Dying Brain Activity

Frontier The prize is a fact nobody can currently obtain: whether a person is still having an experience at the moment a clinician decides they are not. In every other disputed case of consciousness there is a channel to check — a command followed on a scan, a button pressed, a later report. The dying brain is the one condition where a candidate neural correlate can be recorded while every behavioural channel is gone, so the recording is all there is, and the question of what it means is left standing.

Established The magnitude is best given as the gap between what has been recorded and what has been confirmed, because that gap is the whole subject. AWARE-II placed EEG on 85 cardiac-arrest patients and found near-normal electrical activity return in a minority as long as 35 to 60 minutes into CPR, at a mean cerebral oxygen saturation of 43%, deep into a window that textbooks call irreversible brain injury. In the same study, concealed images were shown and sounds played during resuscitation to 28 survivors who could be interviewed, and none identified the image from a set of ten. The recording says something was happening; the recall channel says nothing verifiable came back.

Speculative If the dying-brain signal is experience — which no experiment has shown, and section 7 says what would — then the payoff is a report-free marker of presence, applicable where testimony is impossible. That would change the one decision here that cannot be reversed: the withdrawal of life-sustaining treatment, taken today about somebody a monitor cannot interrogate. The gains accrue to unresponsive patients and their families, to the law and medicine that define the moment of death, and to anyone who wants the boundary between reversible and irreversible drawn by a measurement rather than a convention.

Frontier The condition on all of it is the one section 5 names as binding: an electrical signal is not evidence of awareness without a channel that ties it to an experience, and that channel has been attempted twice at scale and has not yet delivered. The honest upside is therefore conditional and, stated plainly, it is the resolution of an ambiguity rather than the discovery of an afterlife — but it is an ambiguity sitting under a decision made thousands of times a day.

Established When the heart stops, the brain loses its blood supply within seconds and its electrical signature collapses fast. Scalp EEG typically flattens toward an isoelectric line within roughly 10 to 30 seconds of circulatory arrest, because cortical neurons run on a continuous oxygen and glucose supply they cannot store. This flat line is the baseline against which every claim in this brief is measured: activity is surprising precisely because the default expectation is silence. Frontier The window in which any of the disputed activity appears is short — seconds to a few minutes around the arrest, or transiently during prolonged CPR — not a sustained state.

Frontier Against that baseline, several groups have recorded organised electrical activity during dying or resuscitation that was not expected to be there. The signals fall into a few families: a transient surge of high-frequency (gamma) oscillations seen in animal and human brains at or after cardiac arrest; near-normal mixed-frequency EEG (delta, theta and alpha) returning during prolonged CPR in a minority of patients; and brief end-of-life spikes on depth-of-anaesthesia monitors as blood pressure fails. Each is a measurement. None of them, on its own, is a report.

Speculative The interpretive leap the field cannot yet justify is from “this pattern resembles the electrical signature of consciousness” to “this person is conscious.” Gamma synchrony, anterior-to-posterior connectivity and cross-frequency coupling are markers that correlate with wakeful experience in the living, testable brain. Reading them off a dying brain and inferring experience assumes the correlation survives into a state where it can no longer be checked — which is exactly the assumption at issue.

Frontier The one available bridge is survivor recall. A fraction of cardiac-arrest survivors report structured, lucid experiences from the period they were pulseless — the phenomenon the field now calls a recalled experience of death. If such a recollection could be time-locked to a specific recorded electrophysiological window, the signal would acquire a witness. That linkage is the object of the required experiment, and it has never been made.

Handwave Around this measured core sits a large body of claims that the recordings prove survival of consciousness, an afterlife, or mind independent of brain. This brief treats those analytically in section 15: it separates the recorded signal, which is real, from the interpretation, which is not established, and declines the mystical framing entirely.

