s work sparked an immediate and bitter scientific feud that would shape the next century of neurophysics. His rival, Alessandro Volta, a physicist at the University of Pavia, argued that the electricity Galvani observed was not a unique “animal” force, but a generic electrochemical reaction between the two different metals and the moist tissue of the frog’s leg. The debate raged for years, until Volta proved his point in 1800 by building the first modern battery—the voltaic pile—by stacking alternating discs of zinc and copper separated by brine-soaked cardboard. The pile produced a steady current without any animal tissue, proving Galvani’s “animal electricity” was just a special case of general electrochemistry. Both men were right, and both were wrong: Galvani was correct that living tissue generates its own electrical signals, but Volta was correct that the twitch Galvani first observed was driven by the external metal contact, not the tissue’s intrinsic electricity. It would take another 70 years for scientists to prove that neurons themselves generate electrical pulses via the flow of ions across their cell membranes, a mechanism first described by German physiologist Emil du Bois-Reymond in 1843. Du Bois-Reymond demonstrated that when a nerve is stimulated, a measurable electrical current flows along its length, and that this current is the direct cause of muscle contraction. He called this the “negative variation” of the nerve’s resting potential, the first measurement of what we now call the action potential: the brief, all-or-nothing voltage spike that neurons use to communicate with each other, zipping down axons at speeds of up to 120 meters per second.

For decades after du Bois-Reymond’s discovery, the only way to measure neural electrical activity was to insert electrodes directly into or onto the tissue of the nerve or brain—an invasive process that could only be done in animal models or, in rare cases, during human surgery. The core tension of BCI—the tradeoff between signal quality and invasiveness—was already baked into the field’s earliest experiments: direct contact with neural tissue delivered clean, high-resolution signals, but required breaching the body’s protective barriers. For most of the 19th century, no one could figure out how to measure neural activity from outside the skull, let alone from outside the body entirely. That changed, almost by accident, in 1875, when a little-known British physician and researcher named Richard Caton set up a simple experiment in his Liverpool lab. Caton used a device called a galvanometer—a coil of wire suspended in a magnetic field that deflects when an electrical current passes through it—to record electrical activity from the exposed brains of rabbits and monkeys. He found that when the animals were exposed to bright light or loud sounds, the galvanometer’s needle twitched, indicating that the brain was generating measurable electrical signals in response to sensory input. He published his results in a brief paper in the *British Medical Journal*, noting that “the currents evoked in the brain vary with the functional activity of the organ.” His work was largely ignored by the broader scientific community, which still viewed the brain as a “black box” incapable of producing detectable external signals. It would take another 50 years for someone to prove Caton right, not just in animals, but in humans.

That someone was Hans Berger, a German psychiatrist working at the University of Jena. Berger was obsessed with a question that seemed almost mystical to his peers: was the mind a product of the brain’s physical activity, or was it a separate, non-physical entity? He believed that if he could measure the brain’s electrical activity non-invasively, he could prove that mental processes—thought, emotion, consciousness—were rooted in biology, not metaphysics. For years, he tinkered with galvanometers, trying to boost their sensitivity enough to detect the tiny electrical signals that might be leaking through the human skull. His breakthrough came in 1924, when he adapted a device called a string galvanometer, originally designed to measure the heart’s electrical activity (the first ECG), to record from the human scalp. On July 6, 1924, he placed two electrodes on the scalp of his 17-year-old son, Klaus: one on his forehead, one on the back of his head. The needle on the galvanometer jumped, tracing a steady, rhythmic wave of about 10 cycles per second. Berger was stunned. He ran control experiments: when Klaus closed his eyes, the wave grew stronger; when he opened them, it vanished. When Klaus focused on a mental math problem, the wave disappeared again. Berger had discovered the brain’s alpha rhythm, the first measurable signature of human brain activity that varied with mental state. He published his findings in 1929 in a paper titled *Über das Elektrenkephalogramm des Menschen* (On the Electroencephalogram of Man), a 14-page tome that laid the foundation for all modern EEG research.

Berger’s work was met with immediate skepticism. Many of his colleagues dismissed his findings as artifacts—static from the equipment, or electrical signals from the scalp’s muscles, not the brain itself. It took five years for the scientific community to accept his results, after British physiologist Edgar Adrian, who would later win a Nobel Prize for his work on neuron function, replicated Berger’s findings in 1934. Berger’s original EEG machine had a signal-to-noise ratio of roughly 10:1, meaning the brain’s electrical signal was only 10 times stronger than the background noise of the equipment and the environment. To put that in perspective, modern EEG systems have signal-to-noise ratios of 1000:1 or higher, but even Berger’s clunky, finicky device proved a revolutionary point: you could measure the brain’s activity without ever touching the neural tissue itself. For the first time, researchers had a non-invasive window into the living human brain. But that window came with a steep cost: the signal picked up by scalp EEG is a messy, blended average of the electrical activity of millions of neurons, filtered through the scalp, skull, and cerebrospinal fluid. The spatial resolution is roughly 1 centimeter, meaning you can tell which broad region of the brain is active, but not which individual neurons are firing. The temporal resolution is excellent—EEG can pick up changes in activity in milliseconds—but the signal is noisy, easily disrupted by muscle movement, eye blinks, or even the electrical hum of a nearby light fixture. This tradeoff between non-invasiveness and signal fidelity was already on full display: scalp EEG was safe enough to use on any human subject, but it could only tell you so much about what the brain was actually doing.

The next leap forward came not from a psy