Chapter 11

Oxygen Deprivation, Carbon Dioxide, and Chemical Explanations

A. The Critics' Argument: Chemistry Did It

If you have ever felt dizzy standing up too fast, you have had a tiny taste of what happens when your brain does not get enough oxygen. The world tilts. Your vision narrows. For a split second, everything goes gray. Now imagine that sensation multiplied a hundredfold—your heart has stopped, your lungs have quit, and the oxygen supply to your brain is dwindling toward zero. What would your brain do in those final moments?

According to a group of skeptical researchers, your brain would do something remarkable. It would generate a near-death experience. Not because your soul has left your body. Not because you are catching a genuine glimpse of another world. But because the chemistry inside your skull has gone haywire, and a haywire brain produces strange, vivid, unforgettable experiences that you later interpret as something spiritual. That, in a nutshell, is the neurochemical argument against near-death experiences. And it is one of the most common explanations you will encounter when you bring up NDEs with a scientifically minded skeptic.

Michael Marsh devotes a large portion of Chapter 9 of his Oxford monograph to exactly these kinds of chemical explanations. He organizes them into what he calls "intrinsic mechanisms" and "extrinsic factors." The intrinsic mechanism he highlights first is the role of endorphins—the body's own natural painkillers. Dr. Daniel Carr proposed in the early 1980s that the intense stress of dying triggers a massive release of beta-endorphins, which then flood the limbic system (the emotional core of the brain) and produce the feelings of peace, bliss, and detachment that near-death experiencers so consistently report.¹ Marsh takes this proposal seriously. He notes that beta-endorphin receptors are concentrated throughout the hypothalamus, hippocampus, amygdala, thalamus, and reticular formation—all regions deeply involved in emotion, memory, and arousal.² When the body is under extreme physiological stress—from hypotension (dangerously low blood pressure), hypoxia (oxygen deprivation), hypoglycemia (low blood sugar), or severe trauma—these endorphins are released, producing sedation, pain relief, and passivity.

To his credit, Marsh does not oversell the endorphin hypothesis. He acknowledges that Michael Sabom raised a valid objection: endorphins have a prolonged action that far outlasts the typically brief window of an NDE.³ Marsh also concedes that endorphins alone cannot explain the out-of-body experience or the tunnel phenomenon. But he is not willing to dismiss endorphins entirely. He argues they could serve as one important factor in activating what he calls "latent temporal lobe dysfunction," since beta-endorphin is known to cause non-convulsive epileptiform disturbances in the limbic system.⁴ In other words, endorphins might not cause the NDE directly, but they could light the fuse that sets off a chain of abnormal brain activity.

After treating endorphins, Marsh turns to what he considers the "extrinsic factors"—outside forces acting on the brain. The first and most prominent is cerebral hypoxia, which simply means the brain is not getting enough oxygen. Marsh is careful here. He notes that earlier NDE researchers like Moody and Ring dismissed hypoxia too quickly, and he scolds them for failing to distinguish between anoxia (the complete absence of oxygen) and hypoxia (a reduction in oxygen supply).⁵ These are not the same thing, Marsh insists, and the distinction matters medically. He argues that during cardiac arrest or severe circulatory collapse, the brain is almost certainly experiencing some degree of hypoxic insult, and that this could produce aberrant neurological activity as the brain struggles to regain function.

Marsh is careful to concede that hypoxia alone cannot be the sole cause. He notes that induced hypoxia produces "a progressive muddling and confusion of cognitive ability," which stands in sharp contrast to "the sharpness of cognitive insight evident during" NDEs.⁶ He even references the case of a British pilot who experienced both high-altitude hypoxia and a later NDE and reported that the two experiences were completely different.⁷ He also acknowledges Sabom's documentation of a cardiac arrest patient whose blood oxygen level was actually above normal during the NDE event, with carbon dioxide below normal—a case that seems to undercut the hypoxia explanation directly.⁸ Yet Marsh pushes back on this evidence, arguing that peripheral arterial blood gas measurements may not accurately reflect what was happening locally inside the brain at the time.⁹

Next, Marsh turns to hypercarbia—an elevated level of carbon dioxide (CO₂) in the blood. A cardiac arrest can produce hypercarbia as the body stops expelling CO₂ through normal breathing. Marsh references the classic experiments of psychiatrist L. J. Meduna in the 1950s, who gave psychiatric patients and healthy controls a mixture of 30 percent carbon dioxide and 70 percent oxygen.¹⁰ Some of Meduna's subjects reported experiences that bear a surface resemblance to NDEs: frightening figures, complex geometric shapes, feelings of cosmic insight. But Marsh himself admits that many of the features produced by CO₂ intoxication are "notably unlike the phenomenology of" NDEs.¹¹ The subjects experienced bizarre compulsions, illogical imagery, and physical signs of severe neurological distress—not the peaceful, coherent, life-transforming encounters reported by near-death experiencers.

Finally, Marsh treats the anaesthetic gas ketamine (covered primarily in Chapter 16 of this book). He calls ketamine a potential "paradigm" for understanding NDEs and notes that sub-anaesthetic doses of ketamine can produce experiences that share some surface features with NDEs: coloured imagery, a feeling of timelessness, a sense of acquiring all knowledge, emotional highs, and out-of-body sensations.¹² He acknowledges "troublesome differences" between ketamine experiences and NDEs—including the fact that ketamine frequently produces deeply unpleasant, even terrifying reactions and that 60 percent of surgical patients in one study did not wish to repeat the experience.¹³ Nevertheless, Marsh finds these neurophysiological parallels significant and worthy of further investigation.

Fischer and Mitchell-Yellin, while not focusing on the specific chemical mechanisms as Marsh does, provide the broader philosophical framework that makes this chapter necessary. In their discussion of what they call the "Single Explanation" problem, they argue that pro-NDE researchers like Pim van Lommel and Eben Alexander make a logical error: they consider each chemical explanation in isolation, show that it fails to explain every feature of every NDE, and then reject it entirely.¹⁴ Fischer and Mitchell-Yellin counter that a multi-factor physical explanation—combining hypoxia, hypercarbia, endorphins, NMDA receptor activity, temporal lobe disturbance, and other factors working together—might account for NDEs without any need to invoke the supernatural. No single chemical does the job alone, they argue. But perhaps a cocktail of chemicals, acting in concert on a dying brain, produces the full NDE.