Established The anchor animal result is Borjigin and colleagues’ 2013 rat study, which reported a surge of gamma oscillations within about 30 seconds of cardiac arrest that briefly exceeded waking levels. In that model the gamma increase appeared across brain regions, preceded the flattening of the EEG, and came with increased anterior-posterior connectivity and cross-frequency coupling; the coordinated activity outlasted the termination of heart function. It is a replicable measurement in a controlled preparation, and it is the template for everything that followed. Established The load-bearing measures throughout this literature are coherence, cross-frequency phase-amplitude coupling and directed connectivity, not raw gamma power alone — a distinction that matters because it is the coordination, not the amplitude, that is read as a signature of experience.

Frontier The human anchor is Borjigin’s 2023 study of four comatose patients whose ventilators were withdrawn. Two of the four showed a surge of gamma-band activity, accompanied by an increase in heart rate, localised to the temporo-parietal-occipital junction — the posterior “hot zone” that consciousness researchers associate with the neural correlates of experience. The other two showed neither. Both responders had a documented history of seizures, though none in the hour before death, which is a caveat the authors themselves raise.

Established The authors of that study drew the boundary the whole field has to respect: they state that they are “unable to make correlations of the observed neural signatures of consciousness with a corresponding experience in the same patients in this study.” They also cautioned against any global statement on the strength of four patients. The measurement is genuine and the interpretation is explicitly withheld by the people who made it.

Frontier The largest prospective human dataset is AWARE-II (Parnia et al., Resuscitation, 2023): 567 in-hospital cardiac arrests across 25 hospitals, of whom 53 (9.3%) survived to discharge and 28 could be interviewed. Eighty-five patients were monitored with EEG and cerebral oximetry during CPR. In a minority, near-normal EEG activity returned — the peer-reviewed report specifies delta, theta and alpha rhythms “consistent with consciousness” emerging as long as 35 to 60 minutes into resuscitation, at a mean cerebral oxygen saturation of 43%. Frontier By the programme’s own account, nearly 40% of the EEG-monitored patients showed brain activity returning to normal or near-normal at points up to an hour into CPR, well past the roughly ten-minute mark after which permanent damage is conventionally assumed.

Frontier The commission’s anchor described AWARE-II as recording near-normal gamma alongside delta, theta and alpha; that is worth stating precisely. The institution’s own press account lists spikes across gamma, delta, theta, alpha and beta bands, but the peer-reviewed abstract foregrounds delta, theta and alpha. Gamma is centrally the Borjigin finding; in AWARE-II it is a secondary, less emphasised observation. The distinction matters because gamma carries most of the consciousness rhetoric, and the peer-reviewed record leans on the slower rhythms.

Established AWARE-II’s decisive negative result is the one least reported: the concealed-stimulus test returned nothing verifiable. A tablet displayed one of ten images and headphones played sounds during resuscitation. Of the 28 interviewed survivors, one identified the auditory stimulus (3.5%) and none identified the visual image; 11 (39.3%) reported some memory or awareness, and 6 (21.4%) reported a transcendent recalled experience of death, but the objective perceptual channel was effectively empty.

Frontier The precedent for that test, AWARE-I (Parnia et al., 2014), is the reason the field is cautious. Across 15 hospitals and 2,060 cardiac arrests, 330 survived (16%), 140 were eligible and 101 were interviewed; 46% reported some memory, 9% an experience compatible with a near-death experience, and 2% explicit visual or auditory awareness. Roughly a thousand shelves bearing hidden images were installed to test out-of-body visual claims, but 78% of the arrests happened in rooms without a shelf.

Frontier AWARE-I produced one carefully documented verifiable case, and its structure is instructive. A 57-year-old man reported watching his own resuscitation from above and accurately described events — including an automated external defibrillator’s voice prompts and staff — that records confirmed, representing up to three minutes of apparent awareness during arrest. But he was resuscitated in a room with no hidden target, so the strongest available claim rests on auditory recall that could not be shielded, not on the visual test the study was built around.

Established End-of-life electrical surges predate the gamma work. Chawla and colleagues (2009) recorded transient spikes on bispectral-index and related depth-of-anaesthesia monitors in dying patients: after arterial blood pressure was lost, activity declined, then a short-duration high-frequency spike appeared before settling into burst suppression, and in one case raw EEG confirmed the waveform was not artifact. A later characterisation reported such surges in 13 of 28 critically ill patients who were not brain-dead, and in none of 7 who were.

