Chapter 16

Anesthesia, Anoxia, and Drug-Induced Hallucinations

Why the pharmacological hypotheses cannot account for the lucid dying mind

A young Air Force pilot sits in a sealed altitude chamber at Wright-Patterson Air Force Base in Ohio. He is in his late twenties, fit, sharp, the kind of man who handles a fighter jet with the ease most of us bring to a kitchen chair. The chamber will simulate the air pressure at twenty-five thousand feet. He has been instructed to remove his oxygen mask, write his name on a clipboard, count backward from one hundred, and put the mask back on the moment he feels strange. The training officer wants him to learn what hypoxia — lack of oxygen to the brain — actually feels like, so that in a real cockpit emergency he will recognize the warning signs early.¹

The chamber is depressurized. The pilot pulls off his mask. He smiles. He begins to count. One hundred. Ninety-nine. Ninety-eight. He laughs. The instructor over the intercom asks him to write his name. He picks up the pen. He stares at it. He writes a single letter, then another, then a long shaky line. He laughs again. The instructor asks him to put the mask back on. The pilot looks at the mask. He looks at the camera. He gives a thumbs-up. He does not put the mask on. He tries to count again and gets stuck on a number. He looks confused, then mildly afraid, then sleepy. His head tilts. The instructor depressurizes the chamber and clamps the mask back over his face. Within seconds the pilot is back. He looks at the clipboard. He sees his own scribble and blinks. He does not remember most of what just happened.

That is anoxia. That is what a brain does when oxygen is taken away.

Now consider a different scene. A fifty-one-year-old man is on the floor of his kitchen in Atlanta. He is having a massive heart attack. His wife is on the phone with the dispatcher. His heart stops. The paramedics arrive. They begin chest compressions. They shock him — once, twice, three times. He has no pulse for six minutes. His brain is receiving no oxygenated blood. By every clinical measure, his brain is shutting down for those six minutes the way the pilot's brain was shutting down in the chamber. By any pharmacological account that says oxygen-starved brains produce experiences, this man should be having something like what the pilot had — confusion, fragments, mild fear, then nothing.²

That is not what he reports.

When he wakes up in the cardiac unit two days later, he tells his wife and the cardiologist that he was watching from somewhere up near the corner of the kitchen ceiling. He saw the paramedic with the red beard counting compressions out loud. He saw his wife crying with the phone in her hand. He saw the second paramedic crack open a small plastic case and pull out a syringe. He heard the cardiologist say, when he was wheeled into the emergency room, “We are going to lose him.” He felt no fear. He felt enormous peace. He moved through what he describes as a dark passage, then into a light that did not hurt his eyes but somehow knew him. His mother was there — his mother who had been dead for nine years. She put her hand on his cheek the way she used to. She told him he had to go back. He did not want to go back. Then he was on the floor again, and the paramedic with the red beard was leaning over him, and his chest hurt as if a horse had stepped on it.

The cardiologist will later confirm that he did say, on the phone with the ER, the words the patient reports him saying. The paramedic with the red beard will confirm the actions described. The second paramedic will confirm the syringe. None of these things were spoken in the patient's hearing — he had no measurable pulse and no measurable brain activity for the six minutes in question. And yet his account is detailed, sequenced, calm, lucid, and internally consistent.

Two oxygen-deprived brains. Two utterly different experiences.

One pilot, dazed and forgetful, scribbling on a clipboard. One cardiac patient, lucid and luminous, watching his own resuscitation from above and meeting his deceased mother in a place he cannot name. By any honest reading, those are not two intensities of the same phenomenon. They are not two flavors of the same neurochemistry. They are different in kind.

And yet the most popular contemporary physicalist response to NDE evidence is exactly that. Or close to it. The cluster of pharmacological hypotheses — that NDEs are anoxic hallucinations, or anesthesia-related awareness episodes, or surges of endogenous DMT (dimethyltryptamine, a powerful psychedelic compound), or ketamine-like dissociative states — remains the go-to physicalist answer when the dying-brain hypothesis we examined in the previous chapter runs out of road. Pop-science articles repeat it. Skeptical websites repeat it. Even some careful thinkers repeat it. It has the virtue of seeming, at first glance, to dissolve the puzzle. Of course — people have these strange experiences because their brains are flooded with strange chemicals as they die. Move along. Nothing to see here.

The trouble is that the pharmacological hypotheses, examined carefully, do exactly what the pilot in the chamber did: they scribble where they were supposed to write a name, then forget what just happened. They produce a partial overlap with one or two NDE features and then break down on everything else. They are popular because they sound right. They are wrong because they do not survive the data.

This chapter takes them up — anesthesia awareness, anoxic hallucinations, DMT, and ketamine-like dissociation — one at a time. I will give each its strongest case. Then I will show why each fails.

The Pharmacological Cluster and What It Has to Explain

Before walking through the four hypotheses one by one, I want to set the table. Anyone who proposes a chemical or physiological cause for NDE phenomenology is making a falsifiable claim. They are saying: this is the mechanism that produces that experience. To test whether the claim is right, we have to ask three questions about every proposed mechanism.

First, does the proposed mechanism actually produce the phenomenology the NDEr describes? Not a feature here and a feature there, but the whole structured pattern — the lucid awareness, the OBE, the tunnel and light, the encounters, the life review, the moral significance, the peace? If the mechanism produces something that resembles NDEs in only one respect (say, a sense of detachment from the body), then the mechanism is at most a partial cousin, not a parent.

