How Does Anesthesia Turn Off Consciousness?

Solving a 180-Year-Old Puzzle with Pure Logic

An Seungwon · Wonbrand CEO · April 12, 2026

---

Introduction

The answer to anesthesia is not in the brain. It is in the cell. More precisely, it lies in a family of membrane receptor proteins that eukaryotes began to evolve roughly two billion years ago — GABA-A, NMDA, K2P, GPCRs. Anesthetic molecules physically lodge themselves in hydrophobic pockets inside these receptors, disrupting their normal function. That disruption upends the balance of neuronal activity, collapses the reciprocal communication between cortex and thalamus, and in humans manifests as loss of consciousness. In plants it shows up as the cessation of leaf folding; in paramecia as stilled cilia; in fruit flies as motor paralysis. The surface appearance varies by organism, but at the bottom the same event is taking place.

For 180 years anesthesia has been administered to hundreds of millions of people while being treated as a puzzle that couldn’t quite be fully solved. In fact, the core answer can be reached through pure logic. This essay is the record of that path. Instead of accumulating laboratory data inductively, it starts from a single observation — plants, too, can be anesthetized — and narrows in on the answer through a process of elimination. And it confirms that the conclusion reached by this route coincides precisely with the consensus of modern anesthetic molecular biology established by Nicholas Franks in a 1994 Nature paper.

---

1. The Irony of 1846

Anesthesia’s history begins on October 16, 1846, at Massachusetts General Hospital in Boston. The dentist William Morton administered ether to a patient while surgeon John Collins Warren removed a tumor from the patient’s neck. The patient felt nothing. Until then surgery had been literal hell. People were bound to chairs and cut with saws. It is no exaggeration to say that modern medicine begins from that day.

One hundred and eighty years have passed. Over that time, surgical rooms worldwide have performed billions of anesthetic procedures, almost all of them ending safely. Even so, for most of that time anesthesia researchers have said that they “don’t fully know why people lose consciousness.” It is an operating technology whose entire principle remains not entirely transparent — that irony has long been a defining feature of the field. But the situation has shifted considerably during the 2020s. Molecular-level targets have become relatively clear, and most of the ladder running from cellular events up through network-level effects has been connected. What this essay aims to show is that the core of that ladder can be reconstructed through logic alone.

---

2. Dismantling the Common View

The first common view to dismantle is this: “anesthesia shuts the brain off.” That is wrong. When the brain of an anesthetized patient is observed with fMRI or EEG, brain cells are still maintaining metabolism, and primary sensory cortex still responds to input. The components are running.

What changes is connection. A February 2025 review in Pharmacological Research by Yuxi Zhou and colleagues formalized this picture as a disruption of the corticothalamocortical circuit. In the normal waking state, cortex and thalamus exchange reciprocal signals dozens of times per second, integrating sensory information. Anesthetics sever this reciprocity. Cortical neurons still fire locally, but they can no longer hold meaningful long-distance conversations with each other. Information comes in but is not integrated; it is not assembled into experience. This is what loss of consciousness looks like at the network level. It is the picture that MIT’s Emery Brown, Michigan’s George Mashour, and Liège’s Vincent Bonhomme have refined since the 2010s.

But this description is a consequence, not a mechanism. Why does corticothalamic reciprocity break down? What kind of molecular event is amplified into this network-level outcome? Answering that question requires going one level deeper.

---

3. The Decisive Clue from Plants

The clue that takes us one level deeper comes from an unexpected place. Plants.

Charles Darwin had already observed it in his 1878 book The Power of Movement in Plants. Expose a mimosa plant to ether vapor and touching its leaves no longer causes them to fold. Remove the gas and the response returns. The same is true of the Venus flytrap. A 2018 paper in Annals of Botany, from an international team led by Frantisek Baluška at the University of Bonn and Ken Yokawa of Kyoto, reproduced these classical experiments with modern instrumentation and directly measured that plant action potentials themselves fail to be generated under anesthetic gas.

