Chapter 14

Phantom Limbs and the “Illusory Body” Argument

A. The Critic’s Argument: Marsh’s Phantom Limb Analogy

One of the most creative arguments Michael Marsh offers against the reality of out-of-body experiences comes not from brain scans or drug studies but from one of the most fascinating corners of neurology: phantom limb syndrome. You’ve probably heard of this before. A soldier loses his arm in combat, and yet for months—sometimes years—afterward, he still feels the arm. He reaches for things with fingers that aren’t there. He scratches an itch on a palm that no longer exists. The brain, it turns out, doesn’t just accept the loss. It keeps generating the experience of the missing limb as if nothing had happened.¹

Marsh finds this deeply significant for the NDE debate. In Chapter 5 of his book, he devotes considerable space to phantom limb phenomenology—what he calls “the neurophysiology of absence.”² He traces the history of the phenomenon back to its first formal description in 1872, when the term was coined to describe the experiences of ninety American Civil War amputees.³ But his interest is not merely historical. He is building a case. And the case goes like this.

If the brain can create a convincing, subjectively real experience of a limb that no longer exists, then it stands to reason that the brain could create a convincing, subjectively real experience of an entire body that exists outside the physical one. Marsh states his argument with characteristic directness: “if a normal brain can elaborate a non-existent limb, or a torso in someone with a broken neck, it could also manufacture a non-existent ‘body’ thereby generating an out-of-body experience” (p. xxi).⁴ In other words, the out-of-body experience is just a phantom limb writ large. It is the brain doing what brains do—constructing an illusion that feels completely real but has no existence outside the skull.

Marsh builds this argument through several layers. First, he walks us through the basic phenomenology. When a person loses a limb through amputation or trauma, the brain’s internal “body map”—its stored representation of the body’s shape and boundaries—doesn’t automatically update to reflect the loss. The cortical territory that once processed signals from that limb remains active. It keeps generating sensations of touch, pain, heat, cold, and movement in a part of the body that no longer exists.⁵ The experience is not vague or dreamlike. Patients report it as vivid, specific, and unmistakably real. They feel individual fingers. They sense the position of their wrist. Some even experience intense, agonizing pain in the phantom limb—a clenched fist that won’t unclench, a cramped muscle that can’t be relieved.

Second, Marsh extends the phenomenon beyond simple limb amputation. He notes that phantom sensations can arise after the removal of a breast, a tooth, an eyeball, or even internal organs like the uterus, bladder, or rectum.⁶ After bladder removal, patients sometimes continue to feel sensations of fullness, urgency, and evacuation—phantom signals from an organ that is no longer there. The brain is indifferent to the actual state of the body. It generates experience based on its internal model, not external reality.

Third—and this is where Marsh gets especially interesting—he describes phantom limb experiences in people who were born without limbs. This category, called congenital limb absence or aphakia, is particularly striking because these individuals have never had sensory input from the missing limb. There was never a physical arm or leg to generate nerve signals in the first place. Yet they still experience phantom sensations. Marsh cites the research of Melzack and Loeser, who reviewed thirty such patients between the ages of five and eight and found that 17 percent of them reported phantom limb experiences.⁷ One ten-year-old girl who was born without the lower two-thirds of her left arm nonetheless experienced a phantom thumb, fingers, and a faint image of a palm. An eleven-year-old girl born without forearms or hands happily used her phantom fingers to help her count and solve arithmetic problems.⁸

Marsh sees these congenital cases as especially powerful evidence. They demonstrate that the brain contains what he calls “some kind of inherited inbuilt engram”—a pre-installed body map that doesn’t depend on sensory experience for its existence.⁹ The brain comes wired to generate the experience of a complete body, whether the actual body matches that map or not. This is crucial for his argument. If the brain has a built-in template for what a body should feel like, and if it can generate convincing experiences of body parts that don’t exist, then an out-of-body experience might be nothing more than the brain activating this template in an unusual way—projecting the sense of “self” to a location outside the physical body.

Fourth, Marsh extends his analogy from phantom limbs to phantom bodies in paraplegia and quadriplegia. When the spinal cord is completely severed, the brain loses all sensory input from the body below the injury. But the brain’s internal map of that body remains intact. Paraplegic patients report feeling phantom sensations in their paralyzed limbs—sensations of position, movement, even pain. Marsh writes that when patients’ eyes are open, the phantom and reality are “fused and perfectly coordinated”; dissociation occurs only when the body is moved without the patient’s knowledge (p. 103).¹⁰ He draws the comparison explicitly: if a quadriplegic patient can experience a complete phantom body below the neck despite having no sensory connection to it, then the brain is clearly capable of generating the experience of a whole body that exists independently of physical input. And if it can do that, it can do what OBE experiencers describe: generate the experience of a body floating above the physical one.

Marsh wraps up his phantom limb discussion by connecting it to the broader theme of illusory consciousness. He writes that phantom limb phenomenology “expresses the overwhelming belief that this is ‘me,’ this is what ‘I’ am actually experiencing and perceiving, despite the imaginary phantom being brain-engineered” (p. 105).¹¹ The subjective sense of “this is real” is no guarantee of objective reality. Phantom limbs feel absolutely real to the patient—just as OBEs feel absolutely real to the experiencer. But the phantom limb is, by definition, an illusion produced by the brain. Marsh’s conclusion is that OBEs are, too.