Frontier Not every post-mortem signal survives scrutiny, and the honest record includes the artifacts. Norton and colleagues (2017) recorded EEG through withdrawal of life support in four patients; in three, electrocerebral inactivity preceded cardiac death by intervals of 1 minute 30 seconds to 10 minutes 10 seconds. In one patient, single delta-wave bursts of about 4.52 microvolts persisted for 10 minutes 38 seconds after the heart and blood pressure had stopped — a finding the authors themselves flag as possibly artifactual, because it followed a prolonged loss of circulation.

Frontier The most recent replication attempt failed to reproduce the signature that matters. A 2025 Scientific Reports study recorded local field potentials in 39 Sprague Dawley rats through carbon-dioxide euthanasia and saw a rebound spike across all five frequency bands about three minutes after CO2 onset in freely-moving animals, with high correlation across regions — but did not observe the pronounced high-gamma changes Borjigin reported, attributing the difference to method (CO2 euthanasia versus intracardiac potassium chloride). A signature that depends on how the animal is killed is not yet a robust feature of dying.

Established Underneath all of this is the “wave of death” and the reversibility that frames it. As oxygen fails, neurons undergo anoxic depolarisation — a large triphasic wave, a block of action potentials, loss of membrane properties — after which the EEG is isoelectric; yet rodent work shows repolarisation can be initiated on re-oxygenation within roughly 2 to 3.5 minutes, and human cases exist of flat EEG and lost brainstem evoked potentials recovering, as in a 28-month-old child whose electrocerebral silence reversed within 18 hours. Flatline is not proof the machinery is gone.

Frontier The sharpest open question is whether the recorded gamma is even cortical. Shaw (2024) argues that the perimortem gamma surge may originate in subcortical amygdaloid structures and reach cortical electrodes by volume conduction, mimicking genuine cortical gamma-band activity. If so, the “hot zone” localisation is partly an artifact of where the signal is picked up, not where it is generated, and the consciousness inference weakens further.

Frontier A second question is whether gamma indexes consciousness at all in this setting. Shlobin, Aru, Vicente and Zemmar (2023) note that “conscious experience can change without alterations in gamma activity” and that “the gamma signal is more related to consequences of conscious processing than consciousness itself.” Gamma is contaminated by EMG from scalp, jaw and neck muscle, and there is no clean, exclusive association between gamma and wakefulness even in healthy subjects. Established The muscle-contamination problem is not a conjecture: paralysis studies show that scalp frequencies above 20 Hz are substantially electromyographic, which is precisely the band the consciousness inference rests on.

Speculative A third question is whether the surge is experience or its opposite: the electrical noise of a system losing its brakes. On the disinhibition reading, the loss of inhibitory control and excitotoxic cascades during metabolic collapse produce high-frequency activity and network synchrony that look like a conscious signature but are a symptom of the machinery failing. Kayis (2026) states plainly that “gamma synchronization does not equal direct evidence of conscious experience,” and that the same activity can arise from disinhibition.

Frontier A fourth question is mechanistic and now has a candidate model. The NEPTUNE framework (Martial et al., Nature Reviews Neurology, 2025) proposes a cascade of hypoxia, hypercapnia and acidosis, increased neuronal excitability, and neurotransmitter dysregulation — with serotonin acting through 5-HT1A receptors for calming effects and 5-HT2A hyperactivation for hallucinogenic ones — overlaid on REM-intrusion and dissociation. It is a serious attempt to explain the phenomenology without invoking anything non-physical.

Speculative The deepest question is whether the correlation can ever be closed. Shlobin and colleagues write that “we may never find direct evidence to correlate subjectively reported near-death experiences with the neurophysiological changes in a dying human brain because by default we cannot ask the dying patient whether they experienced a memory recall while they died.” The survivor-recall workaround is the only route around that wall, and its yield so far is one auditory hit in 28.

Frontier A fifth question is topographic resolution. The generators of interest — the posterior hot zone, the brainstem arousal systems — are poorly seen by scalp EEG, and the human recordings that exist come from clinical monitors chosen for other purposes. Whether a dense, well-referenced array during arrest would confirm or dissolve the reported patterns is genuinely unknown.