It is not only plants. Anesthetics work on the 302-neuron nematode C. elegans, on fruit flies, honeybees, octopuses, and on the single-celled protozoan paramecium. A 2024 review by George Mashour in Neuron emphasized that anesthetics act across bacteria, protists, plants, fungi, and animals, and pointed out the limits of a mammalian-brain-centered mechanistic research program.

The logical implication is clear. If anesthesia works in organisms without brains, then the cause of anesthesia cannot reside in the brain. If the brain were the cause, anesthesia should not work in brainless organisms. But it does. Therefore the cause must lie below the brain, in some foundation shared across life broadly. The corticothalamic network collapse observed in mammals is a species-specific amplification of this deeper cause. The brain is not the cause; it is the stage of amplification.

The same molecular-level event appears as leaf-folding cessation in plants, as stilled cilia in paramecia, as motor paralysis in fruit flies, as loss of consciousness in humans. What differs is the surface form, depending on each organism’s complexity.

---

4. Elimination — Four Candidates

If we list the foundations shared by all organisms, there are four. First, the cell membrane itself. Second, ion channels embedded in that membrane. Third, membrane receptor proteins. Fourth, the cell’s internal signaling systems. Which of these is the answer?

Here we use the method of elimination. Let us assume bacteria are not anesthetized. This assumption has two supports. One is the fact that the question of bacterial anesthesia has remained contested for decades and has never been conclusively shown to yield true anesthesia. The other is economy of reasoning. Assuming bacterial non-anesthesia narrows the answer dramatically and makes it testable.

If bacteria are not anesthetized and only eukaryotes (animals, plants, fungi, protists) are, then the answer must lie in something eukaryotes have but prokaryotes lack. The cause has to live inside the difference between the two groups. This is the classical method of elimination formalized by John Stuart Mill in the nineteenth century.

The cell membrane fails. Bacteria also possess a lipid bilayer cell membrane. The specific lipid composition differs slightly, but the basic architecture is the same. If the membrane itself were the answer, bacteria would have to be anesthetized.

Basic ion channels also fail. Bacteria possess potassium channels, sodium channels, and mechanosensitive channels. When Rod MacKinnon solved the structure of the bacterial potassium channel KcsA by X-ray crystallography in 1998, what shocked the scientific community was that the selectivity filter of this bacterial channel was nearly identical to that of human neurons (MacKinnon received the 2003 Nobel Prize in Chemistry for this work). If common ion channels at this level were the target, bacteria would have to be anesthetized.

Internal signaling systems fail for a different reason. Bacteria do have signaling systems (two-component systems, cAMP, simple calcium signaling). But there is a more fundamental problem. Anesthetics are lipid-soluble small molecules, and they have a tendency to remain within the membrane. Hundreds of X-ray crystallographic and cryo-EM structures bear this out. There is almost no evidence that anesthetics penetrate deep into the cytoplasm and bind directly to G proteins or MAP kinases. When these systems are perturbed during anesthesia, it is because the upper layer was perturbed first. They are closer to consequences than causes.

What remains is the third candidate, membrane receptor proteins. And within this category, crucially, there exist subfamilies that eukaryotes possess but bacteria do not. GABA-A receptors, NMDA receptors, AMPA receptors, the entire GPCR family, K2P two-pore potassium channels, TRP sensory channels. These are relatively new receptor families that emerged after the rise of eukaryotic cells, especially during the evolution of multicellular eukaryotes. Bacteria do not possess them.

And here the second surprise arrives. The actual experimental targets of anesthetics correspond exactly to this list. Propofol’s primary target is the β3 subunit of GABA-A. Ketamine and xenon primarily target the NMDA receptor. The primary targets of volatile anesthetics (sevoflurane, isoflurane) include GABA-A, glycine receptors, and K2P channels. Dexmedetomidine targets the α2 adrenergic receptor, a member of the GPCR family. A June 2025 review by Scott Hansen and Shan Sen of Scripps in Annual Review of Biochemistry summed up the situation in a single sentence: “Sites of anesthetic action are located within ion channels and the plasma membrane.”

