Critique of Recent Report of Electrical Activity in the Dying Human Brain

Newswise — On May 1, 2023, the journal Proceedings of the National Academy  of Sciences (PNAS) published an article (Xu et al., 2023) about an  increase in EEG-detected electrical activity in two dying patients’  brains. Because many of the media reports of this study distorted the  significance and implications of these findings, we would like to offer a  more reasoned perspective on this report, as we did with a comparably  misinterpreted study last year (Greyson et al., 2022a, 2022b; Vicente  et al., 2022).  

In the present study, Xu et al. (2023) reviewed the cases of four  comatose patients who died in the University of Michigan academic  medical center’s neuro-intensive care unit since 2014. In two of those  patients, both with known epilepsy, electroencephalographic (EEG)  recordings revealed a sharp burst of gamma waves in one part of the  brain and interconnected electrical activity across both hemispheres.  These findings have been trumpeted in respected media as explaining near-death experiences (NDEs) and continuation of consciousness  after the heart stops (e.g., Reardon, 2013).  

However, the findings of this study must be very carefully inter preted because the researchers reported no evidence whatsoever that  these brain activities were correlated with conscious experiences in  those two patients—and no reason to compare these results with pro spective NDE studies in patients who have survived a cardiac arrest.  

Even though the authors wrote that they demonstrated activity in  the human brain during cardiac arrest, they in fact had not studied  patients in cardiac arrest but, rather, had studied patients in coma  as mechanical ventilation had been withdrawn. These patients had  decreased oxygen, initially even with increased heart rate, although  the authors did not mention cardiac electrical activity during the later  dying process. They did not report on any brain electrical activity when  the patients had definitely died, that is, when the EKG showed terminal cardiac arrest.

In other words, the EEG changes reported in this study were as sociated with a decrease in oxygen after the withdrawal of patients’  oxygen support but not with a total lack of oxygen as is the case in  acute cardiac arrest. The surge in electrical activity was seen only in  the two patients whose heart rates actually increased after mechanical  ventilation was stopped.  

The authors themselves acknowledged that, although they believed  their findings suggested elevated conscious processing in these patients,  no one actually observed any indication of any return of consciousness  in these comatose patients. They therefore further acknowledged that  the measured electrical activity may have been unrelated to conscious  processes.  

As we have pointed out elsewhere (e.g., Greyson, 2021a; van Lom mel, 2010), brain function has been shown in many studies with in duced cardiac arrest in both human and animal models to be severely  compromised during cardiac arrest: Immediately following ventricu lar fibrillation, cerebral blood flow ceases completely (Gopalan et al.,  1999), and the resulting loss of function of the cortex results in the  sudden loss of consciousness and of all body reflexes; the abolition of  brainstem activity, including all brainstem reflexes such as the gag  reflex and the corneal reflex resulting in fixed and dilated pupils (van  Lommel, 2010); and failure of the function of the respiratory center,  located close to the brainstem, resulting in apnea—no breathing. 

Under normal circumstances, a patient in cardiac arrest needs to  be successfully resuscitated and defibrillated as soon as possible, so no  attempt is made to measure EEG because the necessary preparations  for this assessment takes far too much time. In some cases of cardiac  arrest, however, such as during surgery, EEG was part of the surgical  protocol, so electrical activity of the brain was measured: Following  the cardiac arrest (‘no-[blood]flow’), the EEG flatlined after an average  of 15 seconds and remained flat despite external resuscitation (‘low-[blood]flow’; Clute et al., 1990; Hossmann et al., 1973; Losasso et al.,1992; Moss et al., 1980). A persistent flatline EEG during external  CPR has also been shown in animal studies (Birchner et al., 1980).  During induced cardiac arrest in humans, EEG monitoring of the elec trical activity of the cortex has shown that from the onset of cardiac  arrest, the first ischemic changes—reduction in activity—are detected  in 6.5 seconds on average, and this reduction always progresses to  a flatline EEG within 10–20 seconds—15 seconds on average (Clute  et al., 1990; De Vries et al., 1998; Losasso et al., 1992; Parnia et al.,  2002). As long as cardiac arrest continues—that is, until cardiac func tion has been restored by defibrillation—the EEG remains flat (Fisher  et al., 1996; Marshall et al., 2001). In tests on animals during induced  cardiac arrest, auditory evoked potentials can no longer be induced,  meaning that sound stimulation that normally would result in brain stem activity no longer results in any measurable activity (Brantson  et al., 1984; Gua et al., 1995).  

In acute heart attacks, the duration of cardiac arrest in coronary  care units is always longer than 20 seconds, usually at least 60-120  seconds, and in a hospital ward or in the case of an out-of-hospital  arrest it takes even much longer. Therefore, all 562 survivors of cardiac arrest in the four published prospective studies so far (Greyson,  2003; Parnia et al., 2001; Sartori, 2006; van Lommel et al., 2001) must  have had a flatline EEG. However, between 10-20% of those patients  nevertheless reported NDEs, and because of the timing of occasional  verifiable aspects of their experiences, it is clear that their NDEs and  the accurate perceptions they included must have happen during the  period of unconsciousness rather than in the first or last seconds of  cardiac arrest (van Lommel, 2013).  

We have previously published critiques of studies purporting to  demonstrate that NDEs are produced by changes in brain physiology  (e.g., Greyson et al., 2008; Greyson et al., 2012; Greyson et al., 2013;  Greyson et al., 2022a, 2022b; Greyson & Long, 2006). However, these  critiques of the details of brain physiology and their overinterpretation  have missed a vital point: Whereas such studies may contribute to an  understanding of the mechanisms through which the brain processes  phenomena such as NDEs, they do not address the cause of NDEs. 

For example, as you, the reader, read the words on this page, nerve  cells in your eyes send electrical signals to the vision center of the oc cipital lobe of your brain and to the language center of your temporal  lobe. But the electrical activity in your nerve cells did not cause the  words to appear on this page; they merely enabled you to see and un derstand them.

Likewise, regarding the well-documented phenomena of NDEs, un derstanding electrical processes in the brain might elucidate the mech anisms whereby experiencers process their memories and interpreta tions of their NDEs. However, electrical processes do not and cannot explain what enables unconscious patients to see unexpected things  in the material world accurately from an out-of-body visual perspec tive (Holden, 2009; Rivas et al., 2023); to recognize and interact with  deceased persons who, in the material world, were not yet known to  have died (Greyson, 2010a); or to experience greatly enhanced cogni tion and perception during cardiac arrest or general anesthesia when  neuroscientific models deem such complex consciousness to be impos sible (Greyson et al., 2009; Greyson, 2021b; Kelly et al., 2007).