Established The binding constraint is not instrumentation but the absence of a report channel. By definition, a person in cardiac arrest cannot report, so the recorded signal has no contemporaneous witness. Every downstream ambiguity — artifact versus signal, experience versus disinhibition — traces back to this single missing degree of freedom, and no better electrode fixes it.

Frontier The second constraint is that most cardiac arrests are never monitored. EEG during resuscitation is rare; AWARE-II managed it on 85 patients across 25 hospitals over roughly three years, and even then the concealed-stimulus targets often were not in the room where the arrest occurred. Statistical power to link a rare recall to a rare recorded window is therefore vanishingly small under current practice.

Frontier The third constraint is signal quality at the frequency that carries the argument. Gamma is precisely the band most vulnerable to EMG contamination from the scalp, jaw and neck, and CPR is a mechanically violent procedure. Distinguishing genuine neural gamma from muscle artifact and volume-conducted subcortical activity, during chest compressions, is an unsolved measurement problem. Established This is why some end-of-life “surges” on depth-of-anaesthesia monitors are dismissed on visual inspection as probable EMG, and only a raw-EEG check can tell the two apart.

Speculative The fourth constraint is source localisation. Scalp EEG cannot cleanly separate a cortical generator in the posterior hot zone from a subcortical one conducted to the surface, and the intracranial recordings that could are available only in unrelated clinical settings, never during an unplanned arrest. The Shaw volume-conduction hypothesis cannot be settled with the instruments the human studies use.

Frontier The fifth constraint is survivor numbers and memory reliability. Survival to discharge after in-hospital arrest sits near 9% in these cohorts, only about half can be interviewed, and recall is subject to confabulation, expectation and post-hoc reconstruction. A single verifiable case, like AWARE-I’s, is powerful but not statistical, and building a cohort large enough for statistics is a logistics problem measured in hundreds of hospitals.

Handwave The sixth constraint is conceptual: there is no agreed operational definition of “experience” that a machine could certify. Even a perfect recording matched to a perfect recall would establish a correlation, not settle what experience is — a limit this topic shares with the wider field covered in Consciousness Research.

Frontier This brief waits on a marker science it does not own. Whether any electrophysiological index tracks experience rather than mere processing is the central question of Consciousness Research, and until that field delivers a marker validated against something other than verbal report, the dying-brain signal has nothing calibrated to compare itself to.

Frontier It depends on covert-consciousness methods developed for the living. The command-following paradigms and complexity measures used to detect awareness in unresponsive but circulating patients — cognitive-motor dissociation, the perturbational complexity index — are the nearest thing to a report-free probe, and importing them into the arrest setting is a prerequisite for any objective claim. Speculative Even those methods, though, are ultimately calibrated against report in patients who can eventually respond, so they inherit the same circularity the arrest setting makes absolute.

Established It depends on a standardised taxonomy of what survivors report. The 2022 multidisciplinary consensus statement (Parnia, Greyson and 13 colleagues, Annals of the New York Academy of Sciences) defines a recalled experience of death by a specific narrative arc and paradoxical lucidity, distinguishing it from hallucinations, delirium and dreams. Without that common instrument, recall data cannot be pooled across hospitals. Frontier It also gives the field a way to reject the lazy dismissal that these are “just hallucinations,” by specifying the features that separate a recalled experience of death from delirium and dreaming.

Frontier It depends on resuscitation science and depth-of-anaesthesia monitoring for its baselines. The meaning of a perimortem spike depends on knowing the normal electrical trajectory of dying and of anaesthetic loss and return of consciousness — the subject of Anaesthetic Mechanism — which supplies the only well-characterised transitions into and out of unconsciousness the field can study at will.

Speculative It depends on progress in reversible whole-brain preservation and perfusion to extend the observation window. Techniques that restore cellular and electrical function to brains after circulatory arrest, discussed in Postmortem Cellular Recovery, would turn a fleeting, uncontrolled event into a manipulable preparation — though they raise their own interpretive and ethical problems.