He then links this argument to his treatment of autoscopy (seeing a mirror image of yourself), heautoscopy (where consciousness seems to alternate between your real body and a phantom double), and the various body-image disturbances he discusses in Chapter 6 on the temporo-parietal junction. The phantom limb is the foundation; the OBE is the final floor. Everything in between—the Pinocchio illusion, migraine auras, body-image distortions, rubber hand experiments—fills out the staircase.¹²

Marsh’s rhetorical strategy here is effective. He does not simply assert that OBEs are brain-generated. He builds the case step by step, from well-established neurological phenomena to more speculative applications. Each step is individually reasonable. Phantom limbs are real. Phantom sensations in paraplegics are real. Congenital phantoms are real. Body-image distortions from migraine, epilepsy, and tendon vibration are real. No serious neuroscientist disputes any of this. And Marsh stacks these bricks so carefully that by the time he reaches the top—the claim that OBEs are essentially phantom body experiences—it feels almost inevitable. The reader who has followed him through phantom arms, phantom torsos, and the Pinocchio effect is primed to nod along when he says that a floating-above-the-body experience is just one more entry in the same catalogue of neural illusions.

It is a well-constructed argument. But well-constructed arguments can still be wrong if they contain a hidden assumption. And Marsh’s argument contains a large one.

To summarize Marsh’s argument as fairly as I can: The brain has a built-in body map. That map can generate the experience of body parts that don’t exist, complete bodies that can’t be felt, and even bodies that were never there in the first place. Since the brain can manufacture these phantom experiences with complete subjective realism, it can also manufacture the experience of being outside the body. The OBE is a phantom body projected to a different location—nothing more, nothing less. No soul required. No consciousness leaving the body. Just the brain doing what brains have always done: constructing an illusion and making you believe it’s real.

It is a clever argument. Genuinely clever. And it deserves a careful response.

B. Identifying Weaknesses in the Phantom Limb Analogy

Let me start by saying something that I mean sincerely: Marsh’s phantom limb discussion is one of the best parts of his book. The neuroscience is solid. The clinical cases are fascinating. His explanation of how the brain constructs body-image from a complex web of cortical maps, proprioceptive signals, and vestibular input is clear and well-documented. If I were teaching an undergraduate neuroscience class, I would assign this chapter. Marsh knows his material, and he presents it well.

But there is a difference between demonstrating that the brain can create illusory body experiences and demonstrating that all OBEs are illusory body experiences. And that is the gap Marsh never closes.

The central weakness in his argument can be stated simply: phantom limbs don’t give you information about the real world.

Think about that for a moment. A man who has lost his left arm experiences a vivid phantom. He feels the fingers. He senses the position of the wrist. He can even try to “move” the phantom hand. But the phantom arm cannot pick up a cup. It cannot turn a doorknob. And here is what really matters for the NDE debate: the phantom arm cannot reach into the next room and tell the man what color shirt the nurse is wearing.¹³

Phantom limbs are experienced by the patient as being located where the missing limb was—or sometimes in a slightly distorted position. They do not provide the patient with new information about the external world. They do not enable the patient to perceive events happening outside their line of sight. They do not allow the patient to see themselves from above, describe the instruments being used during their surgery, or identify the nurse who removed their dentures while they were clinically dead. Phantom limbs are internally generated illusions that refer only to the patient’s own body. They are self-referential. They are closed systems.¹⁴

Veridical OBEs are something categorically different. In a veridical OBE—the kind that forms the core evidence in this book—the patient reports accurate, verifiable information about events occurring in the physical world during a time when their brain was severely compromised or non-functional. They see things. They hear things. And what they see and hear turns out to be true. This is not what phantom limbs do. This is not what any internally generated illusion does.

Marsh is aware that veridical cases exist. He discusses some of them elsewhere in his book. But in his phantom limb chapter, he treats the OBE as though it were only a felt experience—a sensation of floating, a perception of being outside the body—without addressing the evidential cases where that experience included accurate information about the real world. This is the move that makes his argument seem stronger than it is. If you strip away the veridical element, then yes, an OBE looks like it could be a phantom body. But you cannot strip away the veridical element. The veridical element is the evidence.¹⁵

There is a second weakness worth noting. Marsh repeatedly emphasizes that the brain’s body map is “inherited” and “inbuilt”—that it exists prior to sensory experience. This is true. But he draws from it the conclusion that all experiences of embodiment (and disembodiment) must therefore be products of this innate brain mechanism. That does not follow. The fact that the brain has a built-in capacity to represent the body does not mean that every experience of the body—or of being outside the body—is a product of that capacity. We have a built-in visual system too, but not everything we see is a hallucination. The brain’s capacity to construct experience does not mean it only constructs experience. Sometimes it also receives information from the real world. The question is whether OBEs during NDEs are cases of construction or reception.¹⁶

A third weakness: Marsh’s analogy between phantom limbs and OBEs involves a significant disanalogy in the conditions under which they occur. Phantom limbs arise in functioning brains. The cortical maps that generate phantom sensations require an active, metabolically healthy brain to operate. Neuroimaging studies show that phantom limb experiences correlate with measurable activity in the somatosensory cortex, the premotor cortex, and the parietal cortex.¹⁷ These are not dying brains. These are not brains in cardiac arrest. These are brains that are fully operational, just missing their usual input. But many of the most evidentially significant OBEs occur during cardiac arrest, when the brain shows no measurable cortical activity whatsoever. If OBEs are a species of phantom body experience, they should require the same cortical machinery that phantom limbs require. During cardiac arrest, that machinery is offline.¹⁸

Marsh would likely respond by appealing to his broader thesis—that NDE experiences are generated during the moments of recovery, when the brain is regaining function. We address that argument at length in Chapter 10 on the dying brain hypothesis. But even granting that possibility for the sake of argument, it does not solve the problem. A brain that is barely flickering back to life is not in the same state as a healthy brain generating phantom limb experiences. The cortical maps that sustain phantom sensations require organized, coordinated neural activity. A brain emerging from cardiac arrest is not producing organized, coordinated activity. It is producing chaotic, fragmented signals—the kind that produce confusion, delirium, and disorientation. Not the kind that produce detailed, accurate perception of events in the operating room.