This is a sophisticated argument, and it deserves a careful response. Let us give it one.

B. Identifying the Weaknesses

Before we build the pro-NDE case, we need to lay out the specific problems with the chemical explanations as presented by Marsh and the broader skeptical community. There are several, and they are significant.

Weakness 1: Hypoxia Produces the Opposite of an NDE

This is the single biggest problem with the hypoxia hypothesis, and it is a problem that Marsh himself partly acknowledges but never fully reckons with. Hypoxia does not produce lucid, organized, vivid, coherent, life-transforming experiences. It produces confusion. Disorientation. Muddled thinking. Agitation. Memory loss. Eventually, unconsciousness.

Marsh admits this. He writes that induced cerebral hypoxia reveals "a progressive muddling and confusion of cognitive ability" that stands in contrast to NDEs.¹⁵ But having admitted this devastating problem, he simply moves on. He does not resolve it. He does not explain how a mechanism that is known to produce mental confusion could simultaneously produce the most vivid, clear, structured, and memorable experience of a person's entire life. This is not a minor loose end. It is a gaping hole in the argument.

Chris Carter puts the point with characteristic sharpness in Science and the Near-Death Experience: when the brain is starved of oxygen, "the person becomes progressively more disoriented and confused. This is in sharp contrast to the clarity of thought and perception described over and over again in the reports of near-death experiencers."¹⁶ In thousands of laboratory experiments in which researchers have progressively reduced the oxygen supply to human subjects—including the famous medical school exercises in which students breathed through a carbon dioxide absorber until they slowly suffocated—not one person has ever reported an NDE.¹⁷ Not one. That is a striking absence.

Weakness 2: The Selectivity Problem

If NDEs are caused by the chemistry of a dying brain, then every dying brain should produce one. After all, every cardiac arrest patient experiences essentially the same physiological crisis: the heart stops, blood flow to the brain ceases, oxygen levels plummet, carbon dioxide builds up, and the same cascade of neurological events unfolds. Yet only a minority of cardiac arrest survivors report NDEs.

Pim van Lommel's landmark prospective study, published in The Lancet in 2001, found that only 18 percent of cardiac arrest survivors in his study reported an NDE.¹⁸ Sam Parnia's AWARE study found even lower numbers. Other prospective studies have reported similar ranges, typically between 10 and 20 percent.¹⁹

Think about what this means for the chemical hypothesis. All of these patients experienced the same physiological catastrophe. Their brains all underwent the same cascade of oxygen deprivation, CO₂ buildup, and neurochemical flooding. Yet roughly 80 to 90 percent of them reported nothing at all. No tunnel. No light. No bliss. No life review. Just a blank, an absence, a gap. If the chemistry of the dying brain causes NDEs, why does it only work sometimes?

Marsh never adequately addresses this selectivity problem. He acknowledges the data, but he treats it as though the mere fact that most patients do not report NDEs under the same physiological conditions somehow does not undermine the very hypothesis that those conditions cause the NDE. It does. Profoundly.

Weakness 3: The Blood Gas Evidence

Marsh is aware of Sabom's case in which a cardiac arrest patient's blood gases were monitored during the resuscitative effort—and the patient's arterial oxygen was above normal and his carbon dioxide was below normal.²⁰ This is, on its face, devastating for both the hypoxia and hypercarbia hypotheses. The patient was experiencing an NDE at the very time his blood showed plenty of oxygen and no excess carbon dioxide.

Marsh's response is to suggest that peripheral blood gas levels may not accurately reflect what was happening inside the brain at the cellular level.²¹ Fair enough. That is a legitimate scientific point. But notice what Marsh has done here: he has shifted from claiming that hypoxia and hypercarbia cause NDEs to speculating that they might have been present despite the direct evidence that they were not. That is a significant retreat. And it cuts both ways: if we cannot trust peripheral blood gas measurements to tell us about brain conditions, then the skeptic cannot cite blood gas abnormalities in other cases as evidence for the hypoxia hypothesis either. Marsh wants the uncertainty to work only in his direction.

Weakness 4: The Endorphin Hypothesis Explains Too Little

Even if we grant that endorphins are released during dying, what do endorphins actually do? They produce pleasant feelings. Pain relief. A general sense of well-being. They are the brain's natural morphine. And morphine is a wonderful drug. But morphine does not produce complex, structured, narrative experiences with veridical perception of real-world events. It does not create life reviews in which a person re-experiences every significant moment of their life. It does not generate encounters with deceased relatives the person did not know had died. It does not produce accurate descriptions of resuscitation procedures being performed on the person's body in the operating room.

Marsh acknowledges that endorphins cannot explain the OBE or the tunnel.²² That is a remarkable concession. If endorphins cannot explain the two most dramatic and evidentially significant features of the NDE, then what exactly are they explaining? A pleasant mood? The NDE involves far more than a pleasant mood.

Carter amplifies this point by noting that endorphins have a slow decay over several hours, which does not match the instantaneous return of pain that many near-death experiencers report upon returning to their bodies.²³ If endorphins were flooding the brain and producing the NDE's blissful feelings, the bliss should taper off gradually, not end abruptly the moment the person "returns." But experiencers consistently describe the opposite: one moment they are bathed in indescribable peace, and the next they are back in their body, in agony, gasping for breath.

Weakness 5: CO₂ Intoxication Produces the Wrong Experiences

Marsh's own description of Meduna's CO₂ experiments actually undermines the hypercarbia hypothesis more than it supports it. Meduna's subjects experienced bizarre compulsions, frightening figures and shapes, and complex geometric patterns—features that Marsh himself calls "notably unlike" the NDE.²⁴ Extreme hypercarbia also produces convulsive, sometimes violent, muscle movements—another feature completely absent from NDE reports.²⁵

Neuropsychiatrist Peter Fenwick, who studied over three hundred NDE cases, pointed out an additional difficulty. In a cardiac arrest, any buildup of CO₂ would be accompanied by an equally catastrophic drop in oxygen. As Fenwick argued, a brain experiencing both hypercarbia and severe hypoxia simultaneously would be so profoundly impaired that it could not construct coherent experiences, let alone remember them afterward.²⁶ A failing brain, by definition, produces experiences that are limited, confused, and disorganized. The NDE is the exact opposite.

