Friday, January 21, 2011

Sleep Architecture

At the risk of recreating a wikipedia article, we need to lay some groundwork here and discuss the structure of normal sleep. Of course it is never as simple as that - normal sleep architecture changes throughout the life cycle. However, the basics are similar for everyone except perhaps preemies.

Sleep is defined using EEG readings (where you have a bunch of leads taped to your head, and the electrical output of the brain is measured on a polysomnography - which also includes muscle measurements and eye movement measurements to be completely accurate.) In general, all stages of non-REM sleep are characterized by a slowing and deepening of the waveforms as we get deeper and deeper into sleep. When we are awake but zoning out, or meditating, we have a type of waveform detected on the EEG called an "alpha wave." When the alpha waves start to become theta waves, we've progressed to stage 1 sleep. In general, someone who has been awakened from stage 1 sleep will not think he or she was actually sleeping. (Think Dad nodding in his chair on Sunday afternoon, and when you bug him, he blinks and goes back to watching the game as if nothing happened).

EEG tracing of stage 1 sleep with mostly theta waves and some alpha

A few minutes into stage I, if all goes well, you go to stage 2, and then after another few minutes into stage 3 and 4. These last two stages are characterized by "delta waves" and are commonly referred to as "slow wave sleep" and represent the deep, refreshing sleep. After about an hour or a little more (total for 1-4), normal adults will transition to REM sleep, where eyes move rapidly and large muscle groups are paralyzed. This stage is when a lot of dreaming occurs (though dreaming also happens during slow wave sleep). REM sleep waves are somewhere in between stage I and wakefulness - in fact, the sleep scientists rely on the muscle readings and eye movement readings to distinguish between someone who has woken up, and someone who is in REM sleep.

Once the REM cycle is finished, one drops down through stage 1-4 again. The cycle repeats itself every 90 minutes or so (for newborns, every 45 minutes) throughout the night, though the last couple of cycles one spends more time in REM sleep and less time in slow wave sleep.

Normal sleep architecture

A couple of important things - the first cycle or two are the ones where we spend the most time in the deep, refreshing slow wave sleep, so it is vital that these cycles are of good quality. In classic Major Depressive Disorder, for example, patients will often never reach full slow wave sleep throughout the night, thus the common complaint of insomnia, feeling constant fatigue, irritability, or being too easily awakened. These same patients typically aren't hungry and lose weight. In "atypical" depression, seasonal affective disorder, or bipolar disorder, some patients will actually spend too much time in slow wave sleep. They also feel lethargic, and will be hungry and tend to gain weight.

Alcohol consumed close to bedtime will tend to decrease sleep latency (that is, the amount of time from being awake to being asleep when we go down for the night), increases the length of time to get to REM sleep, and increases slow wave sleep for the first half of the night. However, the second half of the night, there will be more wakefulness, more REM sleep, and less slow wave sleep. Alcohol is the most common go-to substance that chronic insomniacs use to get some shut-eye, and overall it decreases the quality and efficiency of sleep.

All told, sleep problems and fatigue arise from a complicated array of too much or too little in a number of neurochemical systems. There are natural chemicals promoting wakefulness, and chemicals promoting sedation as part of the circadian rhythm, and issues with any of these can lead to complaints of poor sleep, insomnia, or fatigue. I hope to get to more of the details in future posts!


  1. Looking forward to this series, Dr Deans; I know you mentioned you'll address the topic of supplemental melatonin (thanks!), but I was also curious as to how, and if, SSRIs and/or SNRIs affect sleep quality?

  2. Yes, I can do a post on SSRIs and SNRIs and how they affect sleep. Not a problem!

  3. Very much enjoying the information in your "sleep series" Dr Deans.

    You mentioned earlier that you take magnesium oxide before bed. Any reason for the oxide vs. the aspartate form? And what would the recommended dosage be for the aspartate form?

  4. Aspartate can be a excitatory/neurotoxic chemical, so I try to avoid it. In the studies I reviewed for my magnesium series, they used magnesium oxide and found that after about a month of supplementation, people had absorbed enough to be topped off and spill a little extra in the urine. Therefore for an average person, I think the cheap oxide form is just fine as long as you are in it for the long haul. That said, the papers mentioned that if more studies were done, they would recommend the citrate version and a higher dose. If I had mg citrate at my local target, that is what I would likely buy, though my next experiment may be the MOM with selzter water (hopefully the gut will be happy). However, I have most of a bottle of 120 pills of 250 mg pills og mg oxide that I got for $3, so I'll be taking that for a while. In general the amino acid chelates are more bioavailable, wih orotate being the most available and a dose of 25mg being plenty. You can also experiment with sea salt in the bath and transdermal formulations - I have to be careful about my magnesium gel, though, as it can get itchy.

