As every time this happens, allow me to quote Alistair Young:
"Interruptions in consciousness can't cause a break in identity because we sleep, which interrupts our narrative thread of consciousness. Except, [some] argue, it doesn't. Which we could argue and I'd be prepared to argue: we certainly have
some type of consciousness going on in REM sleep, but it gets a lot more dubious in deeper sleep states than that. But in any case, and here's my point: it doesn't
matter, because sleep is only the least of the interruptions in consciousness which can be examined. There are also unconsciousness, anaesthesia, coma (natural), coma (medically induced), various states of suppressed brain activity using TMR, extreme hypothermia simulating brain death, and seizure disorders which may not suppress all electrical activity in the brain, but
do derange it all to hell. [...] In short: people have come back from having a
null electroencephalogram, which is to say
a complete absence of consciousness and indeed dynamic mind-state. (Which is why checking for brain death in a medical context requires a sustained absence of such, not just noticing said absence is present.)"