Chapter 17

The Timing Problem — When Do NDEs Actually Occur?

Picture a courtroom. A witness takes the stand and describes—in calm, vivid detail—what she saw at the scene of an accident. The defense attorney rises. He doesn’t challenge whether she saw it. He doesn’t challenge what she described. He challenges when she saw it. “Yes,” he says, “you remember it. But maybe the memory was actually formed later. Maybe what feels to you like a real-time perception was stitched together moments after the fact, as your mind caught up.”

That, in essence, is the timing objection to near-death experiences. Critics rarely deny that the experiencer remembers something. They rarely deny that the contents of the experience match the operating room. What they deny is the timing. They say: maybe it didn’t happen when it seemed to happen. Maybe it happened during a brief residual flicker of brain activity. Or during the slow climb back to consciousness. Or during the gradual ramp-up of memory as the brain came back online.¹

It’s a clever objection. If it works, it strips the dualist case from veridical NDEs of much of its force. The whole point of cases like Pam Reynolds’s, the dentures man’s, or the AWARE Southampton case is that the experience occurred at a moment when the brain could not, on materialist assumptions, generate it. Slide the experience into a window of recovering brain function, and the puzzle dissolves. So everything depends on whether the timing escape route actually fits the data.

It does not. In this chapter we will see, step by step, why.

A. The Critic’s Argument

Michael Marsh raises the timing question most pointedly in his treatment of the Pam Reynolds case (we cover that case in full in Chapter 5). His basic move is this: Reynolds’s body was cooled to 60°F, her heart was stopped, blood was drained from her head, and her EEG went flat. Then she was warmed back up. Marsh argues that her near-death experience could not have occurred during the deep-cold standstill phase, because at that temperature her brain “would have been too cold to have engendered the events experienced and, more importantly, to have set down the necessary memory traces for later recall.”² So when did it happen? Marsh proposes a window between roughly 1 p.m. and 2 p.m.—during rewarming, when her brain was “still warming up” and could plausibly support some kind of mental life.³ The experience, Marsh insists, “came about not because her brain function was impaired by an acute loss of blood pressure, pulse and cerebral circulation, but because her brain was still warming up.”⁴

That move—relocating the experience from the period of impaired or absent brain function to a period of recovering brain function—is the heart of the timing objection.

John Martin Fischer and Benjamin Mitchell-Yellin develop the same strategy at length. The whole of their chapter 3 is titled, fittingly, “When Exactly Do Near-Death Experiences Take Place?” They argue that the apparent timing of an experience can diverge sharply from the actual timing. A dream may feel like hours; the brain activity behind it may have lasted only seconds, in the moments before waking. Likewise, they suggest, a person who reports a near-death experience that seemed to occur during cardiac arrest may have actually had the experience “at some later time,” perhaps as the brain was “coming back on line.”⁵

They press a second point as well. Even if the brain wasn’t functioning at conscious-experience levels during cardiac arrest, it may have been functioning enough to record information that was only later assembled into a remembered experience.⁶ And third, they appeal to scientific progress: maybe future neuroscience will discover residual activity that we cannot now measure.⁷

Put together, the objection has three layers. First, the experience didn’t happen when it seems to have happened. Second, even if the brain seemed too compromised, perhaps it was doing more than we know. Third, science will eventually fill in the gap. If even one of these moves succeeds, the dualist case from veridical NDEs is in trouble.

B. Identifying Weaknesses

The first problem is that this argument predicts the opposite of what NDEs actually look like.

If NDEs were generated by a recovering brain—an organ struggling back into function—we should expect what every emergency physician knows recovering brains produce: confusion, agitation, fragmented imagery, delirium, disorientation. Patients waking from cardiac arrest, severe hypoxia, anesthesia, or coma routinely come up muddled. They don’t come up with hyper-lucid, coherent, structured narratives that they later describe as “more real than real.” They come up like someone surfacing from a long dive—groggy, confused, with broken pieces.⁸ NDEs simply do not pattern with delirium. They pattern with the opposite of delirium.

The second problem is that Marsh’s timeline does not fit the data even on its own terms. Reynolds described her experience as one continuous, unbroken arc—beginning when the bone saw started up (before standstill) and ending only when she felt her body jolt under the cardiac defibrillation paddles (during rewarming). She didn’t experience two separated episodes. She experienced one. Sabom, who interviewed her in detail, noted that her own testimony placed the “close” of her experience at around 2 p.m.—a span that included the standstill, the flat EEG, and the rewarming.⁹ Marsh has to cut that arc into pieces and dispose of the inconvenient parts. Chapter 5 develops this in full; here it is enough to note that Marsh’s “the NDE happened only during rewarming” reconstruction is not what the patient actually reported.