There is a fourth weakness that deserves mention. Marsh draws a connecting line from phantom limbs through autoscopy and heautoscopy to full OBEs, treating them as points on a single continuum. But this continuum hides a crucial discontinuity. Autoscopic hallucinations (seeing a mirror image of yourself) and heautoscopic experiences (alternating between your real self and a phantom double) share a key feature with phantom limbs: they are self-referential experiences involving the patient’s own body. The patient sees themselves, in a distorted form, from a distorted perspective. The content of the experience is limited to the patient’s own body and its immediate surround. Veridical NDE-associated OBEs break this pattern entirely. They involve perception of other people, other objects, and other events that have nothing to do with the patient’s body image. The continuum that Marsh draws is not, in fact, continuous. There is a categorical break at the precise point where veridical perception enters the picture.

Fifth, Marsh’s argument contains an unexamined assumption about the direction of explanation. He observes that the brain can produce phantom experiences. He then infers that OBEs are phantom experiences. But this inference works only if we already assume that all experiences of being outside the body are brain-generated. If some OBEs are veridical—if they involve accurate perception of real events—then they are not phantom experiences at all, and the analogy fails. Marsh is using an analogy to support a conclusion that the analogy already assumes. This is subtle, but it is important. The phantom limb analogy does not provide independent evidence that OBEs are illusory. It only illustrates what illusory OBEs would look like if they existed. Whether the actual OBEs reported in NDE research are illusory is a separate question that requires separate evidence—evidence that Marsh does not provide in this chapter.

C. The Pro-NDE Response: Why the Phantom Body Analogy Fails

1. The Veridical Problem: Phantom Limbs Don’t See

Let me tell you about a case. During the Lancet study conducted by Dutch cardiologist Pim van Lommel and his colleagues—one of the largest prospective studies of NDEs ever undertaken—a man was brought into the hospital in cardiac arrest, comatose, cyanotic (blue from lack of oxygen), and clinically dead. A nurse named TG was part of the resuscitation team. During the effort to intubate the patient, TG discovered that the man had dentures. He removed them and placed them on the “crash cart”—a wheeled cart with supplies, bottles, and a sliding drawer underneath. Extensive CPR continued for about an hour and a half before the patient’s heart rhythm was restored.¹⁹

More than a week later, TG encountered the patient again on the cardiac ward. The moment the man saw TG, he said: “Oh, that nurse knows where my dentures are.” He went on to describe exactly what had happened: TG had taken his dentures out of his mouth and placed them on the crash cart, in the sliding drawer underneath. He described the room where he had been resuscitated, the people present, and what they looked like. He reported seeing all of this from above—watching his own body being worked on while he was, by every medical measure, clinically dead.²⁰

Researchers Rudolf Smit and Titus Rivas later tracked down TG and conducted extensive follow-up interviews, carefully corroborating the details of the case. Their investigation, published in the Journal of Near-Death Studies, confirmed the accuracy of the patient’s account. Skeptic Gerald Woerlee attempted to explain the case away but, as Smit and Rivas documented, his counter-explanations required him to ignore or distort the established facts.²¹

Now. Try to explain that with a phantom limb.

A phantom limb generates the subjective sensation of a body part. It does not generate accurate perceptions of external events. The “dentures man” did not merely feel as though he were outside his body. He saw specific, verifiable details about his surroundings—the nurse’s identity, the crash cart, the sliding drawer, the room, the people—while his brain was in a state that should have made any conscious experience, let alone accurate perception, impossible by every standard model of neuroscience.²²

This is the problem that Marsh’s analogy simply cannot solve. Phantom limbs are about the patient’s body. Veridical OBEs are about the world. These are fundamentally different categories of experience, and no amount of neurological sophistication bridges the gap between them.

2. The Scope of the Veridical Evidence

The dentures case is far from an isolated example. In 2007, NDE researcher Janice Holden conducted the most comprehensive review ever attempted of apparently veridical perceptions during NDEs. She identified 107 cases from thirty-nine different publications by thirty-seven different researchers or research teams, covering the period from 1975 onward. Using the most stringent criteria—classifying a case as inaccurate if even a single detail failed to correspond to reality—Holden found that only 8 percent involved any inaccuracy whatsoever. By contrast, 37 percent were independently verified as completely accurate by objective sources.²³

As Chris Carter pointedly observes, these results pose a real problem for anyone who wants to classify veridical OBEs as hallucinations. Holden herself remarked that these findings “certainly call into question how an allegedly hallucinatory phenomenon could produce only 8 percent of cases with any apparent error whatsoever and 37 percent of cases with apparently completely accurate content that had been objectively verified.”²⁴ If OBEs are just phantom body experiences—internally generated illusions with no connection to external reality—then this accuracy rate is inexplicable.

Cardiologist Michael Sabom took this a step further. He tested the alternative hypothesis directly. Twenty-five cardiac patients who had not had NDEs were asked to guess what had happened during their resuscitations. Eighty percent made at least one major error. Meanwhile, patients who reported veridical OBE perceptions gave descriptions that were not only accurate but specific to their own procedures—details that would not have correctly described another patient’s resuscitation.²⁵ One patient, for instance, reported seeing doctors give him “a shot in the groin.” This was accurate for his procedure but was not done in other cases. If this were guesswork or confabulation based on general hospital knowledge, we would expect generic descriptions, not procedure-specific ones.

