D7: Sleep matters, but Alexey Guzey is right (Nov 20th, 2019)

I’m publishing my newsletter this week a bit off-cycle, and also publishing it on the web on my Substack.


I’ve had tons of people send Alexey’s essay my way in the past week to ask for my thoughts on it. At first I’ll admit that I was highly skeptical just based off the title, and because from both meeting Matt Walker and reading his book, I’ve found him to be quite credible.

While I’m glad that Matt started a conversation that I think desperately needs to be had, which is how large of an input sleep is into your mental and physical health, Alexey reveals some errors within the first chapter that now put into question the entirety of the book. I do think the errors come more from a zealous belief in the importance of sleep than malice, however some of Matt’s misrepresentations seem to be having adverse side effects, both in terms of the general population but also in the academic community citing his incorrect statements.

Unfortunately the chapter that Alexey chose to really dive into, is one of the less interesting chapters in the book, Matt Walker’s hypotheses about the difference and utility of both deep and REM sleep were the parts of the book that I found the most novel, I may try to find time before the end of the year to dig into these chapters myself and compare Matt’s assertions to the underlying literature. I hope it is less riddled with errors in comparison to this first chapter.

I don’t think my personal philosophy on sleep has or was impacted by either Why We Sleep or Alexey’s essay. I still believe that I should be aiming for 7.5-8.25 hrs of total sleep a night (primarily because this is what makes me feel best), with ideally minimal wakeful events and consistent wake-up times. Minimizing wakeful events is mostly to make sure I’m getting the full amount of REM and deep sleep that my body needs each night, which is what matters far more than the particular duration. I imagine as electroneurostimulation tech advances (two tangential papers on topic [1,2]), we should be able to more efficiently transition from waking → light sleep → deep sleep, which could cut around 90 mins off that sleep time.

The unfortunate side effect that Matt’s book seems to be causing is insomnia amongst a population that was sleeping just fine before with less than 8 hours of sleep, but now have insomnia because they got worried about not sleeping enough. There are plenty of people who can function with less sleep due to genetic mutations [link]. What’s more important is to proactively experiment to find what makes you feel best, and also honestly not worry so much about your sleep/track it too rigorously since that can actually make your sleep worse [link]. I know this sounds like a bit of a dichotomy, to both pay more attention to your sleep, but also not track it as closely. The way I follow these recommendations is by adhering to relatively strict “sleep hygiene”, i.e. minimizing blue light, not drinking alcohol, maintaining a consistent night-time routine and bed-time, but then when I wake up in the morning, I try not to immediately go check my sleep tracker or self-assess how rested I feel, instead I get on with my day and check my sleep data in the evening and consider what changes I should make. Letting yourself get too worked up about sleep, especially when you are near sleep, i.e. in bed after you’ve just woken up, can have adverse effects on your sleep. Better to think about it and study it from afar, i.e. 6p at your desk at work.


One of the points that Alexey really dug into was that short sleep by default meant that you were living less. This curve, which was made by a meta-study of sleep studies, shows that the actual mortality risk curve follows a U-shape, where both more than and less than 7 hours of sleep. However I think this graph, especially on the upper-end of above 8 hours, is actually just conflating poor quality of night sleep with total sleep duration. Two relevant quotes from the study [link]:

“the U-shaped relationship was mainly observed among subjects with lower levels of physical activity in other studies and the interaction with body mass index was also found in another study. In addition, the association of long duration of sleep with all-cause mortality was only observed among subjects taking napping, and the U-shaped relationship was also more pronounced among subjects who napped daily.” 

“Among the studies included, a U-shaped association was found among subjects with one or more chronic diseases but no association was found among participants without chronic diseases in the 45 and Up Study”

What I believe the U-curve actually shows, on the upper end, is the effects of what Matt discusses in his book, which is your deep sleep and REM quality significantly decreasing. Older populations, especially with chronic conditions that interfere with their sleep, try to make up for the poor quality of night sleep with naps which extend the total duration. So I think Alexey suggests the incorrect conclusion of, above 8 hours of eight sleep a night is dangerous to your health, versus I believe the correct phrasing should be, deteriorating health and chronic conditions which are linked to mortality risk cause individuals to spend more than 8 hrs/day attempting to make up for the low quality sleep.

The first quote also brings up an interesting counterpoint to Alexey, where increased levels of physical activity potentially negate the negative effects of the right side, i.e. working out more means your body needs more sleep so 8 hours is perfectly fine. It seems like western society is trending towards increased physical activity, especially amongst the wealthy.

On the lower end of the curve, I think it perfectly validates what Matt Walker is trying to say, which is that less sleep leads to a higher risk of mortality, I think the curve just looks less steep mostly because the other side of the curve is heavily weighted by older populations. This is supported by another study Alexey cites [link]:

Furthermore, for short sleep the effects were comparable between men and women, did not depend on age, socioeconomic status, definition of short duration of sleep (≤ 5, ≤ 6 h, or ≤ 7 h per night), duration of follow-up, or geographic location. In contrast, for long sleep, the estimate of effect was stronger in older cohorts, in studies where long sleep was defined as ≥ 10 h per night, in follow-ups not longer than 19 years, and in East Asian cohorts.


I still mostly agree with Matt Walker’s ideology. Sleep is massively underappreciated and understudied in today’s culture, and I’m glad we are having this conversation. However it’s unfortunate that his zealousness has led to some unforced errors that come across as exaggerations in his book, although honestly, who can blame him, he was writing a pop science book, some exaggeration probably helps with sales, but I do wish his book was not being cited by scientific journal.

Grateful that someone like Alexey was willing to stand up and poke the bear by refuting the claims of an accomplished neuroscience professor and NYT best-selling author, but I still think Matt’s overall message is in the right direction. Although maybe so exaggerated that it is causing some unwanted effects by people who think they aren’t sleeping enough right now when they most likely are.

I’m going to stick to my daily workouts and sleeping 7.5-8.25 hours a night that’s for sure