Frontier The decisive experiment is a prospective, adequately powered, pre-registered study that records dense-array EEG on every monitored cardiac arrest while concealed, time-stamped auditory and visual stimuli play, so that a survivor’s recalled moment can be matched to a specific recorded electrophysiological window. This is the experiment AWARE-I and AWARE-II prototyped and under-powered; the decisive version is not new science but the same design run across enough hospitals, with EEG on every case rather than a fraction, to accumulate the rare coincidence of a recorded window and a verifiable recall.

Established What makes it decisive is that it converts an uninterpretable correlation into a falsifiable claim. If survivors reliably recall the concealed stimuli and their reports time-lock to recorded high-frequency activity, the signal acquires a witness; if, as in AWARE-II, the perceptual channel stays empty while EEG activity persists, the case that the signal is experience is not merely unproven but actively undercut. Either outcome moves the assessment.

Speculative The second experiment is a source-localisation test to settle the volume-conduction question. Simultaneous scalp and intracranial recording in an animal arrest model, or in the rare human depth-electrode patient who arrests, would establish whether the perimortem gamma is generated in cortex or conducted from subcortical structures — the Shaw hypothesis — which decides whether the hot-zone localisation means anything.

Frontier The third is a clean replication of the Borjigin human result with proper EMG control. Recording gamma during ventilator withdrawal with electromyographic channels and paralytic control, in a pre-registered cohort larger than four, would show whether the two-of-four gamma surge is a reproducible feature of human dying or an artifact of muscle activity and small numbers.

Speculative The fourth is a method-invariance test. Because the 2025 rat study failed to reproduce high gamma under CO2 euthanasia where potassium-chloride arrest produced it, a single model run across several modes of death would establish whether the gamma surge is a property of dying or of the manner of killing — a question that must be answered before any human inference is safe.

Frontier The fifth is the natural experiment already running in reverse: the recovered patient. Following the minority who show near-normal EEG deep into CPR and survive, and taking structured testimony calibrated to the consensus taxonomy, is fundable now on existing infrastructure and needs no new hardware — it is the closest available approach to a report-free ground truth, and its current yield defines how far the field really is.

Speculative The sixth is a perfusion-restoration study. Whole-brain perfusion techniques that recover electrical activity after circulatory arrest could let the perimortem window be observed under controlled, repeatable conditions rather than caught by accident — converting the central phenomenon from an anecdote into a preparation, at the cost of profound ethical questions.

Handwave No experiment on this list closes the final gap between a matched correlation and the fact of experience. That residue is the hard problem in miniature, and it is honest to state that even a perfectly executed decisive experiment would establish that the dying brain can register and later report an event, not that the recorded activity was itself felt.

Frontier The core engineering need is a dense-array, artifact-hardened EEG that can run through chest compressions. Real-time separation of neural gamma from EMG and mechanical artifact, ideally with simultaneous electromyography for subtraction, is the instrument the human studies lack; without it the band carrying the argument cannot be trusted during the only procedure available.

Established It needs precise time-synchronisation between stimulus delivery and recording. Matching a survivor’s recalled moment to a recorded window requires the concealed audiovisual stimuli, the EEG, the cerebral oximetry and the resuscitation timeline to share one clock to sub-second precision — ordinary engineering, but not yet standard in resuscitation bays. Established AWARE-II demonstrated the stimulus-delivery half of this with a tablet and headphones, so the remaining work is integration and scale rather than invention.

Frontier It needs source-resolving hardware the clinic does not deploy. Distinguishing a cortical generator from a volume-conducted subcortical one calls for high-density arrays or, in animal work, combined scalp and depth electrodes — the kind of interface the field of Neural Interfaces builds, imported into a setting where placement time is measured in seconds.

Speculative Longer term it needs controlled perfusion apparatus. To observe the perimortem window repeatably rather than opportunistically, the whole-brain perfusion systems discussed under Postmortem Cellular Recovery and, at longer horizons, the stabilisation methods of Brain Preservation would have to be adapted to hold electrical activity in a defined state.