Second, does the proposed mechanism explain the veridical content? This is the crucial question for any physicalist account. As established in earlier chapters, the strongest evidential category in NDE research is the body of cases in which the experiencer reports verifiable knowledge of events distant from the body during clinical death — conversations they could not have heard, equipment they could not have seen, deceased relatives whose deaths they did not know about.³ Pharmacology produces experiences inside the brain it acts upon. It cannot, by its own rules, produce accurate information about events outside the brain it acts upon. A pharmacological account that explains everything about NDE phenomenology except the veridical content has not really explained NDEs. It has explained something else and called it NDEs.

Third, does the proposed mechanism explain the long-term aftereffects? NDErs are characteristically transformed by their experiences in ways that last for decades. Drug-induced experiences, by comparison, leave aftereffects that decay within months.⁴ Any account that proposes a drug-like mechanism for NDEs has to explain why the drug-like mechanism produces unusually durable changes that drug experiences themselves do not produce.

Three questions. Phenomenology, veridical content, durability. Hold those in mind. They will do most of the heavy lifting in what follows.

One more piece of framing. The pharmacological hypotheses are sometimes presented as competitors with each other and sometimes as a cluster. A skeptic faced with one objection to the anoxia hypothesis will sometimes shrug and switch to ketamine; faced with an objection to ketamine, will switch to DMT; faced with an objection to DMT, will switch back to anoxia. This is not necessarily intellectually dishonest — sometimes the skeptic genuinely thinks NDEs might involve a mix of mechanisms. But it does mean we have to be careful. We are not refuting one fixed proposal. We are examining a cluster of related proposals, all sharing a single underlying assumption: that NDE phenomenology is what a sufficiently exotic dying-brain chemistry produces.

The Anoxia Hypothesis

The simplest pharmacological account begins with an obvious physiological fact: cardiac arrest causes oxygen deprivation in the brain. Within seconds of the heart stopping, the brain's oxygen supply is gone. Within ten to twenty seconds, measurable cortical activity flatlines. Within four to six minutes, irreversible neuron death begins.⁵ The brain, in cardiac arrest, is in trouble. Anyone can see this.

The anoxia hypothesis takes the next step: oxygen-deprived brains produce hallucinations. This is also true. The literature on hypoxia and anoxia is massive and well-replicated. Pilots in altitude-chamber training are reliably impaired by mild hypoxia at simulated thirteen to fifteen thousand feet; reliably confused at twenty thousand; reliably unconscious at twenty-five to thirty thousand without supplemental oxygen.⁶ Mountaineers above eight thousand meters — the “death zone” on Everest and the other 8000-meter peaks — consistently report cognitive impairment, hallucinations, and bizarre experiences.⁷ Carbon-monoxide poisoning produces hallucinations of figures and presences. Severe sleep apnea produces dreamlike intrusions. The category of “hypoxic hallucinations” is well-established in clinical neurology.

So the syllogism looks tidy. NDEs occur during cardiac arrest. Cardiac arrest involves oxygen deprivation. Oxygen-deprivation produces hallucinations. Therefore NDEs are anoxic hallucinations.

The trouble starts when you actually compare the two phenomenologies side by side.

What anoxic phenomenology actually looks like

Hypoxic phenomenology has been studied for nearly a century, since the first altitude-chamber experiments in the 1920s and 1930s.⁸ The picture that emerges is consistent across studies, decades, and researchers. Here is what oxygen-starved brains do, in summary form:

They produce confusion. Subjects cannot follow simple instructions. They have trouble forming coherent sentences. Their decision-making collapses. They will look at a clipboard with their own name written badly on it and not recognize that it is their handwriting.

They produce euphoria mixed with anxiety. The pilot in the chamber laughs. He also looks afraid. The two emotional states do not stabilize; they oscillate. There is no characteristic peace or serene clarity. There is, instead, a kind of giddy unease.

They produce narrowing of the visual field. This is sometimes called “tunnel vision,” but it is not the NDE tunnel. It is a real, physiological narrowing of the peripheral field, with grayout or blackout at the edges. It is unidirectional and sensory, not narrative or symbolic.

They produce fragmented imagery. If hallucinations occur at all, they are typically brief, disorganized, and dreamlike — people see shapes, colors, fleeting figures. They do not see structured narratives.

They produce impaired memory. The pilot in the chamber typically cannot recall most of what happened during the hypoxic period. The G-LOC literature on combat-jet pilots who experience G-induced loss of consciousness is even clearer on this point: post-event amnesia is the rule, not the exception.⁹

They produce clinical signs the experiencer cannot mask. The pulse oximeter dropping below 80 percent is associated with measurable cognitive impairment that bystanders can see — impaired motor function, slurred speech, slowed reaction time. Hypoxia is observable from outside.

Now compare. NDEs show enhanced clarity, not confusion. They show structured narrative sequence, not fragmented imagery. They show characteristic peace, not oscillating euphoria-anxiety. They show improved memory — many NDErs report that the experience is more vivid and easier to recall, decades later, than ordinary waking memories from the same period of their lives.¹⁰ They show coherent sequence: the patient leaves the body, perceives the room, moves through a passage, encounters figures, returns. The features come in something like the same order, in case after case after case.

If you laid out a hundred carefully transcribed pilot-hypoxia reports next to a hundred carefully transcribed NDE reports and asked any honest reader to sort them, the sorting would be trivial. The two stacks belong to different categories of human experience. They are not the same thing.