C. The Pro-NDE Response: Why the Chemistry Does Not Add Up

We have identified the weaknesses. Now it is time to build the full case. The chemical explanations for NDEs fail not in just one or two respects but across the entire range of evidence. What follows is a thorough examination of why.

The Hypoxia Problem in Full

Hypoxia—a condition where the brain does not get enough oxygen—is one of the most common skeptical explanations for NDEs. The idea is straightforward: when the brain is starving for oxygen, it malfunctions, and that malfunction produces the strange experiences people report. It sounds plausible in a brief summary. But the moment you look at what hypoxia actually does to people, the explanation falls apart.

In the 1930s, a physician named R. A. McFarland studied the effects of hypoxia on members of the International High Altitude Expedition to Chile. Mountain climbers exposed to the thin atmosphere at extreme altitudes reported mental laziness, heightened irritability, difficulty concentrating, slowness in reasoning, and difficulty remembering.²⁷ These are the hallmarks of oxygen-deprived brains: slowness, confusion, impaired memory. Not vivid panoramic life reviews. Not detailed observations of surgical procedures. Not encounters with deceased loved ones.

The laboratory evidence is even more damning. As Carter documents, it was once common practice for medical students to be shown the consequences of oxygen deprivation by breathing through a carbon dioxide absorber into a spirometer filled with air. The students would slowly suffocate without realizing it, their performance at simple tasks becoming increasingly inept until they lost consciousness. Thousands of people participated in these experiments over the years. Not a single one reported an NDE.²⁸

The military research tells the same story. Pilots undergoing hypoxia training in flight simulators experience confusion and disorientation so severe that some try to land their simulated planes on top of clouds before blacking out.²⁹ They do not report leaving their bodies. They do not see a brilliant light or feel overwhelming peace. They get muddled and pass out.

And then there is the testimony of Allan Pring, a British Royal Air Force pilot who experienced both high-altitude anoxia and, years later, a near-death experience. He was in a unique position to compare the two. His verdict could not have been clearer: there was no similarity between them. The NDE, he said, was "crystal clear" from beginning to end and remained so for fifteen years afterward. The hypoxia experience was nothing like it.³⁰

This kind of direct testimony from someone who has experienced both conditions is extraordinarily valuable. It is not an abstract theory. It is a firsthand comparison by a trained aviator who knew exactly what oxygen deprivation felt like—and who said it bore no resemblance to his NDE.

The G-LOC Research: Whinnery's Centrifuge Studies

Some skeptics have pointed to the research of James Whinnery, a Navy physician who studied acceleration-induced loss of consciousness (G-LOC) in fighter pilots using a human centrifuge. When pilots are subjected to rapid acceleration forces, the blood is pushed downward in the body, away from the brain. At around 5 G (five times the normal force of gravity), the brain's blood supply can be so compromised that the pilot loses consciousness. The U.S. Air Force has collected data on over a thousand such episodes.³¹

Marsh references these studies and finds them significant. He notes that some pilots reported tunnel vision, short "dreamlets" with emotional content and vivid imagery, feelings of floating, and a sense of euphoria upon recovery.³² He argues that these "dreamlets" are "microcosms" of the NDE, occurring during the same kind of rapid cerebral ischemia (oxygen starvation due to reduced blood flow) that happens during cardiac arrest.³³

But the differences between G-LOC experiences and NDEs are far more striking than the similarities. Carter lays them out systematically.³⁴ First, tunnel vision—the loss of peripheral vision caused by reduced blood flow to the retina—is not the same as perceiving a structured tunnel. Tunnel vision is the absence of the outer visual field. It is not a dark passage with a light at the end. Second, the convulsions that accompany about 70 percent of G-LOC episodes have never been reported during NDEs.³⁵ Third, the "dreamlets" Whinnery documented are extremely short, fragmentary, and, in the pilots' own descriptions, indistinguishable from ordinary dreams—complete with illogical content, convulsive body movements incorporated into the narrative (one pilot dreamed he was fishing and yanking a rod because his arms were jerking), and difficulty remembering the experience afterward.³⁶ NDEs are the opposite: they are coherent, logical (within their own frame), vividly remembered for decades, and consistently described as "more real than real."

Fourth, no life review has ever been reported during G-LOC. No encounter with a being of light. No meeting with deceased relatives. No border or point of no return. No decision to come back. And no lasting transformation of values and personality.³⁷ These are the core features of the NDE, and they are entirely absent from the centrifuge experience.

Whinnery himself acknowledged the differences. He described the dreamlets as having "illogical content and organization" and noted that subjects experienced them as indistinguishable from ordinary sleep dreams—just shorter.³⁸ NDE experiencers, by contrast, insist with remarkable consistency that their experience was not a dream and was, if anything, more real than their waking life.

The Selectivity Problem: Why Only Some Dying Brains?

I keep coming back to this because it is devastating for the chemical hypothesis, and Marsh never gives it the weight it deserves.

Pim van Lommel's Lancet study was designed as a prospective study—the gold standard in medical research. He did not go looking for NDE stories after the fact. He systematically interviewed every consecutive survivor of cardiac arrest at ten Dutch hospitals over a period of years. All of these patients had experienced the same physiological crisis: their hearts stopped, they lost consciousness, they were clinically dead for some period of time, and they were then resuscitated. The chemical conditions inside their brains were, by all medical reasoning, essentially the same.³⁹

Yet only 18 percent reported an NDE. Eighty-two percent reported nothing at all.

Van Lommel investigated every variable he could think of—duration of cardiac arrest, medications administered, prior knowledge of NDEs, religious belief, age, sex—and could find no physiological factor that predicted who would have an NDE and who would not.⁴⁰ The chemistry was the same. The physiology was the same. But the experiences were radically different.