  5. Thankyou for your blog Dr. Deans. I can relate to this issue involving sleep. I'm a 34yro male from Australia and have a history of endogenous depression and anxiety, relatively treatment-resistant for the past two years, characterised by fluctuations of asymptomatic periods of varying duration (between three days and two weeks) and symptomatic periods (three days to one week, low mood, inner psychic discomfort).
    Lifestyle: excellent; nutrition: excellent; physical activity: always. Sleep: early morning wakening and restless sleep thereafter, even with Quetiapine 800mg and Clomipramine 175mg prescribed by a psych.doc.

    You mention it safe to take with the above medications? Are there any other remedies that can be considered helpful for early morning wakening?

  6. I've been running a sleep workshop for a couple of years now and saying much the same stuff as you here... Phew! People get very stressed when they wake up during the night. But the typical pattern I have noticed for most is that they will generally sleep OK for the first 2 cycles (majority deep sleep, incl. GH release) and then be a bit broken after that. But when they map it, they are typically waking every cycle, either side of a burst of REM. Providing they don't stress themselves about waking, they generally go back to sleep readily again.

    However, as people are typically staying up too late and relying on sleeping as late as possible and using an alarm clock to pull them out of bed (typically part way through a sleep cycle), they are just not getting through enough cycles per night and are building too much sleep debt - IMHO. They then hit the panic buttons when they do wake up as each night is not as restorative as it needs to be for them.

  7. what do you think about sleeping during daytime (even for a short period) - would it mess up something? if yes, what would be the mechanism?
    on a personal side, i know that a short nap can make me feel miserable (headache, insomnia...) for the next couple of days

  8. Dr. Deans, thank you for another fascinating post. I, too, am enjoying your series on sleep.

    And thank you for the warning about the aspartate.

    I use Milk of Magnesia (the kind without sodium hypochlorite) as underarm deodorant. It works very well, and I am hoping that some of the magnesium gets absorbed.

    Thank you very much for your blog. I always look forward to your post and comments.

  9. Gn - lots of places around the world have a siesta, and mid-afternoon sleep corresponds to one of the dips in the day where is is relatively easy to sleep (rather the opposite of the 8am or 8pm forbidden zones - they occur from 2-4 pm and am). Also, kids of course have a natural napping cycle with nearly constant napping/waking every 2-3 hours or so as newborns, to 3 naps a day, to two naps a day, consolidating to 1 nap a day in the afternoon around 18 months. So I think another perfectly natural way to sleep is maybe 90 minutes during the day and then stay up late for another 6 hours or so at night, siesta style. You have to do it right, though - for adults napping in the morning or sleeping too late can cause issues with the sleep cycle and the circadian rhythm.

  10. Santino - getting the sleep cycle regulated is exceedingly important in depression and bipolar disorder. Without doing a proper evaluation, I really can't comment further on your situation. However, if your kidneys are in good shape, magnesium should be just fine with those medications.

  11. Jamie - should have asked to borrow your notes before writing the post. It has to be something about this modern world (or several somethings) that sleep is such a problem nowadays. Something like 24% of American adults have insomnia, and a good percentage of that is chronic insomnia. But when you think about it, sleep should be natural. It is so vital to our health, and yet we are terrible at it.

    H - I have to say I don't know that aspartate in mg supplementation is absolutely bad - I know that even in the studies they avoided it for the reasons that aspartate could *potentially* be a problem.

  12. With people often living undiagnosed, Sleep Apnea is one of the biggest causes of daytime drowsiness.

  13. Hi Emily,

    On SSRI's and sleep quality; I've taken SSRI's for a very long time and have always had poor sleep, specifically trouble getting to sleep and waking about 2:00am.

    I've been strict Paleo/VLC for a few years now and wondered whether I could reduce my SSRI dosage thanks to the power of paleo and ketosis.

    I tried tapering off of them throughout the whole of 2010, thinking a year would be long enough to reduce a tiny bit by bit.

    To get to the point, reducing the SSRI dosage wasn't a good idea as when I reached a certain level suicidal thoughts began in earnest (I generally have them in the back of my mind most days, but on lower SSRI's I couldn't ignore them), but I did have an increased ease in falling asleep.

    When retiring to bed I had a "groggy" brain which made me not spin thoughts and allowed me to fall asleep much easier. Reminded me of taking Temazepam (although Temazepam was much nicer, I loved the feel of hardly being able to stand!).

    Alas, improved sleep lost out to staying alive so I had to increase the SSRI's. I am on a slightly lower dosage than before which makes me feel a bit better, but the poor sleep persists.


  14. I use magnesium chloride. I read it is well absorbed, though it tastes ugly.

    Melatonin gives me nightmares.

  15. Thank You so much for encouraging us sleep properly. :-)

  16. On the top eeg tracing, is the spike towards the end a vertex potential?