The question at the edge of consciousness
A heart stops. Doctors begin CPR. Two minutes later, a speaker plays sounds in the room. If the patient survives, could any of those sounds be recognized?
In a study published in The Innovation, researchers reported that some cardiac arrest survivors performed better than chance on a later sound-recognition test. The finding adds an intriguing chapter to the investigation of what a person may perceive during resuscitation.
It does not establish that consciousness survives irreversible death. The people studied were revived after their hearts stopped, and CPR can deliver blood to the brain. A silent patient is not necessarily a brain without activity.
The research, led by Álex Escolà-Gascón of Comillas Pontifical University and colleagues in Spain and the United Kingdom, was first published online in March 2026. The Daily Mail reported on the findings on October 31.
What the researchers did
The team says 13 hospitals and clinics in the U.K. and Spain monitored roughly 15,000 people at high risk of cardiac arrest. Of 2,418 patients who went into arrest and received CPR, 677 survived. The final analysis included 142 survivors selected after screening for reported lucid or near-death experiences and other eligibility criteria.
Patients had been assigned to one of three groups. Starting two minutes after cardiac arrest began, speakers played up to five sounds during resuscitation. One group heard natural sounds such as rain, waves and birds. Two others heard newly created melodies.
After recovery, participants took a 10-item recognition test: five sounds had been played during CPR, and five had not. Researchers asked them to distinguish between the two.
The nature-sound group averaged 6.68 correct answers out of 10. The first melody group averaged 6.00; the comparison melody group averaged 4.94, close to the five expected from guessing. Those are group averages, not a guarantee that any particular patient remembered sounds.
What the study cannot tell us
The team did not collect standardized EEG recordings from all participants during their arrests, a limitation the authors acknowledge. Without those brain-wave recordings, the study cannot show precisely what each participant’s brain was doing when sounds played. Nor can it pin every reported experience to an exact moment.
The sample was also selected: the paper describes screening survivors for lucid experiences before arriving at the final group. The results should not be treated as the odds that every cardiac arrest patient will remember CPR.
The researchers also reported an association between lower measured brain oxygen and stronger recall or near-death experience scores. An association is not proof that low oxygen improved memory, and it does not explain how any individual experience formed.
As a practical matter, the authors recommend respectful communication during resuscitation and opportunities for survivors to discuss difficult memories. Those suggestions do not change the medical criteria for declaring death.