If hypoxia causes NDEs, then every patient whose brain was deprived of oxygen during cardiac arrest should have had one. They did not. If hypercarbia causes NDEs, then every patient with a CO₂ buildup should have had one. They did not. If endorphins cause NDEs, then every patient experiencing the massive physiological stress of cardiac arrest should have had one. They did not.

This is not a minor inconvenience for the chemical explanation. It is a fundamental problem. An explanation that works only 18 percent of the time, when the proposed cause is present 100 percent of the time, is not a very good explanation.

The Blood Gas Case: Sabom's Demolition

The case that Marsh tries to wave away—Sabom's documentation of a patient whose blood gases were monitored during resuscitation—deserves a closer look because it directly tests the chemical hypothesis.

The patient was a retired mechanic who suffered cardiac arrest on an emergency room table. During resuscitation, arterial blood was drawn from his left femoral artery (in the groin area) to measure blood gas levels. The patient later described, from an out-of-body perspective, watching the medical staff work on his body. He accurately described the procedure, including the "shot in the groin"—which was actually the needle used to draw the blood sample.⁴¹

The blood gas results showed an arterial oxygen level well above normal and an arterial carbon dioxide level actually below normal.⁴² High oxygen. Low CO₂. The two chemical culprits most frequently blamed for producing NDEs were both absent.

Marsh's response, as I noted, is that peripheral blood gas levels might not reflect what was happening inside the brain. Carter acknowledges this objection but points out that the uncertainty cuts both ways: if we cannot use blood gas levels to test the chemical hypothesis, then we cannot use them to support it either.⁴³ And in any case, we have one direct measurement showing normal oxygen and low CO₂ during an NDE with veridical perception. The burden of proof falls on the skeptic to explain how a brain that was apparently receiving adequate oxygen and had no excess CO₂ still managed to produce what the skeptic claims can only be caused by oxygen deprivation or CO₂ excess.

Endorphins: The Bliss Without the Content

Let me be generous to the endorphin hypothesis and grant everything it claims. Yes, endorphins are released during severe physiological stress. Yes, they produce feelings of peace, calm, and well-being. Yes, they are naturally occurring opioids that could account for the pleasant emotional tone of many NDEs.

But here is the problem: endorphins explain the feelings. They do not explain the content.

An NDE is not just a pleasant feeling. If it were, we could dismiss it fairly easily—plenty of drugs produce pleasant feelings, and nobody claims those feelings are glimpses of another world. But an NDE involves detailed, structured, verifiable content: specific observations of real-world events; encounters with identifiable deceased persons; information obtained during the experience that the experiencer could not have known through normal means; and a coherent narrative that follows a recognizable pattern across cultures, religions, and centuries.

Endorphins cannot produce any of this. They are painkillers. They make you feel good. They do not give you accurate knowledge of what the surgeon is doing to your body three rooms away. They do not introduce you to your deceased great-aunt whose death nobody told you about. They do not produce a panoramic life review in which you re-experience every significant moment of your existence from childhood to the present.

As Carter notes, there are dozens of psychoactive neurotransmitters in the brain that could plausibly be released under stress—serotonin, epinephrine, dopamine, and many others. Several theories have been constructed around one or some combination of these. But at this stage, all such theories are little more than speculation because they are extraordinarily difficult to test directly. We cannot insert probes into numerous brain areas to take tissue samples from dying patients or patients undergoing resuscitation.⁴⁴ What we can do is look at the indirect evidence—and the indirect evidence consistently shows that the experiences produced by these chemicals do not match the NDE.

The CO₂ Problem: Meduna's Experiments Backfire

When you actually read the details of Meduna's carbon dioxide experiments from the 1950s—the very experiments that skeptics cite to support the hypercarbia hypothesis—you discover something the skeptics often fail to mention. The experiments included 150 psychiatric patients and 50 controls who inhaled a mixture of 30 percent CO₂ and 70 percent oxygen. Some subjects did report experiences with a surface resemblance to NDEs: feelings of being outside their bodies, bright light, revival of memories, cosmic feelings.⁴⁵

But the full range of experiences was very different from the NDE pattern. Subjects reported brightly colored geometric patterns, fantasized objects like musical notes floating by, seeing things in duplicate or triplicate, and severe convulsive muscle movements.⁴⁶ None of these are features of the NDE. And notice what is missing from the CO₂ experience: no coherent narrative structure, no encounters with identifiable deceased persons, no life review, no border or point of no return, no lasting transformation of personality and values, and—most importantly—no veridical perception of real-world events.

Fenwick's critique is especially pointed here. In a cardiac arrest, any CO₂ buildup would be accompanied by simultaneous severe oxygen deprivation. The brain is even more sensitive to oxygen loss than to CO₂ excess. So hypercarbia during cardiac arrest comes packaged with severe hypoxia—and we already know that hypoxia produces confusion, disorientation, and rapid descent into unconsciousness. A brain being hammered simultaneously by too much CO₂ and too little oxygen is not a brain that should be able to construct crystal-clear, coherent, structured, life-changing experiences.⁴⁷ As Fenwick put it: "A failing brain, by definition, produces experiences which are limited, confused and disorganized. The very opposite is true of the NDE."⁴⁸

NDEs Without Any Chemical Crisis

Here is a fact that Marsh's chemical framework has great difficulty accommodating: some NDEs occur in people who are not in any physiological crisis at all. Carter documents cases of near-death experiences that occurred before the person was in any medical danger—such as a woman who had a full NDE, including a tunnel and light, before her car was struck by an oncoming vehicle.⁴⁹ Her NDE happened before the accident, before any injury, before any possibility of hypoxia or hypercarbia or endorphin release. Her brain was functioning normally at the time.

If NDEs are caused by the chemistry of a dying brain, how do we explain NDEs in healthy brains? Marsh's entire framework depends on metabolic disturbance as the engine of the experience. But some experiences occur in the absence of any metabolic disturbance at all. This does not fit.

J. Steve Miller, in Near-Death Experiences as Evidence for the Existence of God and Heaven, reinforces this point. He notes that in some cases NDEs occurred before any physiological stress whatsoever—in situations like imminent traffic accidents where the person's body had not yet experienced any injury or oxygen deprivation.⁵⁰ The chemical explanation has no foothold in these cases.