The G-LOC literature is worth pausing on for a moment because some skeptics have tried to use it as a back door for the anoxia hypothesis. James Whinnery's work on G-induced loss of consciousness in fighter pilots subjected to high G-forces in centrifuge testing is sometimes cited as showing that anoxia-like brain states produce NDE-like phenomena.⁴³ The argument runs: when pilots black out under high G, some report brief tunnel-like phenomena and pleasant feelings, and a small minority report something like a brief OBE. So perhaps high-G blackouts and NDEs are the same thing.

Whinnery's data, read carefully, do not support that conclusion. The G-LOC reports are characteristically brief, fragmentary, and easily forgotten. They lack the structured narrative sequence of NDEs. They do not include encounters with deceased relatives. They do not include life reviews. They do not include the moral-evaluative dimension. The pleasant feelings reported are vague rather than the characteristic peace of NDEs. Whinnery himself was careful not to claim that G-LOC and NDEs are the same phenomenon — that interpretation came from later commentators who took his data further than he did. The G-LOC literature shows that hypoxic brain states can produce some NDE-like fragments. It does not show that they produce NDEs.⁴⁴

The timing problem

There is a second problem with the anoxia hypothesis, and it is more decisive than the phenomenological mismatch. NDEs occur in normoxic medical contexts — that is, in patients whose blood oxygen is normal at the time the NDE is reported.

Consider the cases of NDEs reported during severe trauma but without cardiac arrest. A car-accident patient with massive blood loss but with circulation maintained by paramedics may have a classic NDE. A patient in deep grief at the deathbed of a family member, no medical crisis at all, may report a fear-death experience indistinguishable phenomenologically from a cardiac-arrest NDE.¹¹ Mountain climbers in falls report NDEs before they hit the ground, in the seconds after they slip but before any anoxia could have begun.¹² Heim's classic 1892 collection of mountaineer fall-experiences — the original empirical study that, much later, would be recognized as a record of NDEs — consisted almost entirely of reports of experiences during the seconds of falling, not after impact.¹³

If oxygen deprivation were the mechanism, normoxic NDEs should not exist. They do exist. They are documented. They are part of the standard phenomenological literature. The anoxia hypothesis predicts something the data falsifies.

The medical-data problem

There is a third problem, and it is the deepest. The hypoxia chamber and a cardiac arrest are not actually the same thing.

Hypoxia — what the pilot experiences in the chamber — is a gradual reduction in oxygen reaching the brain. The brain is still being perfused with blood; the blood simply has less oxygen in it. Cortical electrical activity is disordered but ongoing. The patient remains, by clinical standards, alive and breathing. The brain is functioning poorly.

Cardiac arrest is something else entirely. The heart stops. Blood flow to the brain ceases. Within ten to twenty seconds, surface EEG goes flat — not disordered, not slow, not weird, but flat. The brain is not functioning poorly. It is not functioning at all, in the sense detectable by clinical instruments.¹⁴

This matters because the hypoxia hallucination literature is built on cases where the brain is still working, just badly. The anoxic-hallucination model assumes some kind of disordered cortical processing producing weird outputs. In cardiac arrest, there is no cortical processing to be disordered. There is no signal at all. To say that NDEs in cardiac arrest are anoxic hallucinations is to say that a brain producing no measurable activity is producing structured, lucid, accurate, narratively coherent experiences. That is not a hypothesis. That is a contradiction.

The cleanest illustration comes from the Pam Reynolds case, which is treated in detail in Chapter 12. Reynolds was placed in standstill arrest — cooled to sixty degrees Fahrenheit, blood drained from her head, her brainstem-evoked auditory potentials confirmed flat by clicking earphones in place — for the surgical removal of a brain aneurysm. She had no measurable brain activity for over an hour. She was, by every clinical standard, neurologically dead during that interval. She reports a detailed, sequenced, lucid NDE with veridical content.¹⁵ No hypoxia model can account for that, because there was no brain activity to be hypoxic. Reynolds's NDE happened in a brain that was, neurochemically speaking, not running.

The DMT Hypothesis

If anoxia cannot account for NDEs, perhaps something more exotic can. The most popular contender is endogenous DMT — dimethyltryptamine, a powerful psychedelic compound that occurs naturally in trace amounts in the human body. The hypothesis was popularized by the psychiatrist Rick Strassman in his 2001 book DMT: The Spirit Molecule, which reported on his clinical research giving DMT intravenously to volunteers in the early 1990s.¹⁶

Strassman is careful and well worth reading. He is a serious researcher and his clinical observations on DMT are valuable in their own right. He himself was open to the possibility — not as a confident claim but as a possibility — that the dying brain might release a surge of endogenous DMT, and that this surge might account for some of the NDE phenomenology. He floats the idea more cautiously than his popularizers do; the popularizers turn his cautious speculation into a confident announcement that NDEs are DMT trips.

The hypothesis has appeal. DMT experiences and NDEs do share certain features — a sense of meaningful encounter, the perception of beings, an altered sense of time, profound emotional resonance. If DMT experiences feel a little like NDEs, and the dying brain might release DMT, then maybe DMT explains NDEs. The reasoning is not crazy.

It is, however, wrong.

What DMT phenomenology actually looks like

DMT experiences — the ones reported in the Strassman volunteers, in independent research, and in the broader recreational literature — have a characteristic shape that is well-documented across decades and thousands of users.¹⁷

The visual content is dominated by intensely geometric, kaleidoscopic imagery. Tunnels of fractal patterns. Spinning mandalas. Cities of impossible architecture. Beings often described as “machine elves,” jesters, insectoid figures, or other entities that have no obvious counterpart in the user's pre-experience life. The aesthetic is bizarre. It is hyper-saturated. It is alien.

