37 Trillion Cells That Learned to Be Aware

The one idea

There is no little “you” sitting behind your eyes watching the show. The unified self that feels so obvious and singular is something your brain constructs moment to moment, a best-guess model assembled from the activity of roughly 86 billion neurons, riding on a body of about 37 trillion cells that cooperate with no boss and no central controller. The self isn’t a fixed object the brain contains. It’s a process the brain runs, and that process can be measured, disrupted, and fooled.

The science

We can count the pieces. A typical adult brain holds on the order of 86 billion neurons. A careful 2009 cell-count by Suzana Herculano-Houzel and colleagues retired the older “100 billion” figure. Each neuron is essentially a tiny battery that holds a voltage and fires a one-millisecond electrical spike, then passes the signal chemically across a synapse to thousands of others. Thought is that electrochemical traffic in motion. We know it isn’t a metaphor because we can intervene on it. With optogenetics, scientists switch specific neurons on and off with light and watch specific behaviors and memories appear or vanish. Damage a region, change the person.

What about the self riding on top of all that firing? The strongest evidence that the self is constructed rather than given comes from how easily it bends. In the rubber-hand illusion, stroking a fake hand and your hidden real hand in sync makes your brain “adopt” the rubber hand as part of your body within minutes, flinch-when-threatened and all. V.S. Ramachandran spent decades on phantom limbs, mirror therapy, split-brain confabulation and neglect, and all of it points the same way. The feeling of being one unified agent is stitched together from many separate brain systems, and when those systems disconnect, the unity comes apart at the seams.

How does any of this become awareness, the feeling that there’s something it is like to be you? Here we hit the honest edge. Two leading scientific theories frame it differently. Global Workspace / Global Neuronal Workspace Theory (Baars, Dehaene) says a stimulus becomes conscious when it wins a competition and “ignites,” broadcasting widely across the cortex so many systems can access it at once. Most processing stays in the dark, and only a tiny slice reaches the shared stage. Integrated Information Theory (Tononi) says consciousness is integrated information, a quantity called phi (Φ) that measures how much a system is more than the sum of its parts. It predicts awareness lives where information is most unified, which is why the cerebellum, with its vast neuron count but modular wiring, contributes almost nothing to it. Both have real experimental traction. The Perturbational Complexity Index, which delivers a magnetic pulse and reads the EEG “echo,” reliably separates awake, asleep, anesthetized, and even some unresponsive coma patients who will later recover. Both also have serious open problems, and neither closes the “hard problem” of why any of it feels like anything. That gap is real and we should name it plainly rather than paper over it.

(Faithfulness note. The book’s neuroscience here tracks mainstream science well. 86 billion neurons, GWT “ignition,” IIT/phi, PCI, and the self-as-construct are all current consensus framing. Two small corrections for the room. The famous “11 million bits in, ~10–50 bits conscious” compression figure is a vivid popular estimate, not a precise measurement, so use it as an order-of-magnitude image rather than a hard number. And “gamma waves surge at death” rests on a few small case reports, not settled fact, so present it as intriguing and preliminary rather than established.)

What this changes about how you picture reality

You are not a passenger. You are the journey. The thing you call “I” is not a stable object that has experiences. It is itself an experience, regenerated continuously, the way a flame is not a thing but an ongoing event. And it’s being run by a confederation of 37 trillion cells, each carrying your full genetic blueprint, each expressing only its own part, none of them in charge, all of them somehow producing the single felt voice reading this sentence.

That should land as awe rather than loss. The most familiar fact in your life, being yourself, turns out to be one of the strangest things matter does anywhere in the known universe. You are roughly 13 billion years of cosmic and biological history briefly arranged into a pattern stable enough to wonder about itself. The self being a model isn’t a demotion to “just” neurons firing. It’s the discovery that ordinary matter, given enough integration and the right history, can fold around and watch itself. And you are sitting inside the only example you’ll ever experience from the inside.

Two ways to see it

Put two genuinely different scientific framings of the same mystery in front of the room. This isn’t a debate to win. It’s two lenses on one thing.

1. The Broadcast view (Global Neuronal Workspace). Stanislas Dehaene’s framing is that consciousness is what gets shared. Most of your brain works in private, in the dark. A thought becomes conscious the instant it “ignites” and is broadcast to the whole workspace. Watch for a Dehaene talk or the GNW “ignition” schematic, a stimulus crossing threshold and lighting up the cortex. The takeaway to surface is that consciousness here is about access and sharing, not raw quantity of neurons.

