Chapter 1: The World Before Brains

Core Idea

Life existed for over three billion years before the first brain, and intelligent behavior began long before brains did. Brains appear only in the most recent 15 percent of life’s story. Bacteria already sensed and moved using proteins. Once oxygen made respiration possible, some life had to hunt, which started an arms race between predators and prey. Animals took the hard strategy of catching and swallowing multicellular prey. That strategy required reflexes with “speed and specificity,” which is why animals evolved neurons and muscles and fungi did not. The features that every neuron shares make those fast, precise reflexes possible: all-or-nothing spikes, rate coding, adaptation, and chemical synapses with excitatory and inhibitory neurons. The first animals had nerve nets, not brains.

Frameworks Introduced

• The timeline before brains:
  • About 4 billion years ago, at hydrothermal vents, a DNA-like molecule began to copy itself. This was the first push against entropy.
  • Lipid bubbles enclosed DNA and formed the first cells. Ribosomes then began translating DNA into proteins (protein synthesis). Here the “first sparks of intelligence” appear, because proteins can move and sense.
  • LUCA (the last universal common ancestor) lived about 3.5 billion years ago and resembled a simple bacterium.
  • Cyanobacteria invented photosynthesis. Their oxygen waste caused the Great Oxygenation Event about 2.4 billion years ago. Oxygen levels rose sharply over one hundred million years, followed by one of the deadliest extinctions (the “Oxygen Holocaust”).
  • Aerobic respiration evolved, and microbes began to eat other microbes. Eukaryotes followed. They were over a hundred times larger, produced a thousand times more energy, and invented phagotrophy (engulfing other cells).
  • Plants, fungi and animals each evolved multicellularity on their own. By about 800 million years ago, life fell into three levels of complexity. The first brain appeared about 600 million years ago.
  • Figure 1.3 (the tree of life) places eukaryotes at about 2 billion years ago and multicellularity in plants, fungi and animals at roughly 1 billion to 800 million years ago.
• Three levels of complexity (Figure 1.4, with sizes):
  1. Level one is single-celled life (bacteria and single-celled eukaryotes), 1–10 microns.
  2. Level two is small multicellular life, 10–1,000 microns. It is large enough to engulf single cells and small enough to move with cellular propellers.
  3. Level three is large multicellular life, 1 inch or more. It is too big to move with cellular propellers, so it forms immobile structures.
  • Use: each level eats the level below in a different way. Phagotrophy works only on level one. Eating level two needed a stomach.
• Waiting versus killing (why fungi lack neurons):
  • Animals and fungi both respire, and they are closer relatives of each other than of plants. Sugar is made only by life, so a large respiring organism can either wait for life to die or kill living things.
  • Fungi chose to wait. They digest food externally, spreading spores that grow only near dying matter. Fungi are “Earth’s garbage collectors” and have about six times the biomass of animals.
  • Animals chose to kill. They digest food internally, trapping level-two prey in a stomach.
  • Plants and fungi also sense and move, but over hours to days and clumsily. Coral moves in seconds with well-timed muscle control. Only killing level-two prey requires fast and specific reflexes.
• Edgar Adrian’s three discoveries (single-neuron recording, Cambridge, early 1920s; Nobel Prize):
  1. Spikes: neurons send all-or-nothing signals (action potentials). They work like a telegraph cable with clicks and pauses, not a power line. Adrian compared them to Morse code.
  2. Rate coding: spikes are identical, and the information is in how many fire per second. Pressure, contrast, smell concentration and muscle force are all encoded as firing rates.
  3. Adaptation: neurons keep remapping stimulus strength to firing rate. A strong stimulus raises the threshold for next time, and a weak one makes the neuron more sensitive.
  • Why it works: adaptation solves the squishing problem. A neuron cannot fire faster than about 500 spikes per second, yet natural variables such as light vary by a factor of over a million. Figure 1.11 shows the fix. Without adaptation, one straight line must cover the whole range. With adaptation, the steep S-curve shifts to sit on whatever range the neuron currently faces (low, medium, high).
  • Failure mode: adaptation means a neuron reports change relative to a baseline, not absolute value. The same stimulus gives fewer spikes on repeat.
• The first neural circuit: “do this, not that” (Figure 1.13, a coral polyp’s mouth):
  • When sensory cells detect food on a tentacle, they excite a neuron that drives the opening muscles. A second neuron inhibits the closing pathway.
  • When sensory cells detect food in the stomach, they drive the closing muscles and inhibit the opening pathway.
  • Excitatory plus inhibitory neurons make sure one action wins and the opposite muscles relax. This is “perhaps the first glimmer of intellect to emerge from circuits of neurons.”

