Given that we inhabit a planet with a 24 hour day, it is no surprise that our bodies have become roughly linked with the day/night cycle, with various hormones being released at certain stages in the rhythm enticing us to sleep or wake.

Not everyone's rhythm is the same, some people have theirs shifted (such as when you are jet lagged). Even a skew of an hour or two can make a huge difference to how easily you are able to wake up for work in the morning, or how late you can stay awake.

There are a couple of online questionnaires to help you determine your own circadian rhythm; I took one from Phillips and the other from the BBC. Both of these decided that my rhythm was skewed, so I decided to make my own measurements and chart them here but after seeing the first 24 hours of data, something interesting emerged...

Establishing Your Circadian Baseline

This is a normal circadian rhythm taken from the Philips website. 

In fact, this is a simplified rhythm, because it is normal to have a small dip in the energy levels around mid afternoon, hence the reason for many countries having an afternoon siesta, and the reason why we tend to slump at our desks in the mid afternoon and reach for a cup of coffee.

This dip can be seen more clearly in my graph from the BBC circadian test.

Core body temperature is a good way of charting your own rhythm because energy levels are reflected by our temperature. Our body temperature is not a static 37.5 C, in fact it fluctuates throughout the day inline with our circadian rhythm. Of course, it is also affected by ambient temperature and how active we are. So, all it would take to chart my own rhythm would be a decent thermometer.

The Things We Do For Science

I decided to get a data-logging thermometer, so that I could see what was happening while I was asleep. I ordered a thermometer capable of recording every 10 seconds (to weed out anomalous readings), and was expecting the probe to be a small thermocouple-type sensor that I could stick to the skin under my arm, but the probe was more like a bullet and wasn't adhesive plus the logger turned out to be a lot bigger than it looked in the photo!

I'd gone this far, so in the name of science, I shaved under my arm and fixed the probe there with Melonin (low adhesive dressing) and Tegaderm (a clear waterproof adhesive dressing). I then wedged the cables into my suit and went off to work.

After 24 hours (and a few weird looks) I downloaded the data to a computer. I noticed that generally the pattern was the same as that belonging to "normal" people, but that it had a few fairly severe drops in temp during the night.

I repeated this over several days. On the graph below, the dip is around 1:30pm and again at 4:00pm. I suspect the 1:30pm dip is my normal afternoon energy dip whilst the 4:00pm dip is environmental because I was working outside at 4:00pm.

It's also interesting that the temperature drops around 8pm. That's pretty much when I start to look forward to sleep! It plateaus at 10pm for a couple of hours in a similar way to the one produced by my answers to the BBC link. This may have been because I was already asleep by then!

I repeated this the following day and the dip was much more pronounced again early at around noon, but I ended the recording early as it was becoming awkward to carry the data-logger around with me, so I can't show a full 48 hour graph.

I also noticed that the drops in temperature during sleep are the most severe. 

If You Can Repeat it, it's Science

When I first saw them, I suspected that they were erroneous readings, but the probe was stuck so tightly to me that there was no way that these were due to poor contact (after all, if that were the case, why didn't this happen during the day when there was more movement?). However, as the data rate on the logger was so high (240 recordings an hour), these were clearly not erroneous as they slowly crept back up to the plateau level of around 36.5°C. I wasn't happy with this result, so I repeated it the following night.

The drops in temperature were still there.

On the second night, to eliminate the possibility that these variations were due to environmental changes, my room temperature was monitored with another probe, and it was a surprisingly constant 18°C for most of the night, so there were no corresponding dips in room temperature. I don't like my bedroom to be too warm so I switch the heating off at night.

I took a look at my sleep-stage data using the Zeo, and was surprised that the seemingly random falls in temperature coincided with awakenings from REM! (Shown as orange "wake periods" on the coloured graph, but as REM/Wake on the detailed graph). So what's going on?

Unwrapping the Mysteries of REM

Along with the muscle paralysis that comes with REM sleep (to stop us acting out our dreams), there is also a decrease in the body's ability to regulate its temperature (thermoregulation). During REM this decrease in ability to regulate our temperature causes us to be largely "Poikilothermic" (meaning that our temperature is governed by the temperature of our environment).

So, it seems that my temperature was dropping, causing me to wake in order to regulate my temperature, pretty much in the same way that my apneas cause me to wake during REM.

If this is true then it means that in order to be able to stay asleep during REM I now have two issues to deal with: apneas AND temperature!

So the key to eliminating this issue is to regulate the temperature of my bedroom better.

I repeated this for two more days, and then concentrated on the nights and gathered more data (total of 5 nights).

