Sleep trackers

You may be interested in this article from The Conversation about the unreliability of sleep trackers. with a variety of related sleep articles underneath.

https://theconversation.com/are-sle...ccurate Heres what researchers currently know

 

Interesting to read the paragraph on sleep anxiety:

This suggests that the ‘sleep police’ who see pushing for ‘normal sleeping patterns’ as self evidently (ie unevidenced) a desirable target in treating ME may not helpful, and certainly overall questions how useful such off the shelf sleep trackers might be for us.

 

I used to get stressed about not getting enough sleep, having been an insomniac since I was a child. Eventually I reduced this stress by covering up my clock with a flannel so I can't see the clock unless I want to. It didn't actually help my insomnia but at least I was less stressed about it. My clock has the time, date, and day of week on it. I now only look at the clock to check the day of the week and the date so I keep some semblance of a connection with reality.

 

If I wake up in the night I don't look to see what time it is anymore. I used to do that all of the time. I read that people who look at the clock when they are having problems sleeping tend to be more tired the next day than people who don't clock watch.

 

I can't help but wonder how many studies have been done to lead to that conclusion, or if it's just some bods theory.

 

A while ago I did a trial of 3 devices to measure sleep, wearing all 3 at the same time:

1) a Fitbit wrist monitor (quite an old one, probably one of the first to monitor sleep at all)
2) a Polar wrist monitor (about 2 years old)
3) an Oura ring (latest version)

I was interested in seeing if there was a correlation between the state of my ME/PEM and the amount of total, deep or REM sleep. I rated my ME/PEM state before looking at the device data to reduce bias. Plus I wanted to see how similar/different the recordings from the different devices were.

The Polar, no matter which wrist I had it on, most nights didn't register any sleep at all.

The Fitbit and the Oura broadly, and quite accurately, registered the same falling asleep and waking up times as well as any longer spells awake during the night, though the Fitbit was also quite prone to falsely declaring me asleep during the day. But very different results for deep and REM sleep.

The Fitbit typically had me in deep sleep for 30-90 minutes and in REM for 60-90 minutes, and there was little correlation with PEM or with how I felt generally.

The Oura typically had me in deep sleep for 5-60 minutes and in REM sleep from 120-200 minutes, so very different from the Fitbit. Initially I thought there might have been a correlation between deep sleep and PEM but after more monitoring I don't think this holds up.

So the main conclusion from my experiment is that technology needs to improve a lot before sleep monitoring with consumer devices tells us anything useful.

But I'm also wondering if part of the reason for the vastly different readings, and the fact the Polar hardly got any readings at all, is that with ME our sleep is disturbed in a way that messes up algorithms designed for healthy people?

 

It's the same if you are standing or sitting and moving your arms around. The Fitbit tracks the movement as steps taken.

 

Yup! Best sleep tracker is common sense; people usually know pretty well whether or not they're getting enough for their current needs.

 

The problem appears to be the quality of the scientists carrying out the studies. There are a lot of soft social science studies but not much else. On my garbage research totaliser I put it at around 90%...so similar to nutrition/gut studies.

the tracker I find the most useful is the one that tells me how many times I have woken in the night, which sort of correlates with feeling a bit more awful in the morning (sleep hygiene has no effect on quality of sleep for me ..it’s random). ...oh and of course there is a big correlation with how much I move and how much sleep I need one average. 6000 steps = 7hr sleep, 2500 steps - 6 hrs.

However averages are pretty pointless when your sleep, heart rate and movement are nowhere near stable or the same as the average population.

realise I’m probably/possibly odd (I’ve left that deliberately open ended).

 

Has anyone tried Apple Watch with the AutoSleep App?
According to them I seldom have ‘deep’ sleep (unsurprising); occasionally 20 min to 1 hour, a very few times 2 hours.
If I get more ( it’s never a lot of, or enough) sleep, with some deep sleep, I’m definitely less exhausted in the morning; otherwise each day has its own effort/exhaustion ratio.
The App does monitor HRVariation though, some people think that’s important...