Established It needs a multi-hospital data and consent infrastructure. The rate-limiting engineering is organisational: standardised protocols, waived or deferred consent frameworks, and a shared registry so that EEG-monitored arrests and their survivor interviews from many sites can be pooled — the arrangement section 10 argues is the real bottleneck. See also the timelines in section 13. Frontier None of this is exotic hardware; the reason it does not exist is that the pieces have never been assembled at one bedside under one clock and one consent regime.

Established Anaesthesiology is the closest neighbour, because it owns the only voluntary transitions into and out of unconsciousness. Depth-of-anaesthesia monitoring supplied the first end-of-life surge recordings, and the induction and emergence trajectories it characterises are the calibration states any dying-brain claim is implicitly compared against; the mechanism is the subject of Anaesthetic Mechanism.

Frontier Sleep and REM research is adjacent through the intrusion hypothesis. REM-intrusion accounts of near-death phenomenology, folded into the NEPTUNE model, borrow directly from sleep neuroscience, and the dreamlike experiences a fraction of survivors report are analysed with the same tools used for dreaming and dissociative states.

Speculative Psychedelic neuroscience is adjacent through phenomenology and pharmacology. The reported overlap between near-death experiences and serotonergic and dissociative drug states drives the 5-HT2A and endogenous-DMT strands of the mechanistic models — while also supplying Shaw’s counter-argument, that psychedelic states which mimic near-death phenomenology show delta and theta dominance and a loss of fast rhythms, the opposite of the gamma model’s prediction.

Established Resuscitation science and critical care own the patients and the timeline. Every advance here rides on improvements in CPR, cerebral oximetry and post-arrest care, and the finding that the brain shows electrical recovery long into ongoing CPR is itself a resuscitation result before it is a consciousness one.

Frontier Organ-donation and preservation science is adjacent and in tension. The timing of death determines organ viability, and any evidence that electrical or cellular function persists or recovers after circulatory arrest — the domain of Postmortem Cellular Recovery — bears directly on when donation may proceed. Established The mandatory stand-off after the heart stops in donation after circulatory death is measured in minutes, exactly the window in which the perimortem recordings claim to see something.

Frontier The rate-limiting institution is a multi-hospital, EEG-monitored cardiac-arrest registry with harmonised protocols. The decisive experiment fails on statistics unless dozens to hundreds of hospitals run the same recording and interview protocol and pool the results, because the coincidence it hunts — a recorded window plus a verifiable recall — is rare per site and only becomes tractable in aggregate.

Frontier Consent is the hardest institutional problem, not a formality. Enrolling patients who are, by definition, unconscious and dying requires deferred or waived consent frameworks, ethics-board approval across jurisdictions, and consent from survivors and families after the fact — the machinery that governs emergency research, applied at its outer edge.

Established The field has one genuine institutional product already: shared standards. The 2022 consensus statement gave the study of recalled experiences a common definition and method, which is the precondition for pooling data and the kind of coordination this area has historically lacked. It is an operating procedure for a contested subject rather than a resolution of it.

Speculative Funding shapes what gets asked. Much of this work has depended on a small number of research programmes and philanthropic funders willing to support a subject adjacent to a stigmatised one; ordinary grant funding rarely buys the multi-site adversarial structure the decisive experiment needs, and if that support pattern stops, the programme stops with it.

Handwave Credibility management is itself an institutional task here. Because the topic sits next to afterlife claims, the institutions doing it spend real effort distinguishing measured electrophysiology from interpretation — the discipline this brief’s flags encode — and a single overclaiming press release can cost the field years of standing. Speculative The recurring tension between the sober peer-reviewed text and the headline-friendly press release — visible in the gamma framing of AWARE-II itself — is the characteristic institutional failure mode of this subject.

Frontier The gravest ethical stake is the withdrawal of life-sustaining treatment. If any recorded signal were taken to indicate residual experience, it could alter or delay a decision that is currently made on prognosis and behaviour; conversely, over-reading an artifact as awareness could inflict harm through futile prolongation. Both errors are live, and the evidence does not yet license either use.