Key Concepts

• Entropy: disorder, which the second law of thermodynamics says always increases. Self-replication beats it collectively: copies survive even when individuals do not.
• Protein synthesis: ribosomes build proteins from genes. In the author’s words, DNA became “life’s blueprint, ribosomes its factory, and proteins its product.”
• Abiogenesis: the process by which nonliving matter became life.
• Anaerobic versus aerobic respiration: the oxygen-free route is fifteen times less efficient. Before oxygen, hunting did not pay. After it, hunting started an arms race in which “offensive innovations led to defensive innovations” that required further offensive ones.
• Gastrulation: almost every animal embryo goes from a single cell (zygote) to a hollow sphere (blastula) to a folded sphere with a cavity (gastrula). Figure 1.6 shows these shared stages branching into frogs, fish and mammals; bugs and crustaceans; and jellyfish, corals and anemones. Gastrulation, neurons and muscles are the three features found in all animals and in no other kingdom.
• Coral polyp: possibly the best model of early animals. It is a stomach with a nerve net, tentacles, and muscles (Figure 1.8). Corals, anemones and jellies are “our most distant animal cousins” (Figure 1.7).
• Neuron: the same cell looks different to each field. To a biologist it is a nervous-system cell. To a machine learning researcher it is a unit that computes a weighted sum of its inputs. It is a sensor to a psychophysicist and a muscle controller to a motor scientist. Signals flow from dendrites (inputs) to the axon (output).
• Synapse: the gap between neurons. A spike releases neurotransmitters that cross the gap and bind to receptors on the next neuron’s dendrite (Figure 1.12). Signals are electrical within a neuron and chemical between neurons.
• Excitatory and inhibitory neurons (John Eccles, 1950s): excitatory neurons trigger spikes in the neurons they connect to, and inhibitory neurons suppress them.
• Nerve net: a spread-out web of independent reflex circuits with no central brain, as in today’s corals and jellyfish.

Mental Models

• Think of intelligence as older than brains. Bacteria use motor and receptor proteins to swim away from harmful conditions and toward good ones.
• Think of neurons as a fixed unit and brains as different wirings. Neurons in a human work the same way as neurons in a jellyfish. “What separates you from an earthworm is not the unit of intelligence itself—neurons—but how these units are wired together.”
• Use the “harder strategy” lens when asking where innovation comes from. Fungi won on biomass, but the harder strategy of killing produced neurons.
• Think of early innovations as constraints. Neurons have kept the same basic form for over 600 million years, and evolution has only rewired them.
• Think of a sensory neuron as reporting relative change, not absolute level. The light from a white page is one million times greater in sunlight than in moonlight, and black letters in sunlight are thirty times brighter than a white page in moonlight. Figure 1.10 shows measured values from 30,000 cd/m² (white paper in noon sunlight) down to 0.005 cd/m² (a dark shirt in moonlight). It links a dark shirt in sunlight to a white page in daylight shade as equally bright.
• Map each neuron feature to its purpose. Spikes give fast, coordinated reflexes. Rate coding scales the response to stimulus strength. Adaptation gives sensitivity to faint stimuli without overload at strong ones. Inhibition gives “do this, not that.”

Anti-patterns

• Assuming intelligence needs a brain: single cells already show primitive intelligence through chemical cascades and proteins.
• Assuming plants and fungi cannot sense or move: they can, but slowly and imprecisely. Neurons added speed and specificity, not sensing itself.
• Assuming the most successful strategy drives innovation: fungi outweigh animals about six to one, yet innovation came from the harder path.
• Assuming the first animals lived as gastrulas: all animals develop through a gastrula, but that does not prove an ancestor lived in that form. The author flags this as controversial.
• Historical errors about nerves: the ancient Greeks thought nerves carried “animal spirits,” and Isaac Newton guessed at vibrations in a fluid he called ether. Scientists also once thought the nervous system was one continuous web, until late-nineteenth-century microscopes showed separate neurons.

Rules of Thumb (author’s wording)

• “The story of life, at its core, is as much about energy as it is about entropy.”
• “What began as a pollutant to one form of life became fuel for another.”
• “But as we will continually see, it is usually the worse strategy, the harder strategy, from which innovation emerges.”
• “The original purpose of neurons and muscles may have been the simple and inglorious task of swallowing.”
• “Here’s the beauty: Adaptation solves the squishing problem.”
• “While neural communication within a neuron is electrical, across neurons, it is chemical.”

Worked Example

Adrian’s frog-muscle experiment.

1. Problem: spikes are all-or-nothing, but we perceive graded strengths of sound, light, smell and pain. How does a binary signal carry a number?
2. Adrian took a muscle from the neck of a dead frog and recorded from a single stretch-sensing neuron in it.
3. He hung different weights from the muscle. The spikes all looked the same, but heavier weights produced more spikes per second (Figure 1.9 shows sparse spikes for the lightest weight and dense spikes for the heaviest). This finding is rate coding.
4. Next problem: firing rates run only from 0 to about 500 spikes per second, while natural stimuli vary by over a millionfold.
5. Repeating the same weight showed the answer. A weight that first produced 100 spikes produced only about 80 shortly after, and fewer with each repeat. This is adaptation.
6. Conclusion: neurons encode change relative to a moving baseline, which keeps them precise across huge ranges. The same features are found in all animals, including brainless corals and jellyfish.

Key Takeaways

1. Intelligence began with proteins for movement and perception, billions of years before neurons.
2. Oxygen made respiration and hunting possible, and the predator-prey arms race drove life toward complexity.
3. Animals evolved neurons because eating multicellular prey requires fast and specific reflexes. Fungi chose waiting and never needed them.
4. Every animal’s neurons share spikes, rate coding, adaptation, and chemical synapses with excitatory and inhibitory neurons.
5. Adaptation lets a neuron with a 500-spikes-per-second ceiling cover stimuli that vary a millionfold.
6. Excitatory plus inhibitory wiring gives the first neural logic, “do this, not that.” The first animals had only nerve nets. Brains came with breakthrough #1.

Connects To

• Ch 0: the five-breakthroughs map. This chapter is the starting point before breakthrough #1.
• Ch 2: breakthrough #1, steering. Nerve nets are rewired into the first brain in bilaterians.
• Ch 3: the origin of emotion, which builds on these reflex circuits.
• Ch 4: associative learning, which changes the synapses introduced here.
• Ch 7: pattern recognition, where the machine-learning view of neurons in networks returns.