The conclusion of the elimination method and the conclusion of the experimental literature meet at the same point. The answer is membrane receptor proteins — specifically, the sophisticated receptor families that eukaryotes evolved anew.

---

5. The Molecular Mechanism — What Anesthetics Actually Do

With the answer narrowed to “eukaryote-specific membrane receptors,” we need to descend one more step. What exactly does the anesthetic do at those receptors?

Membrane receptor proteins are large molecules that thread the cell membrane multiple times as they fold. In the process of folding, an oil-like empty space called a hydrophobic cavity (or hydrophobic pocket) forms inside the protein. That cavity functions as a kind of hinge that the protein uses when it changes shape. When a signaling molecule binds or the membrane voltage changes, the protein toggles between closed and open states, and that movement is realized as a change in the geometry of the hydrophobic cavity.

Anesthetics are lipid-soluble. When they cross the blood-brain barrier into brain tissue, they pass this checkpoint easily, dissolve into the membranes of brain cells, and migrate into the hydrophobic cavities of the proteins embedded in those membranes. Inside the cavity, the anesthetic molecule physically lodges itself. It does not form a chemical covalent bond. It simply takes up residence because its geometry fits the pocket. This is what separates anesthetics from many other drug classes.

This “taking up residence” interferes with the protein’s shape change. When propofol enters the hydrophobic cavity of the GABA-A receptor, the receptor becomes more strongly and more persistently responsive to GABA signals. Inhibition that normally lasted a certain time now lasts two to three times longer. The neuron’s brake is amplified. When ketamine lodges inside the channel of the NMDA receptor, calcium cannot pass through even when glutamate arrives. The neuron’s accelerator stops working. Xenon is even more dramatic. This single-atom noble gas forms no chemical bond whatsoever; it simply settles into the NMDA receptor pocket because its size and shape match, and produces the same effect.

This mechanism has two implications.

First, the mid-twentieth-century orthodoxy of the Meyer-Overton membrane theory — which leapt from the proportionality between lipid solubility and anesthetic potency to the conclusion that “anesthetics dissolve the membrane” — was decisively refuted by Nicholas Franks in a 1994 Nature paper. The reason lipid solubility correlates with anesthetic potency is not that the membrane is the target, but that lipid-soluble molecules can enter the hydrophobic cavities inside proteins. The target is not the membrane but the protein pocket within the membrane.

Second, this mechanism explains why drugs that target different receptors produce convergent results. Propofol pushes the brain toward inhibition; ketamine pushes it toward excitation blockade; both turn off consciousness. The common principle is “lodge in a receptor’s hydrophobic cavity and interfere with its normal shape change.” Even if the direction of interference differs, the result converges on a collapse of neuronal activity balance. When that collapse accumulates, corticothalamic reciprocity cannot be sustained. The network fails. Consciousness goes out.

With this, every rung of the ladder from molecular event to experiential event gets connected. Anesthetic reaches the membrane → lodges in the receptor’s hydrophobic cavity → the receptor’s shape change is obstructed → the balance of neuronal activity collapses → synaptic communication weakens → corticothalamic reciprocity is severed → integrated consciousness vanishes. That entire sequence is the answer.

---

6. Why Only the Brain Goes Out

One natural question remains. Membrane receptor proteins exist throughout the body. They are in the liver, in muscle, in vascular endothelium, in skin cells. Yet when an anesthetic circulates through the whole body, the liver does not stop and the muscles do not dissolve. Why is consciousness the only thing that goes out, and only in the brain? Where does this selectivity come from?

The answer splits into two layers.

