They didn't start from the assumption you claim, their actual report states that the problem they're interested in is that there is such a wide variation in what constitutes "normal" sleep for different people. Despite the fact that a person can possibly subsist on microsleep like you claim, does not make that normal or routine and probably not beneficial for most of us. All you're providing is a contradictory anecdote to counter the assumption that they did not in fact make.
For most people, what constitutes normal sleep is highly variable from person to person, but what makes it variable? Is it 100% environmental? Or is there a genetic component? And if there is a genetic component, then what are the gene(s) or pathway(s) responsible. THAT is what the article addresses.
Further, if there is a genetic component, does it define a minimum sleep requirement, for REM or otherwise. The article does not answer this question, but addresses it further as an area for study. Who is to say what the genetic differences are, if any, in a person that can subsist on 1.5 hours of microsleep twice per day? I don't think it's fair to say that everyone can do that, and if you're saying that I'd be interested to read the literature as I can't find much on pubmed about it, except that people who microsleep frequently complain of tiredness and are at increased risk for general sleep disorders? There is ample research backing up what you say, that sleep "intensity" is more important than sleep duration, but what enables a person to have more intense sleep or what factors influence the "training" I can't find much on, so until you say otherwise I'll assume they're unknown. But this unknown component is something that the paper addresses. For some "long-night" sleepers it is known that production levels of things like melatonin are elevated for longer periods of time, but apart from light therapy I don't really know how you can "train" those sorts of responses.
What is known is that there are circadian rhythms that control biological cycles with day/night. These cycles do have a genetic component, and one of the clues that makes the gene in question in the study interesting is that it is a gene that is associated with circadian cycles (fun fact, the related homolog in fruit flies is called clockwork orange). Furthermore it's a transcription factor, probably responsible for the relative production levels of a number of different proteins. The identification of this genetic effect, the authors again state, is intended as a *starting point* to address these issues. It's not presented as an answer by any means, although the science whizzes at CNN (sarcasm) seem to have screwed it up again.
Sorry if it seems like I'm picking on you, I'm not, there are parts of the CNN article that bother the hell out of me because they really jump way too far about what the data itself can say. I just don't want people to be dismissive about what really is a good body of work that addresses an interesting question.
edit: if anyone's on a university sort of connection, there's a perspectives article by Hor and Tafti about this work that's prolly pubmed searchable by now from the same journal issue. I borrowed alot of info from that article.
editx2: the actual article addresses REM sleep, it's decreased in the mutant mice and the overall REM sleep patterns are "fragmented". They also talk about things like REM theta power that I don't understand at all...so.....?
XI Wiki