The third problem is sharpest: Fischer and Mitchell-Yellin’s dream analogy collapses on a feature dreams do not have. Veridical perception. A dream that compresses hours into seconds at the edge of waking is one thing. A dream that, in those seconds, manages to deliver to the dreamer a correct description of the operating room layout, the conversations of staff, the unusual placement of an instrument, or the appearance of someone the dreamer has never seen before—that is something else entirely. Dreams don’t do that. The dream analogy works only as long as you ignore precisely what makes a veridical NDE a veridical NDE.

C. The Pro-NDE Response

What the EEG Actually Does During Cardiac Arrest

Start with the simple physiology, because the timing argument lives or dies on it. The brain is roughly two percent of body mass and consumes around fifteen percent of oxygenated blood under normal conditions.¹⁰ When the heart stops, blood flow to the brain stops. Within an average of about six and a half seconds the EEG begins to slow. Within ten to twenty seconds, the cortical EEG goes flat.¹¹ The cardiologist Pim van Lommel has stressed this for decades: from the onset of acute cardiac arrest, measurable cortical activity in humans lasts no more than about fifteen seconds, with no spike, followed by a consistently flat EEG.¹²

This matters because Marsh’s timing escape and Fischer’s ramp-up scenario both require a stretch of time when the brain has at least some functional capacity. In acute cardiac arrest, the residual-activity window is extremely narrow—measured in seconds, not minutes. There is no documented spontaneous resurgence after that point in the absence of resuscitation. A few seconds at the front edge cannot account for an NDE that contains structured narrative, multiple scenes, encounters, and verifiable observations of a long surgical procedure.

The AWARE Studies and the Three-Minute Case

Sam Parnia’s AWARE studies were specifically designed to test the timing question. In AWARE I, hidden visual targets were placed in cardiac-arrest rooms; if NDErs really float near the ceiling, some should report the targets. The protocol was rigorous. The results were mixed—targets weren’t reported (mostly because patients didn’t arrest in target-equipped rooms), but the study turned up something more interesting: a 57-year-old social worker at Southampton General Hospital who reported watching his own resuscitation from above. He correctly described people, positions, and an automated voice he heard saying, “Shock the patient, shock the patient”—the sound of an automated external defibrillator. The medical record showed an AED was used. Twice.¹³

The Italian anesthesiologist Enrico Facco analyzed the AED’s audio cycle and concluded that the patient’s report “demonstrates that consciousness persisted for at least three minutes during cardiac arrest.”¹⁴ Three minutes is not fifteen seconds. Three minutes is not a flicker on the way out or the way back. Three minutes of accurate awareness during cardiac arrest is a direct empirical refutation of the residual-activity escape route.

And the Southampton case is not alone. Veridical reports anchored to specific, time-stamped events during arrest—the dentures man’s correct identification of the nurse who removed his upper denture and where she put it (a case we treat in Chapter 4), Maria’s tennis shoe, Al Sullivan’s observation of the surgeon’s “flapping arms” gesture during cardiac surgery—all lock the experience to moments when the brain, by every available measure, was not functioning at the level required for ordinary perception.¹⁵

AWARE II, published in 2023, did not produce the target-confirmation Parnia hoped for. But the study’s authors acknowledged something striking: of the patients with explicit conscious recall of the arrest, including six who reported NDEs, none had EEG data showing markers of consciousness. As one careful reader of the final report put it: “Speculation does not imply association.”¹⁶ The skeptic cannot point to AWARE II as evidence of brain activity during the experience, because the study did not find any.

The Borjigin Rat Study

The most often-cited materialist response is a 2013 study by Jimo Borjigin and her colleagues at the University of Michigan. The Borjigin team induced cardiac arrest in nine rats and observed a roughly 30-second surge of high-frequency neurophysiological activity following the arrest, before the EEG flatlined for good.¹⁷ Critics quickly held this up as the answer: maybe NDEs are produced by such a surge in the dying human brain.

The Borjigin result is interesting. It does not, however, do the work the skeptic needs.

First, thirty seconds of brain activity in a rat cannot account for the structured, often subjectively long, narratively detailed NDEs reported by humans. Even if every second of surge corresponded to a second of inner experience, you would have a half-minute event. Many NDEs feel like long stretches of time and contain extensive content—life reviews, conversations, multiple scenes.¹⁸

Second—and this is decisive—a surge of brain activity, however vigorous, cannot grant the brain access to information it could not otherwise obtain. A burst of cortical coherence does not enable a rat to perceive what is happening in another room. It does not enable a human patient to identify a nurse she has never seen, describe a tool she has never inspected, or hear conversations beyond the reach of her ears. The Borjigin finding, taken at face value, would supply at best a substrate for some kind of internal experience. It cannot bridge from neural activity to veridical perception of distant events. The very feature that makes a veridical NDE evidentially powerful is exactly the feature a Borjigin-style explanation cannot touch.