Penny Sartori’s five-year prospective study at a Welsh intensive care unit produced similar results. Patients who reported OBEs gave significantly more accurate descriptions of their resuscitation events and equipment than patients who were resuscitated but did not have OBEs. She concluded that “this research has demonstrated that those who reported OBEs gave more accurate descriptions of events and equipment used than those who were resuscitated, but did not report a NDE/OBE. This lends further support to the possibility of consciousness existing apart from the brain.”²⁶

Van Lommel’s own analysis of his research found that among NDErs who reported veridical perceptions, 92 percent were completely accurate, 6 percent contained some minor error, and only 1 percent were completely wrong.²⁷ These are not the numbers you get from internally generated illusions. These are the numbers you get from perception.

Let me put this in perspective. If you asked one hundred people to describe what happened in a room they had never been in, during a procedure they were unconscious for, you would expect the vast majority to be wrong. You would expect wild guesses, generic descriptions, and a handful of lucky hits. You would not expect 92 percent accuracy. You would not expect patients to identify specific people, specific objects, and specific actions with the precision that NDE researchers have documented. The accuracy rates in the veridical NDE literature are not consistent with guessing. They are not consistent with confabulation. They are not consistent with phantom body projections or any other internally generated illusion. They are consistent with one thing: perception of real events.

3. Cases That Break the Phantom Body Model

Let me walk through several more cases that demonstrate why the phantom body analogy fails. Each one involves veridical perception of the external world during a time when the brain should not have been capable of any such thing.

In a case documented by Rivas, Dirven, and Smit in The Self Does Not Die, Sue Saunders, a respiratory therapist at Hartford Hospital in Connecticut, assisted in the emergency room with a difficult resuscitation. The patient had lost consciousness and had no heartbeat. Saunders administered oxygen, wearing her yellow work smock with a mask over her face. She left partway through to attend to another duty. A few days later, the resuscitated patient greeted her at the ICU with the words: “You looked so much better in your yellow top.” He described her smock, the mask, and what she had been doing—all while he was clinically dead.²⁸ This is not a phantom limb phenomenon. This is a man with no heartbeat accurately identifying a specific person, her clothing, and her actions during his own resuscitation.

Consider the case of Dr. Manuel Sans Segarra’s patient in Barcelona. The patient was undergoing surgery to remove an esophageal tumor when he went into cardiac arrest. Dr. Sans performed hand cardiac massage and the team administered adrenaline directly into the heart and used defibrillation to restore rhythm. After recovery, the patient described seeing his own body with his thorax opened. He described his cardiac arrest, the hand massage, and the injection into his heart. No one had told him about the cardiac arrest before he shared his experience.²⁹

Or take the case from East Lancashire documented in The Self Does Not Die, where a cardiac arrest patient on the ward was resuscitated by the head nurse. During the resuscitation, the nurse fumbled, dropping a kidney tray containing a full syringe of cardiac stimulants while the attending physician chided him for his clumsiness. Three days later, back on the ward, the patient described the incident to the nurse in detail—the dropped tray, the spilled syringe, the physician’s sharp words. The patient had been in cardiac arrest when this occurred. He had no pulse. He had no measurable brain function. Yet he reported events that were specific, embarrassing, and accurate.³⁰

And then there is the remarkable case of J.S., documented by neurologist Mario Beauregard. This 31-weeks-pregnant woman underwent emergency surgery for an aortic dissection. She was anesthetized, her eyes were taped shut, and she was placed on the operating table. During a deep hypothermic circulatory arrest lasting fifteen minutes—during which her heart was deliberately stopped and her body was cooled well below normal temperature—she had an OBE. Afterward, she described watching a nurse hand medical instruments to the surgeon and identified the anesthesia and ultrasound equipment positioned behind her head. Beauregard’s team verified her descriptions: they were correct.³¹ Her eyes were taped shut. She had never seen the surgical team or the equipment. Her heart was stopped. Her body was hypothermic. And she accurately described what was happening in the operating room.

I could go on. The Self Does Not Die catalogs dozens of additional cases across multiple categories of veridical perception during cardiac arrest and other conditions seemingly incompatible with consciousness. Some involve patients perceiving events in the resuscitation room. Others involve patients perceiving events in entirely different rooms or floors of the hospital. Still others involve patients perceiving events outside the hospital altogether. In every category, the same pattern holds: the patient reports specific, detailed information that turns out to be accurate, obtained during a period when their brain should have been incapable of any conscious experience.

What makes these cases so devastating for Marsh’s phantom body model is not any single case taken in isolation. It is the cumulative weight of the evidence. One case might be a lucky guess. Two might be a coincidence. But when you have dozens upon dozens of verified cases spanning multiple decades, multiple countries, multiple research teams, and multiple medical contexts—all showing the same pattern of accurate perception during clinical death—the probability that all of them are products of a brain-generated phantom body shrinks to a vanishing point. At some stage, the most parsimonious explanation is the simplest one: these patients perceived what they perceived because their consciousness was actually operating apart from their physical brains.

None of these cases can be explained by phantom limb phenomenology. A phantom body, by Marsh’s own account, is a brain-generated projection of the patient’s innate body map. It is a felt sense of embodiment projected to another location. What it is not is a perceptual system. It does not see. It does not hear. It does not gather information about yellow smocks, dropped syringes, surgical instruments, or crash cart drawers. The phantom limb model explains sensations. It does not explain perceptions. And the NDE evidence includes perceptions—accurate, verified, often astonishing perceptions—that the phantom body model simply cannot account for.