The first layer is receptor distribution. The primary targets of anesthetics — GABA-A and NMDA — are effectively restricted to the central nervous system. GABA itself is predominantly a transmitter of the brain and spinal cord, and NMDA is a glutamate receptor involved in learning and memory, also concentrated in the central nervous system. Only trace amounts exist in liver or muscle. That is why propofol does not stop the liver. There is nothing there to hit. This explains “why the body is left untouched and only the brain is hit.”

The second layer is the selectivity within the brain. Even within the brain, the consciousness-related regions (cortex, thalamus) go out first, while the survival-related regions (the respiratory and cardiac centers of the brainstem) remain intact. Both regions possess the same GABA-A receptors, yet the outcomes differ. The difference comes from receptor density and subunit composition. Cortex and thalamus carry extremely dense GABA-A populations, and in particular the density of extrasynaptic GABA-A receptors — those positioned outside the synapse to regulate the baseline excitability of neurons — is overwhelmingly high in these regions. It has been experimentally confirmed that propofol selectively potentiates these extrasynaptic receptors. Meanwhile, the respiratory centers of the brainstem contain the same receptors but at lower density, and are dominated instead by other signaling systems such as acetylcholine and serotonin. That is why, at clinical anesthetic doses, cortex and thalamus are affected first and the brainstem is affected only much later.

This density differential creates anesthesia’s therapeutic window — the narrow zone where consciousness goes out but survival continues. Too little, and cortex and thalamus are insufficiently affected and consciousness is not lost. Too much, and the brainstem is reached and breathing stops. The core skill of the anesthesiologist is managing the width of this zone.

Together, these two layers fully account for the phenomenon of “only the brain, and only consciousness.” The first layer explains “why only the brain”; the second explains “why only consciousness.” Quantitative details at each stage are still being worked out, but the overall picture is already in hand.

---

7. Conclusion — The Answer Is Already Here

To summarize. The answer to anesthesia is not the brain but the cell. The target is the new family of membrane receptors that eukaryotes have accumulated over two billion years. Anesthetics are lipid-soluble; they dissolve into the cell membrane, lodge themselves in the hydrophobic cavities of membrane-embedded receptor proteins, and obstruct those proteins’ shape changes. That molecular event disrupts neuronal activity, the disruption weakens synaptic communication, the weakening leads to the collapse of corticothalamic reciprocity, and that collapse presents in humans as loss of consciousness. Because the target receptors are almost brain-exclusive, the body is not affected; because receptor density is overwhelming in cortex and thalamus, even within the brain only the consciousness-associated regions are selectively shut down. Survival is preserved. An anesthesiologist precisely manages this narrow window — that is what clinical anesthesia is.

This answer can be reached from the laboratory, and it can also be reached from logic. The first path was the one Nicholas Franks consolidated in his 1994 Nature paper, synthesizing decades of crystallography and electrophysiology. The second path begins from a single observation — plants, too, are anesthetized — and uses elimination to narrow the four candidates down to one. The two paths meet at the same point. This convergence is not a coincidence. It happens because the structure of anesthesia is clear enough that any route of approach lands at the same conclusion.

The image of anesthesia as a 180-year-old unsolved mystery is no longer an accurate description. The big mystery has largely been solved. What remains is quantitative refinement — which receptor subunit in which organism responds at what concentration and how, how to model the temporal dynamics of network-level collapse — work of that order. These are still active research themes, but they are of a different character from “we don’t understand the principle.” We do understand the principle. We are simply refining the details.

Boston’s Morton in 1846 would not have been thinking about any of this. He simply wanted surgery to stop being hell. His wish came true, and in the course of its coming true, humanity gained one of the closest tools it will ever have for approaching the phenomenon of consciousness. The internal architecture of that tool is now fairly clearly visible. The most ancient sensory apparatus shared by plants and paramecia and humans — membrane receptor proteins — was the answer. It took 180 years. The answer is already here.

---

References

Seungwon is the founder of Wonbrand, where he builds a portfolio of hardware, software, and API products as a solo operator. This essay is a record of personal inquiry and is not medical advice.