Fainting Studies and the Lempert Experiments

Marsh also draws on the fainting studies of Thomas Lempert, who induced syncope (temporary loss of consciousness through fainting) in 42 healthy subjects. Of those, 60 percent reported visual and auditory experiences during the brief unconscious period. Some experienced colored patches, formed scenes, even perceptions of being in another world. Twenty percent reported encountering "preternatural beings." Sixteen percent reported something resembling an out-of-body experience.⁶³

Marsh finds this data compelling. These experiments, he argues, show that a brief period of cerebral ischemia (reduced blood flow to the brain) can produce phenomenological experiences strikingly similar to NDEs, all within a carefully timed 12-second window of unconsciousness.⁶⁴ The implication, for Marsh, is clear: if we can produce NDE-like experiences by briefly cutting off blood flow to the brain, then real NDEs must work the same way—just on a larger scale.

But the comparison does not hold up under scrutiny. Miller has examined these claims carefully and found that every time he went back to the original data, the similarities that skeptics trumpeted turned out to be superficial at best.⁶⁵ The Lempert subjects experienced greyness, colored patches, vague scenes, and noises—not the structured, coherent, narrative NDE with its recognizable sequence of out-of-body experience, tunnel, light, encounter with deceased relatives, life review, border, and return. The experiences were fragmented and dream-like, with "illogical content" and difficulty remembering afterward. The subjects themselves described their episodes as similar to ordinary dreams, not as encounters with a transcendent reality.

Most critically, the Lempert experiments produced no veridical perception. No subject accurately described events happening in the room while they were unconscious. No subject acquired information they could not have known through normal means. The very feature that gives NDE evidence its teeth—the verifiable accuracy of the experiencer's perceptions—is completely absent from these induced syncope experiments. Producing vague, dreamy imagery through oxygen deprivation does not explain how a cardiac arrest patient accurately describes a surgical instrument she has never seen.

The "Realer Than Real" Quality

One of the most consistent and remarkable features of NDEs is that experiencers describe them not as dreamlike or hallucinatory but as "more real than real"—more vivid, more coherent, and more deeply meaningful than ordinary waking consciousness. This is the exact opposite of what any chemical mechanism should produce.

Jeffrey Long, whose NDERF (Near-Death Experience Research Foundation) database contains thousands of cases, has documented this "realer than real" quality with considerable rigor. In his surveys of NDE experiencers, an overwhelming majority described their experience as more vivid and more real than anything they had experienced in their waking lives.⁶⁶ These were not people using vague, impressionistic language. They were struggling to convey that their NDE felt more real than the conversation they were having with the researcher at that very moment.

Long also found that NDE experiencers who had separately experienced hallucinations, drug-induced states, or lucid dreams consistently and emphatically distinguished their NDE from all of these. The NDE was, in their assessment, a fundamentally different category of experience.⁶⁷ This is important testimony, because it comes from people who are in a position to compare. They know what hallucinations feel like. They know what drug-induced experiences feel like. And they say the NDE was not like those things. It was something else entirely.

This creates a serious problem for the chemical hypothesis. If NDEs were produced by the same mechanisms that produce drug hallucinations or hypoxia-induced confusion, then the experiential quality should be similar. Hallucinations feel like hallucinations. Dreams feel like dreams. Drug experiences feel like drug experiences. Hypoxia-induced episodes feel confused and fragmentary. But NDEs feel more real than any of these—more real, even, than ordinary waking consciousness. Whatever is producing them is not producing a degraded or distorted version of normal experience. It is producing something that exceeds normal experience in clarity, coherence, and vividness.

The research of Thonnard and colleagues, published in 2013, provides further empirical support. They compared the memory characteristics of NDEs with those of real events and imagined events, finding that NDE memories had more characteristics of memories of real events than of imagined or fantasized events.⁶⁸ The memories were richer in sensory detail, more emotionally intense, and more consistent over time—all features associated with memories of real experiences rather than with confabulated or hallucinated ones.

Marsh's Anaesthetic Gas Detour

An interesting but ultimately unhelpful section of Marsh's Chapter 9 deals with the historical use of anaesthetic gases—ether and nitrous oxide (N₂O, or "laughing gas")—and the phenomenological experiences they produced. He recounts Oscar Wilde's experience under ether anaesthesia in 1895, in which Wilde reported a diminishing circle of light and a sense of all problems being solved. He cites the work of Ernest Dunbar, whose ether subjects reported racing thoughts, a sense of cosmic unity, and a feeling that all logical relations were laid bare. Eighty percent of surgical patients interviewed by Dunbar recalled rushing into a dark tunnel, hearing singing in their ears, and seeing flashing lights.⁶⁹

Marsh argues that these early anaesthetic experiences have been unjustly neglected by NDE researchers and that they demonstrate the capacity of the brain to generate NDE-like phenomenology under chemical influence. He is right that these accounts are fascinating historical curiosities. But they do not support the chemical hypothesis the way he thinks they do.

First, the experiences induced by ether and nitrous oxide were recognized by the subjects themselves as chemically induced. They did not believe they had left their bodies or visited another realm. They recognized the experience as a drug effect. NDE experiencers, by contrast, are almost universally convinced of the reality of their experience—and this conviction typically deepens rather than fades over time.

Second, the anaesthetic experiences lacked the features that make NDEs evidentially significant: veridical perception of real-world events, encounters with specific deceased persons, acquisition of verifiable information. They were rich in subjective phenomenology but empty of objective content. A woman who hallucinated a Caesarian section under nitrous oxide—and then woke up to deliver her baby naturally—had an interesting experience, but her experience was demonstrably not real. The NDE patient who describes the surgical procedure being performed on her body and gets the details right is in a fundamentally different evidential category.

Third, and most importantly, the availability of these anaesthetic gases is now severely limited in modern medical practice. Marsh himself notes that "modern anaesthetic techniques have now rendered the possibility of retracing those early experiences redundant."⁷⁰ So these historical curiosities, whatever their interest, do not explain why NDEs continue to occur in contemporary medical settings where entirely different drugs and techniques are used.