It seems that these overnight drops in temperature aren't always correlated with REM sleep (sometimes they are just before and sometimes just after), so until my more accurate thermometer arrives, I decided to calculate the mean temperature from the week's recordings to get a better idea of my rhythm and to try to eliminate the overnight drops.

This gives a much nicer picture and goes some way to reducing the environmental effects. The data fits well with how I feel throughout the day. I tend to get a bit tired at 1pm, then start considering bed around 8-9pm.

Compared to the BBC graph, it lines up fairly well, although my "energy drop" happens earlier in the day at a more accurate time.

All that said, I'm not comfortable with this experiment as there is much that could be improved on:

• Better temperature sensor (on it's way)
• Eliminate environmental temperature effects (maybe on a day where I stay at home)
• I still am not convinced whether these changes in temperature are an effect of activity. After all, it stands to reason that when we are more active our bodies produce more heat, so it would be interesting to do this on a day where I forced myself to do nothing, stayed at home and forced myself to stay awake past my usual bedtime, that way I could see whether these changes are still evident on the graph. That said, some of the nights were recorded with an 11pm bedtime although my temperature still dropped as usual around 8pm, so it could be accurate.

One thing that it has shown me is that it is essential to have your sleeping environment just right for you.

A good sleep routine involves making sure that your bedroom is the right temperature. It should be cool but not cold. In days gone by, the outside temperature dropping in the evening acted as a cue for our ancestors' bodies to sleep. Of course, in days where our homes are heated, this confuses the body and can lead to poor sleep.

The same is true with light.

Nowadays we have artificial light filling our homes and offices, and that also interferes with our sleep/wake cycle as we don't receive the nighttime cues. 

So sunlight during the day and a cool, dark bedroom goes some way to remedying this.

For perhaps 6-8 months, possibly more, my wife had been showing signs of cognitive impairment. Memory loss, lack of ‘presence,’ unable to follow complex instructions, etc. It was bad enough that a few weeks ago when a new client asked her our address, she couldn't remember.

She also developed a bilateral hand tremor.

We went from her internist to a neurologist. I'd noticed that she had taken to snoring, and my thought (with which the neurologist agreed) was that it could be obstructive sleep apnea. So we did the whole sleep study routine, and yes indeed, she had obstructive sleep apnea. Tried CPAP, she couldn't stand it. She switched to Provent (small nasal strips with a one-way valve that restricts exhalation to raise pressure in the oral cavity and thus do some or all of what a CPAP machine does). That eliminated the snoring, and a recording pulse oximeter showed improved oxygenation. Great!

But her cognition didn't improve.

I then bought the Zeo, to see if she had a more or less normal sleep pattern. Well, lo and behold, after about 2-3 weeks of measurements, my wife had registered a grand total of no REM sleep.

Zero.

Not one minute in the entire period.

Since my wife was on antidepressants, I thought that her medication might have something to do with her issue. So I decided it was time to hit the research literature. Fortunately, I've had training as a medical writer and researcher, so this wasn't unfamiliar territory for me.

In poking around, I came across info that that said that SSRIs (a.k.a selective serotonin reuptake inhibitors) suppress--often entirely--REM sleep. As I delved further into the literature, I found robust support for this.

Apparently, it's a well-known (if rarely mentioned) side effect of the antidepressants. I contacted an old grad school friend, Joe Graedon. He immediately said that not only do the SSRIs suppress REM as I'd seen in the literature, but so do ALL the antidepressants, including the first generation ones.

It seems that in order to have REM sleep, you have to deplete serotonin (and have melatonin kick in) and if you're on an antidepressant, the antidepressant is upping your serotonin levels, thus lowering your melatonin levels, which is counterproductive to getting good REM sleep.

So, with the approval of my wife’s physician, we started to slowly ween her off her medication. For about three weeks, nothing dramatic happened. A week ago yesterday, she woke up one morning and within minutes I knew she was ‘back.’ She was just remarkably different. Much more animated, much more engaged, much livelier.

And SHE knew the difference.

That night she used Zeo again and she showed about 10 minutes of REM sleep.

Hooray! Not huge, but at least SOMETHING.

It has continued to increase, up to 15 minutes two nights ago, 20 minutes the night before last, and 45 minutes last night! She has continued to be much more energetic, totally present and responsive, with her memory working as it should. The hand tremors also entirely disappeared.

Again, research literature points out that antidepressant drugs can have a noticable affect on sleep quality. 

NONE of the doctors raised the meds as a possibility.

The sleep study center should have mentioned it in their report.

The neurologist should have flagged it.

Without Zeo, we would not have had the information we needed to make the necessary change.  Since tapering down the medication, my wife's been consistently getting some REM every night. 

We're hoping that this trend will continue - and that our experiences will help others assess their own medications and sleep quality.