Established Prognostication is the near-term risk. The demonstrated point — that electrical recovery can occur far into CPR, and that flat EEG after hypoxia is sometimes reversible — is already enough to caution against declaring futility too early, independent of any consciousness claim. This is the safest and most defensible use of the record.

Speculative There is an ethics of hope and grief to manage. Families confront these findings in the worst moments of their lives, and the gap between “the brain showed activity” and “your relative was aware” is precisely where false comfort or false anguish is manufactured. Honest communication of the signal-versus-experience distinction is an ethical obligation, not a stylistic one.

Frontier Research consent at the edge of death is ethically loaded. Studying the unconscious dying, and later interviewing survivors about a traumatic event, requires safeguards against re-traumatisation and against enrolling those who cannot meaningfully consent, even under deferred-consent rules.

Speculative Organ donation creates a direct conflict of timelines. Any claim that meaningful brain function persists after circulatory arrest interacts with the criteria and timing of death used in donation after circulatory determination, where minutes matter and public trust is fragile. Frontier A finding used carelessly here could suppress donation rates without actually establishing that anything was experienced, which would be harm on both sides of the ledger.

Handwave Finally, there is a duty not to weaponise the ambiguity. The recordings are readily conscripted into metaphysical and commercial claims about survival of death; the ethical posture this brief takes is to hold the measurement and refuse the extrapolation, in both directions.

Frontier The largest civilizational stake is the definition of death itself. Modern death is defined by irreversible cessation of circulatory or brain function, and evidence that organised electrical activity can appear after circulatory arrest, or that flat EEG can reverse, presses on the word “irreversible” that does the legal and medical work.

Established Death is already understood clinically as a process, not an instant. The resuscitation literature frames it as a potentially reversible process after severe illness rather than a single moment, and the perimortem recordings are one more reason the boundary is a chosen line on a gradient rather than a natural discontinuity.

Speculative Organ-transplant systems inherit any shift in that line. Donation after circulatory death depends on a defined stand-off period after the heart stops; a well-founded claim that the brain remains electrically active during that window would force a re-examination of protocols that supply a large share of transplanted organs.

Handwave The cultural and religious stakes are immense and easily inflamed. Recalled experiences of death are read by many as evidence of a soul or afterlife, and any high-profile scientific finding here is absorbed into pre-existing belief on contact; a civilization that cannot hold the distinction between a recording and its interpretation will misread whatever the science produces.

Speculative If a report-free marker of experience were ever validated, its reach would extend past the dying. The same instrument would bear on unresponsive patients, on non-human animals, and eventually on artificial systems — connecting this brief to the moral-status questions of Consciousness Research and Integrated Information Theory — but that marker does not exist, and this brief’s honest position is that its absence is the whole story.

These horizons track cohorts, replications and adjudications rather than hardware, because in this field the recordings already exist and the interpretation does not:

• 10 yr: Frontier Expect further prospective AWARE-style cohorts and clean, EMG-controlled replications of the Borjigin human result, most likely confirming that near-normal EEG returns in a minority deep into CPR while the concealed-stimulus channel stays sparse. Established Expect the reversibility findings to enter resuscitation guidance well before any consciousness claim is settled. Speculative The most probable genuine advance is the recovered-patient cohort study, which is fundable now.
• 25 yr: Speculative If a multi-hospital registry is built, the field may accumulate enough recorded-window-plus-verifiable-recall coincidences to make a statistical statement for the first time; if it is not, expect consolidation around the mechanistic NEPTUNE-type models and the disinhibition reading. Frontier Expect the volume-conduction question to be settled in animal models, deciding whether hot-zone localisation survives.
• 50 yr: Speculative Controlled perfusion-restoration preparations may let the perimortem window be studied repeatably, turning the central phenomenon from anecdote into experiment. Handwave Alternatively, the correlation problem proves closed by argument — the dying patient cannot be asked — and the field stabilises with better recordings and the same unanswerable core.
• 100 / 250+ yr: Handwave At this range the open question is whether “was the dying brain experiencing anything” survives as a well-formed question at all, which depends on whether the broader science of consciousness ever produces a report-free criterion. Handwave If it does not, the honest terminal state is a rich, replicated electrophysiology of dying attached to a permanently withheld interpretation.

