What unites the human and the jellyfish? The remarkable power of neurons.
For billions of years, Earth teemed with life before the first brain emerged. Brains are newcomers, appearing only in the final 15 percent of life’s epic timeline.
Life’s story is one of constant innovation. While some organisms remained as simple single-celled microbes, others evolved into eukaryotes—sophisticated cells a hundred times larger, generating a thousand times more energy, with vastly more complex internal structures. These eukaryotes became nature’s first predatory masterpieces, engulfing other cells and digesting them within their cellular boundaries. From these advanced cells emerged the first plants, fungi, and animal precursors. While fungi and animals continued as hunters and respirers, plants returned to photosynthesis. The early animals developed a revolutionary adaptation that functions identically whether in a jellyfish or a human brain: the neuron.
Neurons wear many masks depending on who describes them. To biologists, they’re the essential cellular building blocks of the nervous system. Machine learning researchers see them as computational units that process weighted inputs within neural networks. Physicists might describe them as sophisticated sensors that detect and translate external reality. In popular imagination, neurons are transformed into mystical electrical conduits that somehow generate consciousness itself.
Whether in a human brain or a jellyfish’s rudimentary nervous system, individual neurons function according to identical principles. The distinction lies not in the cellular machinery itself, but in the intricate architecture of their connections.
Edgar Adrian, a young neurologist working in England in the early 1920s, revolutionised our understanding of neurons through discoveries that would eventually earn him the Nobel Prize.Â
His first breakthrough challenged the prevailing notion that neural signals flowed in continuous waves. Instead, Adrian demonstrated that neurons communicate through binary “spikes”—discrete electrical pulses that function like biological switches, either firing completely or not at all. The parallel between these neural impulses and the dots and dashes of Morse code struck Adrian immediately.
Adrian’s mind turned to the meaning behind these sequences of neural spikes. Just as Morse code transmitted messages through patterns of dots and dashes, these neural signals might encode biological information. Indeed, Adrian pioneered the application of information theory to neuroscience, becoming the first researcher to conceptualise neural signals as carriers of information. To test this hypothesis, he designed an elegant experiment that would help decipher the neural code.
Using a muscle extracted from a dead frog’s neck, Adrian connected his recording apparatus to one of the muscle’s stretch-sensitive neurons—cells equipped with specialised receptors that activate when the muscle tissue elongates. He then suspended weights of varying heaviness from the muscle tissue. Adrian sought to answer a fundamental question: Would different weights trigger different types of signals from these neurons, or would the neural response follow some other pattern entirely?
Adrian’s recordings revealed a surprising consistency: each spike was indistinguishable from the next—perfect duplicates regardless of the stimulus strength. What varied, however, was their frequency. As he increased the weight, stretching the muscle, the neuron fired more rapidly, producing a denser pattern of identical pulses. This phenomenon, which Adrian termed “rate coding,” became his second groundbreaking discovery.
Adrian’s third discovery proved most astonishing. Imagine trying to represent the entire spectrum of light brightness—from the faintest starlight to direct sunshine—using only the limited range of a neuron’s firing speed. It’s like attempting to map the vastness of the Pacific Ocean onto a bathtub. How do neurons solve this fundamental mismatch? Rather ingeniously, they refuse to maintain static relationships between external stimuli and their internal signalling. Instead, neurons constantly recalibrate, adjusting their sensitivity to match current conditions. When you step from bright sunlight into a dimly lit room, your visual neurons immediately begin shifting their response curves, gradually enabling you to perceive details that were initially invisible. This remarkable process—what neuroscientists formally term “adaptation”—constituted Adrian’s most profound insight.
Neuroscience blossomed during the late nineteenth and early twentieth centuries. Edgar Adrian operated at the same time as Santiago Ramón y Cajal, Charles Sherrington, Henry Dale, and John Eccles – these other remarkable Nobel Prize awardees. Among their most profound revelations was the mechanism of neural communication: neurons don’t actually touch but speak across microscopic gaps called synapses. When an electrical spike reaches the end of one neuron, it releases chemical messengers—neurotransmitters—that sprint across this gap in nanoseconds, dock with receptor proteins, and trigger ion channels to open, altering the receiving neuron’s electrical state.
This duality in neural communication—electrical within neurons and chemical between them—reveals a critical blind spot in the development of artificial intelligence.
Since the 1950s, AI architects have largely modelled their systems on the electrical aspects of neural computation, as exemplified by Ray Kurzweil’s 2012 work “How to Create a Mind. The Secret of Human Thought Revealed.” Yet this approach neglects the rich chemical symphony of neurotransmitters that may hold the key to why machines, despite advances in computation, still cannot learn with the nuance and adaptability of biological brains.
In his book, Kurzweil explicitly sidesteps the chemistry of neural communication—dopamine, serotonin, and their kin—to concentrate on computational abstractions and pattern recognition within the neocortex. This narrow focus reflects a broader trend among AI pioneers, who have largely extracted computational principles from brain architecture without fully investigating how these structures generate our internal models of self and world. The humbling reality, acknowledged by only a handful of researchers in the field, is that contemporary artificial intelligence systems function at cognitive levels comparable to early vertebrates—with a long evolutionary journey ahead before reaching even feline intelligence. The neural enigma that is human thinking remains only partially decoded.