4. The Filter/Transmission Model: A Better Fit for Both Phantom Limbs and OBEs

There is, I believe, a better way to make sense of both phantom limbs and veridical OBEs—one that does justice to Marsh’s neuroscience without surrendering the NDE evidence.

Marsh assumes—as most materialist neuroscientists do—that the brain produces all conscious experience. On this model, called the production hypothesis, consciousness is generated by neural activity the way steam is generated by a kettle. Damage the kettle, and you get less steam. Kill the kettle, and you get no steam at all. On this view, phantom limbs are produced by cortical maps that persist after the limb is lost. OBEs are produced by cortical maps that malfunction during metabolic stress. Everything is brain-generated. Everything is illusion.

But there is an alternative model with a long philosophical pedigree. William James articulated it in his 1898 Ingersoll Lecture at Harvard. The philosopher Ferdinand Schiller and the Nobel laureate Henri Bergson developed it independently. It is called the filter or transmission hypothesis, and it proposes that the brain does not produce consciousness but rather transmits and filters it—the way a prism transmits and filters light, or the way a television set receives and transmits a broadcast signal.³²

James pointed out that when we observe a correlation between brain states and mental states, we are observing concomitant variation—when one changes, the other changes too. But correlation does not tell us the direction of causation, and it certainly does not tell us whether the relationship is one of production, transmission, or permission. James wrote that “we are not required to think of productive function only; we are entitled also to consider permissive or transmissive function.”³³ The brain may be the instrument through which consciousness operates in the physical world, not the factory where consciousness is manufactured.

As psychologist Cyril Burt summarized: “The brain is not an organ that generates consciousness, but rather an instrument evolved to transmit and limit the processes of consciousness and of conscious attention so as to restrict them to those aspects of the material environment which at any moment are crucial for the terrestrial success of the individual.”³⁴

On this model, phantom limbs make perfect sense. The brain’s body map is part of the filtering and transmission apparatus. When a limb is lost, the map doesn’t update because the map is a built-in feature of the biological instrument. The instrument keeps generating the sensation of the missing part because the cortical representation of that part is still intact. The phantom limb is a malfunction of the instrument, not a proof that consciousness is produced by the instrument.

And on this model, veridical OBEs also make sense—but in a very different way. If consciousness is not produced by the brain but only normally filtered through it, then when the brain’s filtering function is severely disrupted (as in cardiac arrest), consciousness might become less constrained rather than extinguished. It might expand beyond the body rather than shutting down. And it might perceive things that the brain’s normal filtering would ordinarily block from awareness. This is precisely what NDE research suggests: patients report expanded awareness, heightened clarity, and accurate perception of events they could not have perceived through normal sensory channels.³⁵

Chris Carter makes this point powerfully in Science and the Near-Death Experience. He notes that neuroscientist Gary Schwartz of the University of Arizona has drawn a telling analogy: the same methods used to study the brain—correlation studies, stimulation studies, and ablation studies—could be applied to a television set with parallel results. Damage the TV, and the picture degrades. Stimulate certain circuits, and you get specific effects. But no one concludes from this that the television creates the signal. The evidence is equally consistent with reception and transmission.³⁶

Marsh never considers this alternative. His entire phantom limb argument assumes the production model. He assumes that because the brain can generate illusory body experiences, all body experiences are brain-generated. But if the brain is a receiver rather than a generator, then phantom limbs are a transmission artifact (the receiver continues to output based on its hardware configuration even when the input signal from the limb has stopped), while veridical OBEs represent moments when the receiver is so degraded that it stops filtering consciousness, allowing perception to operate beyond its normal boundaries.

Think of it this way. When your car radio has a damaged antenna, you get static, distortion, and fragments of signal. That is what a dying brain typically produces: confusion, fragmented imagery, delirium. But what if the radio breaks in a very specific way—say, by losing the tuner that normally restricts it to a single frequency? In that case, instead of static, you might briefly pick up signals from multiple stations at once. The output would be richer, not poorer. Stranger, not simpler. And it might contain information you would never normally receive.

That is what veridical NDEs look like. The brain is breaking down. The normal filtering mechanisms are failing. And instead of producing chaos, the result is a period of extraordinary clarity during which the patient perceives things they should not be able to perceive. The production model predicts chaos. The filter model predicts exactly what NDE researchers have documented.

Both phenomena are real. Both involve the brain. But they have fundamentally different relationships to the external world. The production hypothesis can explain phantom limbs but not veridical OBEs. The transmission hypothesis can explain both.

5. The Brain-State Problem: OBEs During Cardiac Arrest

I want to press this point further because it is so important. Marsh’s phantom limb analogy requires a functioning brain. The cortical maps that generate phantom sensations are located in specific, identifiable brain regions—the somatosensory cortex, the premotor cortex, and the parietal association areas.³⁷ These maps require metabolic energy. They require blood flow. They require oxygen. They require glucose. Without these, the maps go silent. The phantom limb disappears along with everything else when the brain shuts down under general anesthesia.

During cardiac arrest, the brain is not merely impaired. It is, for practical purposes, offline. The cortex flatlines on EEG within an average of about fifteen seconds after the heart stops.³⁸ As cardiologist Sam Parnia has documented, all brainstem reflexes are lost immediately after cardiac arrest and do not typically return, even with CPR, until after the heart has been restarted.³⁹ Van Lommel emphasized the point: “It seems rational to assume that all 562 survivors of cardiac arrest in several recently published prospective studies on NDE should have had a flat EEG, because no patient had been resuscitated within 20 seconds of cardiac arrest onset.”⁴⁰

The question is straightforward. If OBEs are a form of phantom body experience generated by cortical body maps, how are they being generated when the cortex is flatlined? Phantom limbs require an active cortex. Cardiac arrest produces an inactive cortex. You cannot have the first without the second. Yet veridical OBEs during cardiac arrest are among the best-documented phenomena in NDE research.