The phenomenology is characteristically perceptually distorted. Time bends in odd ways. Objects melt and re-form. The user's body feels strange — sometimes monstrous, sometimes absent, sometimes reconfigured. There is often a sense of being “launched” or “blasted off” rather than gently lifted out.

The encounters tend to be with beings that are not the user's deceased relatives. The machine elves of DMT folklore are not Aunt Edna. They are not the user's father. They are not Jesus, or Mary, or Buddha. They are something else — something the user has never seen before, often described as a kind of intelligence, sometimes playful, sometimes threatening, often inscrutable. The user does not feel known by them. The user feels confronted by them.

The experience is short. Intravenous DMT lasts ten to fifteen minutes; smoked DMT (the recreational form) lasts five to eight minutes. The user comes back with a sense of having visited somewhere strange, often profoundly. But the experience does not have the leisurely, sequential, narrative pace of an NDE.

And the aftereffects, while sometimes meaningful, do not have the durable life-restructuring shape of NDE aftereffects. Strassman's volunteers reported that their DMT experiences felt important; they did not consistently report that those experiences reorganized their lives.¹⁸

Set this next to NDE phenomenology. NDEs are not characteristically bizarre — they are characteristically familiar. The visual content is recognizable: hospital ceilings, family members, deceased relatives the experiencer knew, sometimes religious figures consonant with the experiencer's prior tradition. The setting is often the experiencer's actual hospital room, then a transitional space, then a destination that has the texture of a real place rather than a fractal one. The encounters are with beings the experiencer recognizes — Mom, Dad, the older brother who died of leukemia at twelve. Even the “Being of Light,” treated more fully in Chapter 21, is described as deeply known by the experiencer, not as alien.

The endogenous-DMT-at-death claim

The DMT hypothesis depends on a specific empirical claim: that the dying human brain releases a surge of endogenous DMT sufficient to produce the experience. This claim has not been established. It remains, as of this writing, a conjecture in search of data.

The evidence sometimes cited is a 2019 study by Strassman and colleagues that detected small amounts of endogenous DMT in dying rat brains.¹⁹ The study is real, and the finding is interesting. But the leap from “small amounts in dying rat brains” to “sufficient quantities in dying human brains to produce a full-spectrum psychedelic experience” is enormous. The rat amounts were tiny, far below what is required for psychoactive effect even by liberal estimates. They were not measured in the human brain. They were not correlated with any specific phenomenology, because the rats were not asked what they were experiencing.

The honest scientific position, here, is that we do not know whether the dying human brain releases significant DMT. We have not measured it. We do not have good methods for measuring it during the brief window of cardiac arrest. The hypothesis, as a hypothesis, is interesting. As an established mechanism, it does not yet exist.

The Ketamine Hypothesis

The third pharmacological account is closer to the mark than DMT, and so it deserves more careful treatment. Ketamine is a dissociative anesthetic widely used in emergency medicine and pediatrics. In sub-anesthetic doses it produces a characteristic experience: a sense of separation from the body, time distortion, sometimes a tunnel-like visual phenomenon, sometimes a sense of meeting beings or visiting other places. Karl Jansen, a psychiatrist who studied ketamine clinically and wrote about it for both professional and popular audiences, argued that ketamine and NDEs are essentially the same phenomenon, produced by the same neurochemistry — the binding of NMDA receptors blocking the release of glutamate, with the dying brain producing endogenous ketamine-like activity.²⁰

The ketamine hypothesis has more going for it than the DMT or anoxia hypotheses. The phenomenologies overlap more substantially. The OBE component, in particular, is shared — ketamine reliably produces a sense of being outside the body. The time-distortion is shared. The tunnel-like imagery is at least sometimes shared. The semantic-similarity research bears this out.

The Martial study and what it actually found

In 2019, Charlotte Martial and her colleagues at the University of Liège published a careful study comparing 625 NDE accounts with 165 ketamine-experience accounts and a large sample of accounts from many other psychoactive substances.²¹ They used semantic-similarity analysis on the textual descriptions, asking which class of altered-state experience most resembled NDEs at the level of language and content.

The headline result was widely reported in pop science as “ketamine experiences resemble NDEs.” That report is true as far as it goes. Of all the substances examined, ketamine was the closest. Closer than DMT, closer than psilocybin, closer than LSD.

What was less widely reported is that ketamine and NDE accounts were still systematically distinct in the semantic-similarity analysis. The two clusters overlapped substantially but did not coincide. Ketamine experiences were the closest of all the chemically-induced altered states — and were still recognizably, statistically different from NDEs. The distinguishing features included: NDEs had higher emotional valence (more peace, less dysphoria); NDEs had more references to deceased loved ones; NDEs had more references to a being or beings of overwhelming love; NDEs had more references to moral significance and life review; NDEs reported more cognitive clarity and less perceptual distortion.

The Martial study is, ironically, often cited by skeptics as supporting the ketamine hypothesis. Read carefully, it does not. It says: of all the substance-induced experiences we have data for, ketamine is closest to NDE in semantic content — and is still distinguishable. That is not a vindication of the ketamine hypothesis. It is a vindication of the claim that NDEs are not even ketamine experiences, despite ketamine being the nearest chemical analogue.

What ketamine experiences are like, in detail

Beyond the Martial data, the clinical and recreational literature on ketamine paints a picture that overlaps with NDE phenomenology in places but diverges in characteristic ways.