The "Multi-Factor" Response: Fischer and Mitchell-Yellin's Escape Hatch

Fischer and Mitchell-Yellin offer what initially sounds like a reasonable philosophical argument. They say NDE researchers make the mistake of evaluating each chemical explanation in isolation and then, when each one fails alone, rejecting them all. Fischer and Mitchell-Yellin argue this is like dismissing a gas burner as a cause of a house fire because a gas burner alone does not cause a fire—ignoring the possibility that a gas burner combined with, say, a grease spill and a missing smoke detector might produce one.⁵¹

This is a fair point in the abstract. Multi-factor explanations can be perfectly valid. Nobody disputes that. But the argument has a critical problem: it is an argument from possibility, not from evidence. Fischer and Mitchell-Yellin do not actually present a multi-factor explanation that accounts for the full range of NDE evidence. They simply assert that such an explanation might exist. They argue that the failure of each factor in isolation does not prove that a combination of factors could not succeed. True enough. But they never show us the combination that works.

And here is the deeper problem. The individual failures of each chemical explanation are not just quantitative (each one explains some features but not all). They are qualitative. Each chemical mechanism, when studied in controlled settings, produces experiences that are fundamentally unlike NDEs—not just less complete versions of them, but qualitatively different kinds of experiences marked by confusion, fragmentation, distortion, and disorganization. Combining multiple mechanisms that each produce confused, fragmented experiences does not logically yield a single coherent, hyper-lucid, veridical experience. Mixing several wrong answers together does not produce the right answer.

Think of it this way. If you have a radio that produces static, and you combine it with a second radio that produces static, you do not get a clear signal. You get more static. If hypoxia produces confusion, and CO₂ excess produces bizarre geometric patterns and convulsions, and endorphins produce a pleasant haze—why would combining all three produce the crystal-clear, narratively coherent, veridically accurate, life-transforming NDE?

Veridical Perception: The Unscalable Wall

Every chemical explanation for NDEs crashes into the same immovable obstacle: veridical perception. No chemical mechanism can explain how a patient with no heartbeat, no measurable brain activity, and closed eyes accurately describes specific events happening in the operating room, in other rooms of the hospital, or in locations miles away.

We dealt with the veridical evidence extensively in Chapter 4, and I will not repeat the full catalogue here. But it is essential to note, in this context, that no amount of endorphins, no degree of hypoxia, no level of CO₂, and no combination of neurochemicals can produce accurate perception of real-world events that the patient had no normal way of perceiving. Endorphins do not give you eyes on the ceiling. Hypoxia does not transmit operating room conversations into your unconscious brain. CO₂ does not allow you to describe the shoe on the third-floor windowsill that you have never visited.

The chemical explanations are explanations of the phenomenology of NDEs—the subjective feel of the experience. They are attempts to explain why NDEs feel the way they do. But they cannot explain the informational content of NDEs: the verified facts that patients obtained during their experiences. This is a fatal limitation. An explanation that accounts for how something feels but cannot account for what it reveals is, at best, radically incomplete.

Holden's 2009 analysis in The Handbook of Near-Death Experiences showed that 92 percent of apparently non-physical veridical perception claims during NDEs contained only accurate information, with only 8 percent containing some error and a mere 1 percent being completely wrong.⁵² If these experiences were chemically generated hallucinations, we would expect a random distribution of accurate and inaccurate details. We do not find that. We find overwhelming accuracy. Chemistry cannot account for that.

NDEs and Cardiac Arrest: The Brain Activity Question

I want to address one more crucial piece of the puzzle, because it affects every chemical explanation simultaneously. For a chemical to produce an experience, the brain must be functioning well enough to process that chemical and generate the experience. But during cardiac arrest, the brain's capacity to do this is profoundly compromised.

When the heart stops, blood flow to the brain ceases. Measurable electrical activity in the brain's cortex (the part responsible for conscious experience) typically flatlines within about 10 to 20 seconds—sometimes even faster.⁵³ The EEG goes flat. From the standpoint of every known model of brain function, there should be no conscious experience occurring during this period. Not a confused experience. Not a fragmented experience. No experience at all.

Yet this is precisely when many NDEs appear to occur. Patients report accurate perceptions of events that took place while their hearts were stopped and their brains showed no measurable electrical activity. As the authors of The Self Does Not Die carefully document, the complex, coherent contents of NDEs simply cannot be explained by the kind of brain activity (or lack of it) that is present during acute cardiac arrest.⁵⁴

This is the ultimate problem for every chemical explanation. It is not just that the chemicals produce the wrong kind of experience. It is that during the period when many NDEs occur, the brain should not be producing any experience at all—because the brain is not functioning. You cannot explain an experience in terms of brain chemistry when the brain's chemistry has effectively shut down.

Van Lommel was particularly forceful on this point. In his Lancet study and in his subsequent book Consciousness Beyond Life, he argued that the NDE occurring during cardiac arrest represents a fundamental challenge to the physicalist assumption that consciousness is produced by the brain. If the brain produces consciousness, then no brain activity should mean no consciousness. Yet patients report hyper-vivid consciousness during precisely the period when their brains were flatlined.⁵⁵

Marsh's response to this—that the experiences occur not during the deepest period of cardiac arrest but rather during the approach to unconsciousness or during recovery—is addressed in detail in Chapter 17 of this book, which covers the timing problem. For now, it is enough to note that even if we grant the timing objection (which we should not grant too quickly), a brain that is losing function or regaining function after cardiac arrest is not a brain that should produce heightened lucidity, veridical perception, and the most vivid experience of a person's entire life. A recovering brain should produce confusion and disorientation—which is, in fact, exactly what it does in the overwhelming majority of cardiac arrest survivors who do not report NDEs.

The Ketamine Comparison: Close but No Cigar

Although the ketamine argument is treated in full in Chapter 16, it is worth noting briefly here how it fits into the chemical landscape, since Marsh treats it as part of his Chapter 9 discussion.