Marsh might appeal to the possibility that some residual brain activity, below the threshold of EEG detection, persists during cardiac arrest. This is technically true—a flat EEG does not absolutely rule out all neural activity at every level. But as van Lommel carefully explains, the issue is not whether there is any electrical flicker anywhere in the brain. The issue is whether there is enough organized, coordinated activity across the cortex, brainstem, hippocampus, and thalamus to support conscious experience—let alone the vivid, detailed, accurate perceptions that NDErs report. The answer, according to everything we know about neuroscience, is no.⁴¹

There is also the telling fact that the experiences people have during cardiac arrest are the opposite of what a degraded brain produces. Hypoxia—oxygen deprivation to the brain—causes confusion, disorientation, and fragmented thinking. It does not cause the hyper-lucid, structured, coherent experiences that NDErs describe. Patients recovering from cardiac arrest without NDEs are typically groggy, confused, and amnesic. Patients who had NDEs report experiences that are, as many describe them, “more real than real.”⁴² This is not the signature of a malfunctioning cortical body map. This is something else entirely.

6. The Specificity Problem: Why Phantom Bodies Don’t Explain What OBEs Contain

Let me push the analogy one more step to show just how far it stretches before it breaks.

Phantom limbs have a consistent phenomenological profile. They are experienced as being in or near the position where the missing limb was. They involve proprioceptive sensations (position, movement, tingling, pain) rather than exteroceptive ones (vision, hearing). They do not include the perception of external objects or events. A man with a phantom arm does not “see” with his phantom hand. He feels it. The experience is entirely somatic—it is about the body, not about the world.

OBEs during NDEs have a radically different profile. The experiencer reports perceiving from an elevated vantage point, typically above the body. They see their own physical body below, often during medical procedures. They hear conversations among medical staff. They observe specific actions—instruments being used, items being placed, people entering and leaving. They accurately report details that are subsequently verified by medical personnel or other witnesses.⁴³

Notice the qualitative difference. Phantom limb experiences involve a narrow range of somatic sensations: touch, pressure, position, pain. NDE-associated OBEs involve a full range of perceptual experiences: vision (often described as sharper than normal), hearing (often described as clearer than normal), spatial orientation, and in many cases a remarkable capacity to attend to and remember fine details that surprise even the medical professionals involved. The experiencer doesn’t just feel a disembodied limb. They see the crash cart. They hear the physician’s words. They watch the nurse fumble with the syringe. The modalities involved are entirely different from anything in the phantom limb literature.

The content of phantom limb experiences is limited to the body itself—the felt shape, position, and sensations of the absent part. The content of veridical OBEs extends to the entire surrounding environment and includes information that the patient demonstrably could not have obtained through any normal sensory channel. Asking the phantom body model to explain veridical OBEs is like asking a thermostat to explain television. Both involve electrical signals. Both involve a kind of “detection.” But the outputs are categorically different, and the analogy collapses under the weight of the difference.

7. Ramachandran’s Own Work Undermines the Analogy

It is worth noting that the neuroscientist whose research Marsh draws on most heavily for his phantom limb discussion—V. S. Ramachandran—is himself cautious about the philosophical implications of phantom limb research. Ramachandran’s classic work Phantoms in the Brain, co-written with Sandra Blakeslee, explores the extraordinary range of phantom experiences: phantom pain, phantom movement, the “telescoping” of phantom limbs, and the famous mirror box therapy for relieving phantom pain.⁴⁴

Ramachandran has shown that phantom limbs arise from the brain’s neural plasticity—its ability to reorganize itself when deprived of sensory input. When a hand is amputated, the cortical region that formerly processed hand sensations doesn’t just go dark. Neighboring cortical regions expand into the vacated territory. This remapping produces strange cross-wiring effects: stimulating the face can produce phantom sensations in the missing hand because the face and hand representations are adjacent on the cortical body map.⁴⁵

But Ramachandran is a careful thinker. He does not extend his findings to claim that all experiences of disembodiment are neural artifacts. His work demonstrates that the brain is remarkably creative in constructing body representations. It does not demonstrate that the brain is the sole source of all consciousness or that experiences occurring during cardiac arrest are merely cortical remapping events. These are separate claims that require separate evidence. Marsh treats them as if they follow naturally from each other. They do not.

There is also a further point worth making. Ramachandran’s phantom limb research focuses on how the brain represents the body to itself. The cortical remapping he documents is an internal process—the brain adjusting its own models in response to changed input. At no point in Ramachandran’s research does a phantom limb enable the patient to perceive something in the external environment. The mirror box therapy that Ramachandran pioneered works precisely because it provides an external visual input to trick the brain into updating its internal model. The brain, left to itself, is stuck with its phantom. It cannot reach beyond itself. This is exactly the limitation that veridical OBEs appear to transcend.

8. What Phantom Limbs Actually Support

Here is an irony that Marsh may not have anticipated. Phantom limb research, rather than undermining the case for consciousness beyond the brain, may actually support it—in a way that phantom limb researchers rarely discuss.