Ketamine experiences are characteristically dysphoric — that is, they often involve fear, dissociation that feels alarming, paranoid ideation, and a sense of being trapped in something. The recreational term “K-hole” refers to the experience of dissociative paralysis combined with the inability to exit the experience.²² NDEs are characteristically euphoric in the sense of being marked by extraordinary peace; the few that are distressing (treated in Chapter 28) tend toward fear-of-judgment or sense-of-loss rather than ketamine-style dysphoric paralysis.

Ketamine visual content is characteristically pattern-rich, geometric, and what users sometimes call “drug-like” — recognizable to anyone familiar with recreational psychedelics as belonging to that family of phenomena. NDE visual content is characteristically environmental and recognizable — the OBE phase shows the actual hospital room from above, with accurate detail.

Ketamine encounters, when they occur, often involve abstract beings or strange presences that the user does not recognize. NDE encounters involve specific deceased relatives, the experiencer's own personal history, or religious figures from the experiencer's prior tradition.

Ketamine experiences lack the sustained moral-significance dimension that characterizes NDEs. The life review — the comprehensive moral self-examination treated in Chapter 20 — has no clear ketamine analogue. Users sometimes report meaningful insights, but they do not report a structured ethical accountability of the kind that NDErs describe.

Ketamine experiences are short and chemically bounded. NDEs sometimes feel timeless to the experiencer but occur within minutes of clock time, and unlike ketamine experiences, they are reported as more vivid than ordinary memory rather than as a drug-like fog.

The endogenous-ketamine claim

The Jansen hypothesis depends on the claim that the dying brain releases endogenous ketamine-like compounds. Like the DMT claim, this is a conjecture rather than an established mechanism. The brain does have NMDA receptors; under stress and anoxia, glutamate dynamics are disturbed; in principle some endogenous NMDA-blocking activity could be involved. But no one has measured a surge of endogenous ketamine-like activity in dying human brains during NDE-producing intervals. The hypothesis remains, like the DMT hypothesis, a possibility framed by analogy rather than by data.

And the same point that defeats the DMT hypothesis defeats the ketamine hypothesis with equal force: even if the dying brain were releasing ketamine-like compounds, the resulting experience would still happen inside the dying brain. It could not, by the rules of pharmacology, produce accurate observation of events in the next room or the parking lot or the waiting room across the hospital. That is not what NMDA receptors do. The veridical content remains unexplained.

The Anesthesia-Awareness Hypothesis

The fourth and final pharmacological account requires less treatment because it does not survive its first contact with the data. The anesthesia-awareness hypothesis holds that NDEs reported by surgical patients are episodes of intraoperative awareness — cases where the patient was inadequately anesthetized, retained some level of consciousness, and either heard or sensed events in the operating room that were later reconstructed into something that looked like an NDE.

Anesthesia awareness is a real phenomenon. It is documented in the surgical literature, with an estimated incidence of one to two cases per thousand procedures.²³ Patients who experience it report partial wakefulness, often combined with paralysis from the muscle relaxant, often with significant trauma. It is a serious complication and the anesthesiology profession has worked hard to reduce its incidence with monitoring tools like the BIS (bispectral index) monitor.

Could intraoperative awareness account for surgical-NDE reports? In some specific cases, perhaps. A case where a patient reports a snippet of overheard conversation that turns out to have been at a very high volume and during a period of light anesthesia is at least conceivable as an awareness episode misclassified as an NDE.

But the hypothesis breaks down across the broader data in three immediate ways.

First, anesthesia-awareness phenomenology is characteristically traumatic, not peaceful. Patients who experience awareness report fear, helplessness, and significant post-traumatic stress. NDEs are characteristically marked by peace and absence of fear. The two phenomenologies do not match.²⁴

Second, anesthesia-awareness episodes are typically experienced from the body, not from above it. The patient hears or feels events from where their body is on the operating table. NDEs include consistent OBE phenomena where the experiencer perceives the procedure from a vantage above the body — Pam Reynolds describing the Midas Rex bone saw from above and slightly behind the surgeon, the cardiac patient describing the resuscitation from a corner of the ceiling. These are not the phenomenologies of awakening on the table.

Third, and most decisively, NDEs occur in surgical contexts where awareness is not possible — including, as already mentioned, the Pam Reynolds standstill case, where the patient was at sixty degrees Fahrenheit body temperature, on full cardiopulmonary bypass with all blood drained from her head, and where intraoperative awareness was clinically and electrically ruled out by the auditory-evoked potentials registering flat. The patient cannot have been “aware” in the awareness sense in a brain that is not generating signal.

The anesthesia-awareness hypothesis can perhaps account for some surgical reports. It cannot account for cardiac-arrest NDEs in non-surgical contexts. It cannot account for blind-NDE reports of accurate visual content (treated at length in Chapter 11). It cannot account for non-medical NDEs in mountaineering falls or grief-bedside experiences. As a comprehensive hypothesis, it is one of the weakest of the pharmacological cluster.

The Pharmacological Cluster as a Whole

Let me step back. We have walked through the four major pharmacological hypotheses — anoxia, DMT, ketamine, anesthesia awareness. Each has a real basis in clinical phenomenology. Each captures something true about altered states the human brain can produce. None, at the end of careful examination, accounts for NDEs.

Three observations are worth pulling together.