• Depends on This brief waits on adjudications made elsewhere on the map. Whether any electrophysiological index tracks experience rather than processing is the open question of Consciousness Research; whether the right integrated structure suffices for experience is contested at Integrated Information Theory; and the only well-characterised transitions into and out of unconsciousness come from Anaesthetic Mechanism. No typed depends-on edge is claimed, because none of those produces the specific result this brief is blocked on, which is a report-free marker.
• Requires (not on this map) First, a neural marker of experience that is validated against something other than verbal report, which no field currently produces and without which the dying-brain signal has nothing to be compared to. Second, deferred or waived consent frameworks that let the unconscious dying be enrolled and their survivors interviewed, without which the cohort cannot be built. Third, a multi-hospital registry running one harmonised EEG-and-interview protocol, without which the rare recorded-window-plus-recall coincidence never reaches statistical mass. Fourth, a source-localisation method that separates a cortical gamma generator from a volume-conducted subcortical one, without which the hot-zone interpretation cannot be defended.
• Enables A validated dying-brain marker would feed prognostication in critical care, the timing of death in organ donation, and the moral-status questions raised for unresponsive patients and, eventually, non-human and artificial systems. No typed enabling edge is claimed, and the reason is the finding: the marker does not exist, so nothing measurable is yet enabled.
• Adjacent Anaesthesiology and sleep medicine, which supply the calibration states; resuscitation and critical care, which own the patients and the timeline; organ-preservation science via Postmortem Cellular Recovery and Brain Preservation; the instrumentation of Neural Interfaces; and on the interpretive side Consciousness Research and Integrated Information Theory.

Handwave “The dying-brain gamma surge proves consciousness survives death, or that there is an afterlife.” This is the dominant popular claim and it is unsupported. The recordings show electrical activity; they show no experience, because the only channel that could — a survivor matching a recollection to a recorded moment — has returned essentially nothing verifiable. The authors of the human gamma study explicitly decline to correlate their signatures with any experience.

Speculative “A gamma surge means the patient was conscious.” Gamma correlates with wakeful experience in the living brain, but the association is neither exclusive nor clean, it is heavily contaminated by muscle artifact, and it may in this setting be volume-conducted from subcortical structures or a symptom of disinhibition as inhibitory control fails. Gamma synchronisation is not direct evidence of conscious experience.

Established “A flat EEG means the brain has switched off.” Isoelectric EEG is consistent with a dead brain but does not prove irreversibility: repolarisation can begin on re-oxygenation within minutes in animal models, and human cases exist of flat EEG and lost brainstem potentials recovering. Flatline is a signal absence at the scalp, not a certificate of finality.

Speculative “Near-death experiences are just hallucinations, so there is nothing to study.” This under-claim is as unearned as the over-claim. The consensus taxonomy finds recalled experiences follow a specific narrative arc with paradoxical lucidity, distinct from hallucination, delirium and dreaming; that they may have physiological drivers does not make them uninteresting, it makes them a phenomenon with a mechanism worth finding.

Handwave “AWARE proved out-of-body perception is real.” It did not. The hidden-target test was defeated by logistics — most arrests happened in rooms without a target — and the one carefully verified case rested on auditory recall that could not be shielded, not on the visual test the study was designed around. AWARE-II’s concealed images were identified by none of 28 survivors. Established The correct summary of the visual out-of-body test across two decades of AWARE work is that it has never once returned a positive result.

Frontier “We have recorded the moment of death.” What has been recorded is electrophysiology across the perimortem window in tiny, heterogeneous samples, some of it plausibly artifact, some of it not reproducible across methods of death. Death is a process, the recordings are sparse, and calling any single trace “the moment of death” overstates both their number and their meaning.

Speculative “This tells us machines could be conscious, or that unresponsive patients definitely are.” Any such reach depends on a report-free marker of experience that does not yet exist; until it does, the dying-brain work bears on those questions only by sharing their central unsolved problem, not by answering it.