Consider the congenital phantom limbs that Marsh himself highlights—the cases where children born without limbs nonetheless experience phantom sensations. As Marsh notes, these children have never had sensory input from the missing limb. There was never a physical arm to generate nerve signals. Yet they experience a phantom arm complete with fingers and a palm. Marsh interprets this as evidence of an “inherited inbuilt engram”—a genetically encoded body map.⁴⁶

Fair enough. But consider what this means from a broader philosophical perspective. The brain contains a template for a body that was never physically present. It generates the experience of something that does not exist in the material world. The patient experiences a reality that has no physical counterpart. Now, Marsh uses this to argue that the brain is capable of wholesale illusion. But a dualist could equally note that the brain appears to be wired to represent something beyond the immediate physical body—that there is an inherent “reach” to consciousness that extends beyond what is physically available. The innate body map suggests that the brain is designed to interface with something more than just the current state of the physical body.⁴⁷

I don’t want to overstate this point. It is suggestive rather than conclusive. But it is worth noting that the same evidence Marsh uses to support his materialist reading has a plausible dualist reading as well. The brain’s body map could be evidence that the brain is a generator of illusion—or it could be evidence that the brain is an interface between the immaterial self and the physical body, pre-wired for a relationship that transcends the merely physical.

9. Fischer and Mitchell-Yellin’s Related Approach

It is worth briefly noting that Fischer and Mitchell-Yellin, the philosophers whose Near-Death Experiences: Understanding Visions of the Afterlife we engage throughout this book, take a different but related approach to the OBE problem. They do not build their case on phantom limbs specifically, but they do argue that the subjective experience of being outside the body can be explained without invoking an immaterial consciousness. In their treatment of the dentures case, for instance, they suggest that the patient might have constructed his visual experience from nonvisual sources—piecing together sensory fragments registered while unconscious, combined with information obtained after regaining consciousness in the hospital.⁴⁸

Their specific proposal about the dentures man is that he might have felt his dentures being removed and heard the drawer opening, registered these tactile and auditory cues unconsciously, seen similar equipment during his hospital stay, and then retroactively assembled all of this into a coherent visual narrative that happened to be accurate. They call this “incredibly felicitous” but not impossible.⁴⁹

The problems with this should be obvious. It requires multiple independent lucky guesses to converge into a single, detailed, accurate account. It requires unconscious sensory registration at a level of detail and integration that is itself extraordinary, given that the patient was in cardiac arrest. It requires that a man with no heartbeat, no blood pressure, and no measurable brain function somehow registered and stored tactile and auditory information that he later assembled into a visual scene—including the correct identification of a specific nurse he had never consciously met. And it must be applied not just to this one case but to the dozens of similar cases in the literature. At some point, “incredibly felicitous” becomes “incredibly implausible.”⁵⁰

Fischer and Mitchell-Yellin themselves acknowledge the force of the veridical cases, writing that they “appear to pose a powerful challenge to the claim that the mind is entirely physical.”⁵¹ Their subsequent attempt to explain these cases away relies on the same kind of ad hoc speculation that characterizes Marsh’s phantom body approach: the evidence looks like consciousness is operating apart from the brain, but maybe, just maybe, there is a complicated alternative explanation that doesn’t require us to revise our physicalist assumptions. The question is whether this “maybe” is genuinely more plausible than the straightforward reading of the evidence. I don’t think it is.

D. Counter-Objections

“But phantom limbs show that the brain creates compelling illusions—so OBEs could be illusions too.”

This is the most common version of Marsh’s argument, and it deserves a direct response. Yes, the brain can create compelling illusions. No one disputes this. Phantom limbs are real. Rubber hand illusions are real. The Pinocchio effect is real. The brain’s capacity for constructing experience is extraordinary. I have acknowledged this multiple times, and I mean it.

But the argument doesn’t stop there. The fact that the brain can create illusions does not mean that every experience is an illusion. We don’t conclude from the existence of mirages that no one has ever really seen water. We don’t conclude from the existence of optical illusions that vision is unreliable. The question is not whether the brain can produce false experiences but whether it is producing false experiences in this particular case. And the veridical evidence tells us that in NDE-associated OBEs, the experience frequently turns out to be accurate. Phantom limbs are illusions; they have no correspondence to external reality. Veridical OBEs are not illusions; they have verified correspondence to external reality. Saying “the brain creates illusions, therefore OBEs are illusions” is logically equivalent to saying “cameras sometimes malfunction, therefore no photograph is reliable.” The existence of error does not prove that accuracy is impossible.⁵²

“Maybe the veridical information was acquired through normal channels—hearing, residual vision, or post-hoc reconstruction.”

This objection applies to some cases more than others, and researchers have been careful to address it. In the dentures case, the patient was in deep cardiac arrest throughout the resuscitation—no heartbeat, no blood pressure, and cyanotic. TG, the nurse, confirmed in later interviews that the patient was completely unconscious and unresponsive during the entire procedure.⁵³ In the case of J.S., the patient’s eyes were taped shut and she was under general anesthesia during deep hypothermic circulatory arrest.⁵⁴ In case after case, the conditions under which the veridical information was allegedly acquired are conditions that should have precluded any sensory input whatsoever.

The post-hoc reconstruction argument has been tested and found wanting. Greyson’s research demonstrated that NDE memories remain remarkably stable over decades—unlike ordinary memories or confabulated memories, which degrade and change over time.⁵⁵ Furthermore, in many cases, patients reported their experiences immediately upon regaining consciousness, before there was any opportunity for reconstruction. And the level of specific, accurate detail reported—the identity of particular nurses, the description of particular instruments, the location of particular objects—goes far beyond what any post-hoc guessing could reliably produce.

“If consciousness can operate apart from the brain, why do brain injuries affect consciousness?”