First, the four hypotheses do not actually fit together. They propose different mechanisms operating under different conditions producing different surface features. If a skeptic invokes anoxia for cardiac-arrest NDEs and DMT for grief-bedside NDEs and ketamine for surgical NDEs and anesthesia awareness for the rest, the skeptic has not produced a theory of NDEs. The skeptic has produced a list of substitutes, each invoked when the previous one failed. This is what I called earlier the “cluster strategy.” It is the signature of empirical strain. When one explanation needs to multiply into four to cover a single phenomenon, that is a sign that the underlying paradigm is overextended.

Second, all four hypotheses fail at the same two points — the points I named at the beginning. None of them explains the veridical content. None of them explains the durable transformative aftereffects. These are not minor footnotes to NDE phenomenology. The veridical content is the empirical heart of the case. The aftereffects are the longest-lasting evidence we have that NDEs are not just impressive subjective experiences but life-restructuring events.

Third, even where the hypotheses partially succeed — ketamine's overlap with the OBE phase, for instance — their partial success is compatible with substance dualism rather than fatal to it. This is a point that deserves its own development.

What the Hypotheses Get Right and Why It Doesn't Help Them

Ketamine produces a sense of separation from the body. NDEs include a sense of separation from the body. There is real phenomenological overlap there. The dualist reading does not deny this overlap. It says something the physicalist reading cannot say: both phenomena involve the body's chemistry interacting with the soul's awareness, and the body's chemistry partly co-shapes the resulting phenomenology in both cases.

Think about it this way. The body, on the dualist account, is the soul's instrument during embodied life. It is not a cage, not a costume, but the means by which the soul interacts with the physical world. When you think a thought, the thought correlates with brain activity — not because the brain produces the thought but because the brain is the body's vehicle for translating the soul's thinking into the embodied life of the person. When the body's chemistry changes, the resulting phenomenology changes too. A drunk person experiences the world differently from a sober one not because the soul has changed but because the body's instrument is being played differently.

So when ketamine produces a sense of separation from the body, the dualist account does not need to deny that the chemistry is doing real phenomenological work. It just denies that the chemistry is producing the experiencer. The chemistry is shaping how the experiencer experiences. The experiencer remains the soul. And in the case of clinical death, the chemistry is doing something more: it is releasing the soul from the body's instrument-function, while the soul's own continued existence carries on the awareness in a way the body normally constrains.

This is why a partial chemical analogue does not, on its own, threaten the dualist case. The dualist is not committed to the claim that NDE phenomenology is chemistry-independent. The dualist is committed to the claim that the experiencer is not produced by chemistry. Those are different claims. The pharmacological-hypothesis literature, when it succeeds, supports the first claim and tells us nothing about the second.²⁵

This is also why the strict identity-claim of physicalism — consciousness just is brain activity, full stop — cannot retreat to “some chemistry we have not yet identified” without conceding the central point. If consciousness is identical to brain activity, then in the absence of brain activity there is no consciousness. The pharmacological hypotheses, taken seriously, all require brain activity of one specific sort or another to produce the relevant chemistry. In cardiac arrest with documented EEG flatline, that activity is not present. The hypothesis cannot run.

The Veridical Problem, One More Time

I keep coming back to the veridical content because every honest physicalist eventually has to. Whatever pharmacological account a skeptic favors, the same question applies: how does the chemistry inside Pam Reynolds's brain generate accurate information about a surgical instrument she has never seen that the surgeon is holding behind his head?²⁶

It cannot. Pharmacology operates on the substrate it is acting on. It does not, and cannot, generate accurate information about events outside that substrate. If the cardiac patient on the kitchen floor in our opening scene is generating his perception of the paramedic's red beard and the cardiologist's words and the syringe-handling from brain chemistry alone, then the brain chemistry has to be sourcing accurate information from somewhere outside the brain, by some mechanism nobody has identified. At that point, calling the mechanism “chemistry” is just renaming the problem.

The dissertation database includes 1,114 cases of accurate distant observation during clinical death (treated at length in Chapter 10) and 33 cases of accurate visual perception by congenitally blind experiencers (Chapter 11). These cases are not all decisive. Some have flaws. Some have weak documentation. But the cumulative pattern across 5,278 cases, with 30.7 percent meeting the dissertation's Strong or Exceptional thresholds (treated at length in Chapter 14), is what we are talking about.²⁷ The pharmacological hypotheses do not address this dataset. They cannot address this dataset, by their own structural rules. Pharmacology does not produce veridical content. The case for veridical NDEs is the case the pharmacological hypotheses cannot touch.

The Physicalist's Deeper Problem

I have engaged each pharmacological hypothesis on its own terms. Now I want to step back and look at the physicalist program as a whole, because something interesting is happening at the level of the program rather than at the level of any single hypothesis.

The pharmacological hypotheses began, decades ago, as a confident and unified picture. NDEs are anoxic hallucinations, full stop. That was Susan Blackmore's position in Dying to Live (1993).²⁸ When the anoxia hypothesis came under empirical pressure, the picture started to fragment. Some physicalists held the line on anoxia. Some moved to DMT. Some moved to ketamine. Some moved to a layered account — anoxia for cardiac-arrest cases, DMT for some others, ketamine for the OBE component, anesthesia awareness for surgical cases. The unified picture broke apart into a cluster of partial accounts.

This is what an empirically strained paradigm looks like in practice. When a single phenomenon requires four or five different explanations, depending on which feature is being addressed, the underlying paradigm is overextended. Thomas Kuhn's classic description of the way paradigms accumulate ad-hoc adjustments in the run-up to their replacement applies here.²⁹ The physicalist program on NDEs is not yet at the point of being abandoned. But it is at the point of multiplying entities to keep the basic commitment afloat.