Marsh calls ketamine a potential paradigm for understanding NDEs because sub-anaesthetic doses can produce some experiences that superficially resemble NDE features: colored imagery, timelessness, out-of-body sensations, emotional highs.⁵⁶ But the differences are substantial. Ketamine frequently produces disturbing, chaotic, terrifying experiences—nothing like the consistent pattern of peace and love that characterizes most NDEs.⁵⁷ In Barbara Collier's clinical study, 60 percent of patients who received ketamine anaesthesia did not wish to repeat the experience.⁵⁸ NDE experiencers, by contrast, almost universally describe their experience as the most meaningful and positive event of their lives.

Perhaps most tellingly, even Karl Jansen—the psychiatrist who developed the most detailed ketamine model of NDEs—eventually changed his mind. After twelve years of studying ketamine, Jansen wrote: "I now believe that there most definitely is a soul that is independent of experience. It exists when we begin, and may persist when we end. Ketamine is a door to a place we cannot normally get to; it is definitely not evidence that such a place does not exist."⁵⁹

When the leading proponent of the ketamine model of NDEs reverses his own position, that is worth paying attention to.

And again: ketamine cannot explain veridical perception. A chemical that produces hallucinations cannot explain how patients accurately report events they had no normal way of perceiving. Similarity of subjective experience does not prove identity of cause. A person who dreams of flying and a person who actually flies both have the experience of flight—but the mechanisms are utterly different. And only one of them can look down and accurately describe what the landscape looks like.

D. Counter-Objections

Counter-Objection 1: "We Just Haven't Found the Right Chemical Yet"

A skeptic might respond: "You have shown that endorphins, hypoxia, CO₂, and ketamine each fail as complete explanations. But neuroscience is a young field. Perhaps we simply have not yet identified the right neurochemical mechanism. The fact that we cannot explain NDEs chemically today does not mean we never will."

This is a legitimate point, and I want to treat it honestly. It is true that neuroscience is young. It is true that our understanding of brain chemistry is incomplete. And it is true that future discoveries might change the picture.

But there is a difference between scientific humility and a blank check. The "we just haven't found it yet" argument is unfalsifiable. No matter how many chemical explanations fail, the skeptic can always say, "But the next one might work." At some point, the accumulation of failed explanations becomes evidence in its own right—evidence that the phenomenon under investigation may not have a chemical explanation at all.

Moreover, the problem is not merely that we have not found the right chemical. The problem is that the entire category of chemical explanation faces a qualitative mismatch with the evidence. Chemicals that impair brain function produce impaired experiences. NDEs involve enhanced function. This is not a gap that a new chemical is likely to fill. It is a fundamental incompatibility between the proposed mechanism and the observed result.

And no future chemical discovery can explain veridical perception during cardiac arrest. No chemical can give an unconscious patient accurate knowledge of events happening in other rooms. The veridical element is not a problem of finding the right molecule. It is a problem of category: chemicals act on brains, and if the brain is not functioning, chemicals have nothing to act on.

Counter-Objection 2: "The Multi-Factor Explanation Is Still Viable"

Fischer and Mitchell-Yellin might respond: "You are still evaluating each chemical in isolation. We are arguing for a combination. Perhaps hypoxia contributes the tunnel, endorphins contribute the bliss, temporal lobe activation contributes the vivid imagery, and some as-yet-unidentified process contributes the narrative structure. Together, these could produce the full NDE."

This is the strongest version of the chemical argument, and I respect its philosophical sophistication. But it has three problems.

First, as I argued above, it is entirely hypothetical. No one has actually constructed a multi-factor chemical model that accounts for all the features of the NDE. Fischer and Mitchell-Yellin explicitly acknowledge that they do not possess the scientific expertise to build such a model.⁶⁰ They are arguing for a possibility, not presenting an explanation. And in the meantime, the substance dualist explanation—that consciousness actually does separate from the body—accounts for the data as it stands, without requiring a yet-to-be-discovered mechanism.

Second, the multi-factor approach does not solve the qualitative problem. If each factor produces confused, fragmented, distorted experiences individually, why would they produce a coherent, hyper-lucid experience in combination? Fischer and Mitchell-Yellin never address this. They assume that combining partial explanations yields a complete explanation, but they never show how combining fundamentally different kinds of cognitive impairment could yield cognitive enhancement.

Third, the multi-factor approach completely fails to address veridical perception. Even if some future combination of chemicals could theoretically reproduce the subjective feel of an NDE, no combination of chemicals can explain how patients acquire accurate information about real-world events that they had no normal access to. The veridical element is the rock on which every chemical explanation breaks.

Counter-Objection 3: "Marsh's Peripheral Blood Gas Point Is Valid"

A skeptic might press Marsh's argument about Sabom's blood gas case: "Peripheral arterial blood gas levels may not reflect intra-cerebral conditions. Just because the blood in the arm shows normal oxygen does not mean the brain was adequately oxygenated."

This is a technically valid point. There are circumstances in which peripheral blood gas levels can diverge from brain tissue levels. But it cuts both ways, as I noted in Section B. If we cannot trust peripheral blood gas measurements to tell us about brain conditions, then the entire edifice of blood-gas-based arguments for the chemical hypothesis collapses. You cannot use blood gas data to support the hypoxia hypothesis in some cases and then dismiss it as unreliable in the one case where it contradicts you.

Furthermore, in Sabom's case, the patient was receiving high concentrations of supplemental oxygen during CPR, which is standard protocol. The high peripheral oxygen reading was exactly what we would expect in those circumstances. And the patient was simultaneously reporting veridical out-of-body perception—accurate details of what was happening in the room—at the very time his blood showed no evidence of hypoxia.⁶¹ To dismiss this, the skeptic must argue that the brain was somehow oxygen-deprived despite the body being flooded with supplemental oxygen and the blood showing high oxygen levels. That is a heavy burden of proof.

Counter-Objection 4: "Correlation Between Brain States and NDEs Does Not Require Dualism"

A final and more philosophically subtle objection goes like this: "Even if some chemical explanations fail, the mere fact that NDEs often occur in conjunction with altered brain states (cardiac arrest, hypoxia, anaesthesia) suggests a brain-based origin. If NDEs were genuinely out-of-body, why would they be triggered by physical events?"