This is a fair question, and it is one that Fischer and Mitchell-Yellin press hard. Their argument is that if consciousness is nonphysical, we should not expect brain damage to impair it. But brain damage clearly does impair consciousness. Therefore, they argue, consciousness must be a product of the brain.⁵⁶

The transmission model answers this directly. If the brain is an instrument through which consciousness operates in the physical world, then damage to the instrument will impair its operation—just as damage to a television impairs the reception and display of the broadcast signal without destroying the signal itself. As Carter puts it, the production hypothesis and the transmission hypothesis are equally compatible with the observed effects of brain damage. The effects of drugs, disease, and injury on consciousness are exactly what we would expect if the brain is the medium through which consciousness manifests itself in the material world. The difference is that only the transmission hypothesis can also account for the cases where consciousness appears to operate more effectively during severe brain compromise—a phenomenon that the production hypothesis cannot explain at all.⁵⁷

The phenomenon of terminal lucidity is relevant here. These are cases—documented repeatedly in the medical literature—where patients with severe brain damage from Alzheimer’s, tumors, or other diseases suddenly regain full mental clarity shortly before death. The brain is more damaged than ever. Yet consciousness briefly shines through with full force. On the production model, this is inexplicable. On the transmission model, it makes sense: the brain’s filtering function is breaking down, and as the instrument fails, what it was filtering is briefly released.⁵⁸

“Isn’t the transmission model just unfalsifiable speculation?”

Some critics dismiss the filter/transmission model as unscientific because it seems unfalsifiable—you can always say “the brain is just filtering” in response to any neurological finding. But this is not accurate. The production hypothesis generates a clear prediction: when the brain is severely impaired or shut down, consciousness should be correspondingly impaired or absent. This prediction is falsified by veridical NDEs during cardiac arrest. The transmission hypothesis generates a different prediction: when the brain is severely impaired, consciousness might become less constrained, potentially enabling perception beyond the body’s normal boundaries. This prediction is confirmed by veridical NDEs during cardiac arrest.⁵⁹

As Carter explains, the production hypothesis has been “proved false by the data.” The transmission hypothesis can accommodate the facts that refute the production theory and has the additional advantage of explaining phenomena—like terminal lucidity, veridical OBEs, and the hyper-lucidity of NDEs—that remain utterly mysterious under the production model.⁶⁰

“The sample sizes are too small to draw conclusions.”

A skeptic might argue that the number of rigorously verified veridical OBE cases is still too small to warrant any philosophical conclusions about the nature of consciousness. After all, we are talking about dozens of strong cases, not thousands. Shouldn’t we wait for more data before abandoning the standard physicalist model?

This objection has a surface plausibility, but it confuses two different types of evidence. In statistical research, you need large sample sizes to detect subtle effects. But veridical OBEs are not a subtle effect. They are a categorical anomaly. If the production model is correct, there should be zero cases of accurate perception during flat-EEG cardiac arrest. Not a few. Not a handful. Zero. The claim is that consciousness depends entirely on organized brain activity. If that claim is true, then consciousness during the absence of organized brain activity is impossible—not merely unlikely. Even one well-verified case is enough to falsify an absolute claim. And we have far more than one.

To use an analogy: if someone claims that no crow is white, you do not need a thousand white crows to prove them wrong. You need one. The veridical NDE cases are the white crows of consciousness research. Their significance lies not in their quantity but in their very existence.

Conclusion

Marsh’s phantom limb argument is one of the most intellectually engaging challenges to the reality of OBEs. He has done excellent work explaining how the brain constructs body-image, how phantom limbs arise from innate cortical maps, and how body-image disturbances can produce experiences of disembodiment. All of this is real neuroscience, and it deserves respect.

But the argument fails where it matters most. Phantom limbs are internally generated illusions that tell us nothing about the external world. Veridical OBEs are experiences that provide specific, accurate, independently verified information about the external world—information obtained during states when the brain, according to every mainstream neuroscientific model, should be incapable of producing any conscious experience at all. The phantom body model explains the feeling of being out of body. It does not explain the seeing—the accurate perceiving of real events by a patient whose brain has flatlined.⁶¹

The analogy between phantom limbs and OBEs is clever. But cleverness is not the same as adequacy. And when the full range of veridical NDE evidence is laid on the table, the phantom body explanation simply cannot carry the weight.

What we have seen in this chapter is a pattern that will recur throughout this book. A skeptical critic identifies a real neurological phenomenon. He explains the mechanism accurately and thoroughly. He then extends the mechanism to cover NDE-associated OBEs. The extension looks reasonable, as long as you consider only the subjective experience of disembodiment. But the moment you bring the veridical evidence into view—the moment you ask not just “Did the patient feel like they were outside their body?” but “Did the patient accurately perceive events occurring in the physical world while clinically dead?”—the neurological mechanism falls short. It explains the form of the experience but not its content. It explains the sensation of floating but not the accurate identification of a nurse’s yellow smock. And the content, the accuracy, the verified reality of what these patients perceived—that is the evidence that demands a different explanation.

The phantom limb teaches us something true and important about the brain. The brain constructs experience. It maintains models. It generates felt realities that may not correspond to physical facts. All of this is true, and none of it threatens the NDE evidence. Because the NDE evidence is not about what the brain constructs. It is about what consciousness perceives when the brain’s construction equipment has been shut off. And the fact that it perceives anything at all—let alone perceives it accurately—is the finding that Marsh’s phantom limb analogy was never designed to explain.

In the next chapter, we turn to another neurological argument against NDEs—the claim that they are nothing more than dreams, hallucinations, or altered states of consciousness produced by a stressed brain. As we will see, this argument faces many of the same problems we have identified here: it can explain the subjective quality of the experience but not its verified connection to external reality.