Third, the human evidence does not match the rat evidence. The Chawla study, sometimes pointed to as a human parallel, looked at a different scenario altogether: gradually dying, deeply comatose patients on withdrawal from life support, not acute cardiac arrest. Roughly eighty percent showed a brief surge of EEG activity at the end of life. Chawla himself acknowledged that he could not say in which part of the brain the spikes occurred, and could not rule out artifact.¹⁹ A 2017 analysis by Loretta Norton and colleagues suggested that some of the apparent activity could be muscular, not neural, in origin.²⁰ Van Lommel has been blunt: in acute cardiac arrest, this kind of surge has never been observed.²¹

Fourth, the much-publicized 2022 case of an 87-year-old patient (Vicente, Zemmar, et al.) is sometimes cited as a possible smoking gun. But the gamma activity in that case occurred before cardiac arrest, not after. Greyson, van Lommel, and Fenwick wrote a careful commentary pointing this out, along with the fact that the type of gamma activity measured is not specifically linked to consciousness and may have been influenced by muscle contractions.²² The patient, moreover, died without ever reporting an NDE—making the speculative bridge to NDE explanation even thinner.

The Lucidity Paradox

Even if we hand the skeptic every benefit of the doubt and grant that there is some unmeasured residual activity during arrest, one problem remains—and it is fatal. A compromised, hypoxic, recovering brain produces confusion. NDEs produce clarity. Bruce Greyson and other researchers have repeatedly noted that NDErs typically describe their experiences as more lucid, more vivid, and more cognitively organized than ordinary waking consciousness.²³ Peter Fenwick has called this the “lucidity paradox”: at exactly the moment the brain should be falling apart, the mind reports the opposite.

Think about how strange that is. Take any other condition where brain function is impaired—general anesthesia coming off, severe hypoxia, a high fever, a stroke, post-cardiac-arrest delirium—and you find the same family of symptoms: scrambled time-sense, disorganized speech, broken attention, fragmentary perceptions. There is no medical condition in which a struggling brain reliably produces, instead of those symptoms, a coherent autobiographical narrative anchored to verifiable external events. That is not what dying brains do anywhere else in medicine. It is, however, exactly what we would expect if consciousness is not produced by the brain but—at moments of crisis—temporarily uncoupled from it.

D. Counter-Objections

A careful skeptic will press one more move: maybe the perception happened in a tiny window of activity, and the rest—the duration, the lucidity, the structure—was generated by reconstruction afterward. The brain caught a glimpse of the room during the front-edge fifteen seconds, and then later, when memory came back online, the mind elaborated that glimpse into a long, coherent narrative.

It’s a tidy story. It runs into two walls. First, the veridical content frequently extends well beyond the front-edge seconds—the dentures man’s nurse moved his teeth at a specific moment of the resuscitation; the AED voice that the Southampton patient heard cycles every three minutes. Second, reconstruction does not normally manufacture details that turn out, on independent verification, to be correct. Confabulation produces plausible-sounding mistakes, not improbable-sounding facts. We treat the broader memory question fully in Chapter 18; the short version is that the lock-and-key fit between specific NDE reports and specific events in the operating room is exactly what reconstruction does not tend to produce.

The other counter-move is the appeal to future science. We don’t know what we don’t know; perhaps future tools will reveal residual activity we cannot now detect.²⁴ This is fair as far as it goes. But it cuts in both directions. The skeptic cannot use a hypothetical future discovery to overturn present evidence; that is an appeal to ignorance dressed in lab coats. We are obliged to follow the evidence we actually have. And the appeal-to-the-future move is one the dualist could play just as well: maybe future science will refine our measurements of the moment of arrest and reveal, more clearly than ever, that the experiences are happening when the brain cannot generate them. The bare possibility that future research could go either way is not a reason to dismiss what current research already shows.

The timing objection, in the end, is a sophisticated last-line defense. It concedes the experiences are real, concedes that they often contain accurate information, and concedes that something needs explaining—but tries to relocate the experience to a moment when the brain might bear the weight. The actual measurements will not bear it. The Borjigin rat cannot explain what the dentures man knew. A recovering brain does not produce hyper-lucid coherence; it produces confusion. The Southampton patient’s three minutes of accurate awareness during arrest cannot be squeezed into fifteen seconds of front-edge brain activity. And every time the skeptic moves the experience to a window of partial brain function, the lucidity paradox surges right back: that brain, in that window, does not do anything like what NDEs do.

Our best evidence keeps pointing the same direction. Consciousness, in these moments, is doing something the brain cannot account for. Which is precisely what dualism predicts.