2. The Integration view (Integrated Information Theory). Giulio Tononi’s framing is that consciousness isn’t about broadcasting at all, but about how unified a system’s information is, measured as phi. Watch for a Tononi/IIT explainer, ideally one that shows the cerebellum paradox. The cerebellum holds more neurons than the cortex, yet you can lose much of it and stay fully conscious, because its wiring is modular rather than integrated. The takeaway to surface is that consciousness here is about how the parts are woven together, not how many parts there are.

There’s an optional third lens if the room is hungry, the rubber-hand illusion clip. It’s a short, almost comedic demonstration that the boundary of “your body” can be redrawn in two minutes. It makes the abstract claim (the self is constructed) physical and undeniable, and it’s funny, which loosens a room up.

Discussion questions

1. The rubber-hand illusion can make you feel a fake hand is yours within minutes. If the brain can be that wrong about where you end, how much should we trust that the rest of our self-sense is accurate?
2. Global Workspace says consciousness is what gets broadcast. IIT says it’s how integrated the information is. Sitting here in your own head, does either one match what being aware actually feels like to you, or does neither?
3. We can measure phi, scan ignition, and read a brain’s “echo” to tell awake from anesthetized, and still none of that explains why there’s something it feels like to be you. Does that gap bother you, or feel fine to leave open?
4. Your 37 trillion cells coordinate with no central controller and produce a single “I.” Where do you think that single voice actually comes from, and is “you” the cells, the pattern, or something else?
5. If the self is a model the brain builds rather than a fixed thing, does that make “you” feel more fragile, or more remarkable? Does it change anything about how you’d live?
6. IIT’s boldest claim is that any sufficiently integrated system could be conscious, not just brains. If a machine someday scored high on these same measures, what (if anything) would we owe it?

Closing question

How do you feel about this science and its understanding of reality?

Take it further

From The Book (the working text (v10), “Neuroscience & Consciousness” section):

• 86 billion neurons, neuron-as-battery, action potential, synapse (L3861–3877).
• The self is constructed. Rubber-hand illusion and body ownership as inference (L4019–4047, L4082–4084).
• Theories of consciousness, from the hard problem to Global Workspace “global ignition” to IIT and phi (L4108–4134).
• Perturbational Complexity Index, or measuring awareness when behavior can’t (L4136–4146).
• Why neuron count isn’t the point, the cerebellum paradox (L4152–4154).
• 37 trillion cells cooperating without a central controller (L2601, L2633, L2687, L2706).

Real external pointers (verify before citing on screen):

• Suzana Herculano-Houzel, The Human Advantage (2016). The actual 86-billion neuron count and how it was measured. That number replaced the long-repeated “100 billion.”
• Stanislas Dehaene, Consciousness and the Brain (2014). An accessible primary source on Global (Neuronal) Workspace and “ignition.”
• Giulio Tononi and Christof Koch on IIT and phi. Koch’s The Feeling of Life Itself (2019) is the readable entry point.
• Anil Seth, Being You (2021). The best single popular synthesis of the self as the brain’s controlled construction, and it pairs naturally with this session.
• Uncertainty flags. The “11 million bits / ~10 bits conscious” figure is a popular estimate, not a hard measurement. “Gamma surges at death” rests on a few small case reports. And the hard problem of consciousness remains genuinely unsolved. Present these as live questions, not settled answers.

Visual notes

Anchor the room on two visuals, in this order.

1. Neuron-count scale (the opener). Lead with the raw number, 86 billion neurons, ~37 trillion cells, on screen, ideally next to a relatable comparison. 86 billion is on the order of the number of stars in our galaxy, and the room should feel that scale rather than just read it. This is the gut-punch that earns the whole session. The count is enormous, yet the count alone is not the answer.
2. GNW / IIT schematic (the pivot). Then show the two competing mechanisms side by side, a Global Neuronal Workspace “ignition” diagram (stimulus crossing threshold, lighting up the cortex) next to an IIT integration diagram. Visually, this is where the session turns from “look how many pieces” to “it’s not the number, it’s how they’re woven.” The cerebellum paradox image (most neurons, least consciousness) is the single clearest picture of that point.

Room layout suggestion. The science-library neuron-count scale clip plays as people settle. The GNW/IIT schematic goes up when you move into “Two ways to see it.” Close on a single still of one neuron, or of the rubber-hand moment, while the closing question sits on screen.