Tracking Deep Sleep-Can It Be Improved?

Deep sleep, the stage of sleep also known as “slow wave sleep” declines during aging. Based on a meta-analysis of 65 studies representing 3,577 subjects (aged 5 years to 102 years; Ohayon et al. 2004), slow wave sleep, expressed as a percentage of total sleep time decreases during aging from 25% in childhood to less than 10% in adults older than 65 years:Screen Shot 2019-02-16 at 5.14.10 PM.png

Decreasing levels of slow wave sleep may increase risk for Alzheimer’s disease (AD), and for AD-related pathology, including an increased brain burden of amyloid-beta (AB). A correlation between decreased levels of slow wave sleep with an increased AB burden in the brain has been reported Mander et al. (2016). For example, in the picture below, during youth, non-rem slow wave activity (NREM SWA) is at its highest (blue line), and AB burden is at its lowest (red line). However, during aging, SWA decreases, and AB burden correspondingly increases:

Screen Shot 2019-02-16 at 9.46.17 PM.png

~20 years ago,  I did a polysomnograph-based, overnight sleep study and found that my deep sleep duration was less than 10% of total sleep time. Although I was relatively young (~26 years old) at the time, I had the deep sleep duration of a 65 year old! Unless I’ve been able to improve my deep sleep since then, I’d bet that my risk for Alzheimer’s Disease would be high. Not good!

I regularly blood test using the metabolic panel + CBC, which covers the health and functioning of many organ systems.However, brain health, including AB burden is not included! With this in mind, I’ve been tracking my sleep, including the different sleep stages-light, REM, and slow wave sleep (SWS) since August 2018 with the WHOOP Strap 2.0. So can deep sleep duration be improved?

First, one might think that if you sleep more, you’ll get more SWS. However in my case, more total sleep does not equal more deep sleep (p-value = 0.71). Shown below are my values for total sleep time plotted against SWS duration from August 2018 through March 20, 2019 (n=224 days):

sleep deep

As many know, studying the microbiome is part of my job as a scientist. Interestingly, based on my gut microbiome, the microbiome-analysis company, uBiome, suggested that increasing my gut levels of Bifidobacteria may improve sleep, based on their ability to produce the neurotransmitter, GABA (Yunes et al. 2016).

Furthermore, intestinal levels of Bifidobacteria decrease during aging, from 60-70% of total bacterial abundance in infants-toddlers, to 30-40% in young adults, to ~10% in middle-aged, to less than 5% in older adults (Arboylea et al. 2016):


Based on uBiome’s suggestion, the finding that Bifidobacteria decrease during aging, and also, that my gut intestinal levels of Bifidobacteria have never been good (more on that in a minute), I decided to try a Bifidobacteria-based probiotic to seed my microbiome for the third time. Previous attempts to colonize my gut with Bifidobacteria failed, as their levels were almost undetectable in my stool samples. To do a better job of getting them in my gut, this time I supplemented with a much higher dose (6 billion colony forming units, CFU). In addition, I previously used Yacon root powder (a rich source of inulin+ fructoligosaccharides, FOS) as the prebiotic for their growth, and for whatever reason, despite publications demonstrating the efficacy of inulin and/or FOS for stimulating the growth of Bifidobacteria, it didn’t work. This time, I used 60 grams of dry chickpeas (then cooked to an undetermined weight) as the prebiotic to stimulate their growth. Galactooligosaccharides from chickpeas have been shown to stimulate the growth of Bifidobacteria, albeit in mice (Dai et al. 2017).

So, did it work? Did my Bifidobacteria levels increase in response to supplementation + the prebiotic? Shown below is the species-level relative abundance of Bifidobacteria in my gut microbiome. In March 2016, I had literally 0%. In June 2017, they were at 0.25%, and in October 2017 they were 0.19%. What happened this time?

I supplemented with a Bifidobacteria-based probiotic in conjunction with the chickpeas for 23 days (Jan-Feb 2019). I then sent my stool for gut microbiome analysis 1 week after stopping probiotic supplementation, to see if the Bifidobacteria would be completely washed out from my gut. They weren’t-my gut Bifidobacterial levels jumped to 1.18%, a 54-fold increase!bifid ub.png

Did my deep sleep duration, and the percentage of time that I spent in SWS increase during and after the Bifidobacteria experiment? Shown below are those data. First, based on data for 158 days from August 2018-January 8, 2019, my average deep sleep duration (SWS%) was ~10% (red column).  As I mentioned earlier, this amount is what’s found in adults older than 65y, and my goal is biologic youth, so I’d like to increase this value! Supplementing with Bifidobacteria (yellow column) did not significantly increase my % SWS (p=0.14), as it remained at ~10% of total sleep time. However, when I completely stopped the probiotic, but kept the chickpeas, my % SWS jumped significantly to ~13% of total sleep time (green column), a value that was significantly different when compared to the Bifidobacteria-supplementation period, and when compared with the pre-supplementation period!my sws pct.png

Why would my % SWS increase after stopping probiotic supplementation? It’s possible that I was taking too high of a dose, 6 billion CFUs/day. Then, when their levels (potentially) dropped over time without being consumed, my gut may have settled on an amount that was good for my deep sleep. If this is true, would I see further improvements at a lower dose of Bifidobacteria? I’ve ordered more, and will try this experiment, so stay tuned!