By contrast, the dualist account is far more parsimonious. The experiencer is the soul. The body is the soul's instrument. During the dying process, the body's chemistry varies enormously — cardiac arrest, anesthesia, drug effect, anoxia, traumatic shock, peaceful old age — but the experiencer remains the same. The varying body chemistry partly co-shapes the phenomenology in each context. The same underlying experiencer carries through. The veridical content is explained because the experiencer is not confined to the brain's substrate. The cross-cultural consistency is explained because human nature is the same across cultures even though chemistry varies. The transformative aftereffects are explained because the experiencer encountered something real and was changed by the encounter.

One commitment, simple, predictive, explanatorily adequate. Versus a cluster of four or five partial mechanisms, none of which explains the central evidential category. This is the situation honest readers find themselves in when they survey the literature carefully. The dualist account is the simpler, stronger reading.

The Charitable Concession the Dualist Owes

I want to be careful not to overstate the case. The pharmacological literature does have something to teach us, and the dualist who refuses to acknowledge it is being uncharitable.

The brain matters. The dying body's chemistry does affect what the dying experiencer experiences. The reason a drunk person's awareness differs from a sober one's is that the body's instrument is being played differently. The reason a person on ketamine experiences dissociation is that ketamine actually does change something about how the soul interacts with the body and through the body with the world. The pharmacological hypotheses are correctly identifying real chemical correlates of altered phenomenology. They are wrong only when they identify those correlates with the experiencer.

Some sub-features of NDE phenomenology may be partially co-shaped by dying-brain chemistry. The vivid visual quality of the OBE phase, the time-distortion effect, possibly the tunnel-and-light imagery, possibly some aspects of the felt-quality of the experience — all of these may have neurochemical contributions. The dualist need not deny this. The dualist needs only to insist that the contribution is co-shaping, not creation. The body shapes the phenomenology; the body does not generate the experiencer.

And here is where I want to be honest about where the evidence runs out. The veridical content cannot be explained by chemistry. The experiencer's continuity across the brain-shutdown of cardiac arrest cannot be explained by chemistry. The transformative aftereffects cannot be explained by chemistry alone (though chemistry may be involved in setting up the conditions for transformation). The deceased-relative encounters where the relative was unknown to the experiencer to be deceased — the “Peak in Darien” cases treated in Chapter 13 — cannot be explained by chemistry. These are the things the physicalist research program has not, after fifty years, found mechanisms for.

What I am saying, in other words, is that the data divide neatly into two categories: the parts the pharmacological hypotheses can partially address, and the parts they cannot address at all. The first category includes some phenomenological features. The second category includes the entire empirical core of the evidence for substance dualism. The pharmacological hypotheses, taken at their charitable maximum, address the first category. They leave the second category exactly where it was.

Where We Are at the End of the Pharmacological Survey

Let me try to summarize what we have established — and what we have not.

We have established that anoxia does not account for NDEs. The phenomenology of oxygen-starved brains is characteristically confused, fragmented, fearful, amnesic. The phenomenology of NDEs is characteristically lucid, sequenced, peaceful, indelible. These are not two intensities of the same phenomenon. They are different in kind. NDEs occur in normoxic contexts where anoxia could not have begun. NDEs occur during full cardiac arrest with flat EEG, where the brain is not “hypoxic” in the chamber-experiment sense but neurologically silent.

We have established that DMT does not account for NDEs. The phenomenologies overlap in some respects but diverge in characteristic ways. DMT experiences are bizarre, fractal, hyper-saturated, often alien-encounter. NDEs are familiar, environmental, often deceased-relative-encounter. The endogenous-DMT-at-death claim remains a conjecture rather than a measured mechanism. Even if DMT were released, the resulting experience would still happen inside the dying brain, leaving the veridical content unexplained.

We have established that ketamine, despite being the closest substance analogue to NDE phenomenology, is not the same as NDEs. Ketamine experiences are characteristically dysphoric, drug-like, abstract-encounter, paralytic. NDEs are characteristically peaceful, environmental, specific-relative-encounter, lucid. The Martial study confirmed ketamine as the closest of all psychoactives and confirmed it as still distinguishable from NDEs. The endogenous-ketamine-at-death claim is, like the DMT claim, a conjecture rather than a measured mechanism.

We have established that anesthesia awareness does not account for surgical NDEs in any general way. Awareness episodes are characteristically traumatic, body-located, partial. NDEs are characteristically peaceful, OBE-located, comprehensive. NDEs occur in deep anesthesia and barbiturate sedation contexts where awareness is clinically and electrically ruled out.

We have not established — and the evidence is silent on — the precise mechanism by which the soul's awareness, during the brain's failure, perceives accurate distant events. We do not know how Pam Reynolds saw the Midas Rex bone saw. We can say with confidence that no current pharmacological hypothesis explains it. We can say with confidence that the simple identity-claim — consciousness is brain activity, period — cannot accommodate it. We cannot yet say what the positive mechanism is. The dualist account predicts that the soul's perception is not bound to the brain's substrate; the dualist account does not yet specify the precise dynamics by which a disembodied or partly-disembodied awareness perceives the physical world. That work remains for theology and philosophy of mind, and it is not the burden of this chapter to complete it.

What this chapter has established is narrower but firm. The pharmacological hypotheses are not strong enough. They never were. The physicalist research program has, for fifty years, been trying to find a chemical cause for NDE phenomenology, and the four candidates examined here are the strongest the program has produced. Each of them captures something. None of them captures the whole.