This is actually a reasonable question, and the answer takes us into one of the most important philosophical insights for understanding NDEs. The correlation between altered brain states and NDEs is entirely compatible with substance dualism—if the brain functions not as the generator of consciousness but as the filter or transmitter of consciousness.

This is the filter model (sometimes called the transmission model), proposed in various forms by William James, Henri Bergson, Aldous Huxley, and more recently by Bernardo Kastrup and the team behind Irreducible Mind.⁶² On this model, the brain does not produce consciousness the way a factory produces a product. Rather, the brain filters, limits, and channels a consciousness that exists independently. When the brain is impaired or shuts down, the filter is removed, and consciousness may actually expand rather than diminish—which is exactly what NDEs consistently show.

Think of it like a radio. When you damage a radio, you change the sound that comes out of it. Static. Distortion. Eventually, silence. A naïve observer might conclude that the radio generates the music. But we know better: the radio receives and transmits a signal that exists independently. The music is not inside the radio. And when the radio breaks, the music does not cease to exist—it is simply no longer being transmitted through that particular device.

On the filter model, the fact that NDEs are triggered by extreme brain events is not surprising. It is exactly what we would predict. When the brain's filtering function is severely disrupted or shut down entirely, consciousness is freed from its usual constraints. The result is not confusion (which is what a malfunctioning generator would produce) but expansion—heightened clarity, vivid perception, access to information beyond the normal senses. This is precisely the pattern we see in NDEs.

We will explore the filter model in much greater depth in Chapter 23, which addresses the hard problem of consciousness. For now, the key point is this: the correlation between brain states and NDEs does not prove that the brain causes NDEs. It is equally consistent with the brain constraining consciousness that exists independently—and the NDE representing what happens when that constraint is temporarily removed.

Counter-Objection 5: "These Are Just Argument-from-Ignorance Claims"

A scientifically minded skeptic might raise a broader charge: "All you are doing is pointing to things we cannot currently explain and filling the gap with 'the soul.' That is an argument from ignorance—the same kind of reasoning people have always used to attribute natural phenomena to supernatural causes before science catches up."

This objection sounds powerful, but it mischaracterizes the argument. The pro-NDE case is not built on gaps in scientific knowledge. It is built on positive evidence. The argument is not: "Science cannot explain NDEs, therefore souls exist." The argument is: "We have specific, verified, documented cases in which patients acquired accurate information about real-world events during periods when their brains were not functioning. This positive evidence is better explained by consciousness existing independently of the brain than by any proposed neurochemical mechanism."

There is a crucial difference between an argument from ignorance and an inference to the best explanation. An argument from ignorance says, "We do not know, therefore God (or the soul, or the supernatural)." An inference to the best explanation says, "We have data that needs explaining, and among the available explanations, this one accounts for the data most completely." The pro-NDE case is of the second kind. We are not saying the soul must exist because science has failed. We are saying that the best available explanation of the full range of NDE evidence—including veridical perception, the selectivity problem, the "realer than real" quality, the fundamental mismatch between chemical effects and NDE features, and the transformative aftereffects—is that consciousness can and does function apart from the physical brain.

Moreover, the skeptical position itself rests on a gap. As we will explore more fully in Chapter 23, physicalism has not explained consciousness. The hard problem—why and how physical brain processes give rise to subjective experience—remains unsolved. The skeptic who says "we just need more science" is making their own argument from ignorance: they are assuming that a future discovery will vindicate physicalism, despite the fact that every current attempt to reduce consciousness to brain chemistry has failed. If arguments from ignorance are bad reasoning, the skeptic should apply that standard to their own position as well.

Conclusion

The chemical explanations for near-death experiences are, taken individually, each unable to account for the full NDE. Taken together as a "multi-factor" hypothesis, they remain hypothetical and face the qualitative problem of explaining how mechanisms that produce confusion can combine to produce clarity. And none of them—singly or in combination—can account for the veridical perception that is the evidential heart of the NDE case.

Marsh is to be credited for his careful treatment of each chemical mechanism. He is more cautious than many skeptics, and he acknowledges the limitations of each explanation more honestly than some. He recognizes that hypoxia alone is insufficient. He admits the endorphin hypothesis has significant gaps. He concedes that CO₂ intoxication produces experiences "notably unlike" NDEs. He notes the "troublesome differences" between ketamine and the NDE pattern. These are honest concessions from a serious scholar. But his caution does not save the argument. Having acknowledged all of these limitations, Marsh still arrives at the conclusion that these mechanisms, taken collectively, account for the NDE. That conclusion does not follow from the evidence he himself presents. If each mechanism fails individually, and Marsh has offered no specific model showing how they succeed in combination, then the collective conclusion is an assertion, not a demonstrated finding.

The chemical explanations are not so much wrong as they are radically incomplete. They may explain fragments of the NDE—perhaps the pleasant feelings, perhaps the tunnel-like visual effects during retinal ischemia, perhaps some of the emotional intensity. But they cannot explain the whole. They cannot explain the coherence. They cannot explain the lucidity. They cannot explain the "realer than real" quality that distinguishes NDEs from every known altered state of consciousness. They cannot explain the transformation that reshapes people's lives for decades afterward. And above all, they cannot explain the veridical perception—the documented, verified cases in which patients accurately reported events they had no normal means of perceiving.

When we step back and survey the full landscape of evidence, a pattern emerges. Every chemical explanation predicts confused, fragmented, distorted experiences. NDEs are coherent, structured, and hyper-lucid. Every chemical explanation applies to all dying brains equally. NDEs occur in only a minority. Every chemical explanation operates within the brain. Veridical NDEs involve perception of events outside the brain's sensory reach. The gap between what the chemical explanations predict and what NDEs actually deliver is not a small gap that future research might close. It is a chasm.

The evidence, when taken as a whole, points in a direction that the chemical hypothesis simply cannot accommodate: that during near-death experiences, consciousness is not being produced by a dying brain. It is being freed from one. And if that is true, then the implications are enormous—not just for neuroscience and philosophy of mind, but for how we understand ourselves, our mortality, and the reality of the world beyond the physical. We will continue to explore those implications in the chapters ahead.