It’s important to note that other factors may also affect my average daily average % SWS. Included on that list are whether I exercised with weights (or not), my calorie intake/my heart-rate predicted calorie intake, my salt intake, and various other dietary factors. More on that in another post!


If you’re interested, have a look at my book!


Arboleya S, Watkins C, Stanton C, Ross RP. Gut Bifidobacteria Populations in Human Health and Aging. Front Microbiol. 2016 Aug 19;7:1204.

Dai Z, Lyu W, Xie M, Yuan Q, Ye H, Hu B, Zhou L, Zeng X. Effects of α-Galactooligosaccharides from Chickpeas on High-Fat-Diet-Induced Metabolic Syndrome in Mice. J Agric Food Chem. 2017 Apr 19;65(15):3160-3166.

Lucey BP, McCullough A, Landsness EC, Toedebusch CD, McLeland JS, Zaza AM, Fagan AM, McCue L, Xiong C, Morris JC, Benzinger TLS, Holtzman DM. Reduced non-rapid eye movement sleep is associated with tau pathology in early Alzheimer’s disease. Sci Transl Med. 2019 Jan 9;11(474).

Mander BA, Winer JR, Jagust WJ, Walker MP. Sleep: A Novel Mechanistic Pathway, Biomarker, and Treatment Target in the Pathology of Alzheimer’s Disease? Trends Neurosci. 2016 Aug;39(8):552-566.

Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV. Meta analysis of quantitative sleep parameters from childhood to old age in healthy individualsdeveloping normative sleep values across the human lifespanSleep. 2004 Nov 1;27(7):1255-73.

Yunes RA, Poluektova EU, Dyachkova MS, Klimina KM, Kovtun AS, Averina OV, Orlova VS, Danilenko VN. GABA production and structure of gadB/gadC genes in Lactobacillus and Bifidobacterium strains from human microbiota. Anaerobe. 2016 Dec;42:197-204.

9 thoughts on “Tracking Deep Sleep-Can It Be Improved?

  1. Great article! Thank you for the information. I look forward to hearing your further experimentation on sleep. Pam L.

    1. Great question. My normal bedtime atm is ~1030 PM. When I stay up later than that (after midnight), anecdotally, I remember a few days where my deep sleep amount was terrible. Whether it was from staying up later than normal, I don’t know.

      Also, considering that humanity probably spent millions of years going to sleep close to when the sun went down, I’d bet that bedtimes closer to sunset would be optimal for deep sleep. However, doing that in the near future is a major challenge for me!

  2. Interesting article. I think you are being generous when you say “sleep stages can be predicted”. I also use Whoop (for 30 months now) and have tried other devices to track sleep. I would say that it is nothing other than a guess as to what stage of sleep you might be in. Whether you move a lot or not seems to be the major factor in determining what stage of sleep you are in.

    Whoop has limited accuracy detecting when I’m asleep or simply lying in bed reading. It will often give me an extra hour or two of sleep, simply because I choose to read in bed.
    Equally, I’ve seen enough Whoop HR data from other users to know that is not particularly accurate at tracking Heart Rate itself. If it struggles to report HR accurately then how can anyone have confidence regarding the other data.
    It also consistently under-reports HRV compared to a heart rate strap worn for the same period.
    I can also remove the strap and it will still report my heart rate and it will fluctuate as if measuring HR, even though I’m not wearing the device.
    I love the concept of Whoop and really want them to be successful. I hope that the next generation of device will make a major step forward in accuracy.

  3. Hmmm. So do you think it was it the chickpeas or the probiotic?

    Regarding sleep, I’m curious whether you’ve ever tried cryotherapy. I had my first whole body cryo session yesterday and I slept like a baby last night. I’m curious what your experience has been. The owner/operator of the cryo studio I went to said cryotherapy increases lifespan. I didn’t ask to see the research paper, but if there was one, I figured you would know about it!

    Thanks Mike!

    1. It could be, chickpeas and Bifido. In terms of cryo, you’d need a lot more trials to see if it’s a real effect. It could just be a coincidence that you got great sleep on that day. No data on cryo and lifespan!

Leave a Reply