One Last Honest Note on Cumulative Evidence

I want to be honest about where the careful reader should land at the end of this chapter. I am not asking you to conclude that anoxia produces no relevant phenomenology. I am not asking you to conclude that ketamine and NDEs are unrelated. I am not asking you to conclude that the brain has no role in shaping NDE experience.

I am asking you to conclude something narrower. Pharmacology, taken at its charitable maximum, can account for some sub-features of NDE phenomenology. It cannot account for the full pattern. It cannot, in particular, account for the veridical content. Anyone who claims otherwise is overreaching what the data actually show.

And I am asking you to consider what the cumulative pattern looks like when you combine this chapter with the previous one. The dying-brain hypothesis (Borjigin's gamma-surge work, treated in Chapter 15) does not account for veridical content. The pharmacological hypotheses (treated in this chapter) do not account for veridical content. The chapter that follows this one, Chapter 17, will pull together the cumulative case against the entire physicalist response and develop the broader REM-intrusion, temporal-lobe, and cultural-conditioning hypotheses with the same care. The pattern, by the end of Chapter 17, will be clear. Each physicalist hypothesis succeeds at explaining one phenomenological feature and fails at explaining the central evidential category. Together, they form a cluster, and the cluster strategy itself is what we are calling out.

The dualist case is not concluded by this chapter. It is built up, brick by brick, across the cumulative argument. Chapter 23 will draw the conclusions together. But this chapter has done its part. It has shown that the pharmacological cluster — arguably the most popular and most-often-cited physicalist response to NDE evidence — does not survive careful examination.

Why This Matters for the Christian Facing Death

This is a chapter about hypotheses and chemistry and academic debates, and I am aware that it has been long and at times technical. I want to close where I always try to close: with the dying person, and the people who love them.

Why does any of this matter to a Christian who is facing death, or to the family of someone who is dying, or to the pastor or chaplain who walks with them?

It matters because the question of what is happening inside the dying body has practical consequences for how the dying are accompanied and for how their experiences are received.

If a strict physicalist account were right — if NDEs really were just anoxic confusion, or DMT trips, or ketamine-like dissociation — then the dying Christian's deathbed reports of meeting the deceased, of seeing a loved figure approach, of feeling the pull of light, would be brain noise. They would be the cognitive equivalent of fever dreams. The pastoral response would be to offer comfort but not to take the experiences as evidence of anything beyond the patient's neurochemistry. The pastor at the bedside would treat the report kindly, the way one treats any product of a confused mind, and would be inwardly sure that nothing real was being perceived.

If the pharmacological hypotheses fail — as I have argued, fail substantially — then the situation is different. The dying patient's reports may not be brain noise. They may be early-stage perceptions of a real reality that the dying body's instrument is, for the first time, no longer obscuring. They are not infallible. They are not Scripture. They are not free of cultural conditioning, expectation effects, or interpretive overlay. But they may not be nothing either. They may be the soul's first encounter with what comes next.

I have sat with families whose loved one, in the days before death, looked into a corner of the room and said: your father is here. He's saying it's all right. Or: I see them now. They are coming for me. Or simply: the light. The light. I have heard pastors trained in pharmacological skepticism gently dismiss these reports as confusion. I understand the reflex. I share the desire not to overpromise.

But the careful examination of the pharmacological hypotheses, the kind I have walked you through in this chapter, gives the pastor permission to receive these reports differently. Not credulously — we are not bound to believe everything a dying person says is veridical. With Christian discernment, in light of Scripture, with appropriate caution, but not as automatic delusion. The pastor can stand at the bedside and say to the family: what your mother is reporting may be more than confusion. What she is seeing may be the first edges of what the apostle Paul called “far better.”³⁰

And for the dying Christian themselves, the implication is profound. The dying body's chemistry is not your enemy. It is shaping the texture of an experience that, on the evidence, is more than chemistry. The peace that washes over the dying as the breathing slows, the sense of presence, the visions of loved figures — these are not pharmacology pretending to be God. They are, in the dualist reading the evidence supports, your soul's first apprehension of a reality the body had been keeping at a distance. The chemistry shapes how you perceive it. The chemistry is not what you are perceiving.

There is one more pastoral note I want to add. Some of you reading this have a complicated relationship with the medical environment in which a loved one died. There were drugs. There was anesthesia. There was confusion. The death was not as you had hoped. And you have wondered, sometimes painfully, whether your loved one's last conscious moments were obscured by chemistry rather than illuminated by Christ.

The pharmacological hypotheses, examined honestly, do not require that conclusion. The evidence consistent across thousands of cases is that the dying person's awareness, even in chemically complex contexts, can perceive realities the chemistry cannot account for. Your loved one's last conscious moments may have included — almost certainly did include — awareness of things that the medication cannot have produced. The Christian hope at the deathbed is not undermined by morphine. It is consistent with what every careful examination of the data shows.

And this brings us to the broader argument the next chapter will gather and the chapters after that will build on. The physicalist hypotheses, examined one by one, do not survive. The cumulative case for substance dualism — the case that the experiencer is the soul rather than the brain — grows stronger with each hypothesis we test and each case we examine. Chapter 17 will draw the cumulative refutation together. Chapter 18 and the chapters that follow it will turn from the negative work of refuting alternatives to the positive work of describing what the dying actually report. We are leaving the laboratory. We are returning to the bedside, where the evidence we have been weighing in the abstract is being lived out, in real bodies, in real rooms, by real people whose families wait outside.

The lucid dying mind is real. The pharmacological hypotheses cannot make it disappear. The Christian who walks toward death has more than chemistry to expect.