 

So glad to hear that the automatic, unthinking, ‘sleep hygiene’ mumble is being questioned.

 

So no one else uses duvet disruption as an indicator?

e.g. no movement, once the area lifted to get out of bed has been restored then it's as if no one had slept in the bed, or conversely, duvet shows evidence of having been rotated on a horizontal plane 180 degrees during the night - as measure of activity whilst asleep.

In combination with other things obviously.

 

I did say I was odd

Already with our two posts we have moved the field on by leaps and bounds though so well done us...however to fit in with the body of research my next correlation needs to involve obesity I think and or heart disease and stroke. That way I can line my pockets with a load of research funding to set me up for the next 5 years.

 

For most basic purposes a sleep tracker is overkill, I agree.

But I do see a couple of uses for them if or when they become good enough to provide detailed and, most importantly, accurate data. For now they aren't up to the task so the following points are hypothetical.

At the individual level: to assess subtle but cumulative effects of something on sleep quality. You probably wouldn't notice a 30% reduction in deep sleep after a night or two but long-term it could still have a negative impact. So if you're feeling inexplicably wiped out and your sleep data flags that you've been missing out on deep sleep for the last month you could try to figure out what you've been doing different lately and then test suspected culprits. It might even help you identify which types of sleep hygiene actually work for you and which are not worth bothering with

At the population level: to pick up patterns worth researching further. Let's say a large number of pwME report in the You & ME app that their REM sleep doubles in PEM (I'm making this up for the sake an example). You could then look into physiological mechanisms known to increase REM sleep and see what, if anything, this might tell us about PEM and ME.

Actually, I have a possible real example from my Oura data. My HRV is always significantly lower at night than midmorning. This is the exact opposite of what happens in healthy people. If the readings are accurate this could be interesting to investigate further. However, this finding could just be me being an oddity.

HRV has in fact been looked at in ME though mostly readings were taken either during the day when awake or at night when asleep, finding depressed HRV for both but not actually looking at diurnal rhythm. I only found one study doing so mentioned in an MEA review and its findings would suggest I am indeed an oddity. Though the study only looked at 35 Fukuda participants (Cvejic et al. 2017**; haven't read it yet).

https://meassociation.org.uk/wp-con...eview-Dysfunctional-ANS-in-MECFS-24.01.18.pdf

So this is an example where lots of people monitoring their sleep & awake HRV with an accurate sleep monitor could tell us if I'm an oddball exception or if a flipped diurnal HRV is actually much more common than the small Cvejic et al. study suggests. And if it is common it would raise questions: why does it happen and what can it tell us about physiological processes in ME?

**Paper briefly discussed here (comes with a BPS warning): https://www.s4me.info/threads/auton...hysical-or-cognitive-challenge.239/#post-3645

 

Have now read the paper (https://sci-hub.se/10.1016/j.jpsychores.2017.10.010). Despite being loaded with BPS authors it doesn't actually lead to tearing your hair out - if you can overlook the fact that they're studying post-exertional fatigue and that it's not clear if the participants also experience full PEM. Alas, all too common in all too many studies...

Anyway, what was interesting is that while they found a normal diurnal pattern with higher readings at night they also found that after exertion the difference between day and night readings was significantly reduced. These were probably quite mild patients in the study.

So mild patients have reduced diurnal variation when experiencing post-exertional fatigue/malaise, largely due to a reduction in the typical nighttime rise. Would it be reasonable to hypothesise that severity plays a role here? So that in moderate patients that nighttime rise doesn't happen at all and they end up with a flat curve? And that in severe patients there is a nighttime reduction instead of a rise, resulting in a flipped curve? If only we had accurate consumer-level sleep monitors to find out!

PS: Any further discussion of the study would be better made in its dedicated thread. Just wanted to make the point here that there would be good uses for sleep monitors if they actually worked properly.