REM Rebound

Looking back at my post about Sleep Debt, it seems illogical to me that I can be carrying round a lifetime's worth of sleep debt...

Why on the December 1st graph, didn't my brain arrange for all the debt to be paid back in one night, why do I not have solid nights of REM or SWS to catch up on all the lost sleep of my life?

It's clear to see from the following graph that I roughly caught up on the REM sleep that I missed out on, the area below the mean REM on the debt days is roughly equal to the area above the mean on the payback days.

In fact the figues bear this out too. Assuming that my 220 day mean figure of 99 minutes is my desired amount then on the two debt days I lost 73 minutes of REM, 60 of which I caught up on over the two payback days.

Why don't we just have a night of "super concentrated sleep" without any (N1 or N2) to help us catch up on REM and slow-wave sleep? 

Maybe light sleep (N1 and N2) also serves a purpose and is desirable to the body, hence it makes sure it has some each night, maybe the body actually needs N1 and N2 as a vehicle for REM and slow-wave sleep.

Maybe REM and slow-wave sleep don't need to be paid back on on a minute-by-minute basis.

From looking at the many nights of data that I have, I think it seems clear that I generally get my "normal" amount of slow-wave sleep, and that my issue is more with REM.

Perhaps each of us has an optimal amount of REM and more than that is unnecessary, perhaps it serves a function such as recharging the immune system or cognitive functions and their respective batteries can only be fully charged and no more. There is no point in over-charging a battery, so the body doesn't waste energy trying. Maybe the brain says "Okay, your immunity was low for a day, you were sluggish in thought too, you even dipped into your reserves, so we'll top you up with 'premium' rather than 'regular' tonight"

This is bourn out by this graph of the 13 nights where I had more than 140 minutes of REM sleep...

So, by selection, there is little change in the amount of REM (140 - 153) but there is a large variation in Light (N1 + N2) and Deep (Slow-wave) sleep. This shows something else besides total sleep time governs the amount of REM that the brain desires.

Yes, the total amount of sleep and REM can't be completely independent of each other because the total REM cannot exceed the total sleep, and the brain must also arrange time for slow-wave sleep too, but this suggests that if the brain desired more REM then it could arrange for a night of the "concentrated sleep" that I spoke of in the post about "Sleep Debt"

That describes conventional REM Rebound. However, as I said above I doubt that one is able to get more than a pre-set maximum amount of REM (which will vary for each person).

I think that the weekend of the Zeo Sleep Challenge goes some way to show this. 

The idea was to have a competition where the highest sleep score (ZQ) wins. The ZQ isn't the same as "Sleep efficiency", it can be made higher by the length of time that you sleep. So my idea was to partially deprive myself of sleep for the two nights leading up to it and hope that I caught up on them for the night of the challenge.

I thought that I'd also help my chances by depriving myself of REM sleep during the time that I actually spent asleep, causing more REM to be necessary on the challenge night, hence a higher score because Zeo awards bonus points for REM and slow-wave sleep. In essence I was hoping to make use of "REM Rebound".

William Dement did this with students who volunteered in his early work in the lab. He used an EEG to determine when they were in REM sleep. As soon as he saw evidence of REM he woke them. Then let them go back to sleep until he saw another REM period. Needless to say this annoyed many of the volunteers, and some gave up.

How could I manage to do this without employing someone to watch over me?

There is an excellent piece of software called ZeoScope that has been written by a Zeo user, Dan.

It was designed to work with the Zeo Raw Data Library, so that it receives the data from the Zeo and displays it on a PC. He also added in a function where the software can sound an audio alert when you have been in a certain sleep phase for a given period of time.

The purpose of this is to use as a trigger for a lucid dream (a dream in which the dreamer knows that they are dreaming and can consciously control the dream as if they were both the author and main character of the story).

It works well for this, if you select a nice quiet tone, such as raindrops or a soft voice telling you that you are dreaming.

I found a loud alarm clock tone and set it to play after I had 2 minutes of REM. I chose the alarm clock sound because I didn’t want to have a sound that I would ignore and incorporate into the dream.

After a couple of hours things seemed okay and I’d only been woken a couple of times, but as the night wore on the sounds became more frequent and persistent until I got annoyed and chose to get up early.

It was like being ripped out of another world, only to realise that you are in your bed and flapping an arm around trying to silence a non-existent alarm clock!

If I nodded off again quickly then I went straight back into REM and was greeted by an alarm again, so I quickly worked out that it was better to make sure that I was fully awake before going back to sleep.

It got to the point that as the flashing colours and shapes of dozing (hypnagogia) began to form coherent images the alarm would sound. Once I began to lucidly dream and remember thinking "Oh, I'm somewhere new... anytime now it'll go off..." and it did.

It turned out to be pretty effective...

I only clocked up 20 minutes of REM compared to my usual 1 hour and 40 minutes – mission accomplished!

I attempted to deprive myself of REM for the following night. 

I slept for longer that night and ended up with 40 minutes of REM, still an hour short of my usual amount. So if the hour-for-an-hour payback idea is correct then I would owe 2:20h of REM, which means that on the night of the challenge, all being well I'd end up with 2:20h plus my normal amount of 1:40h giving a total of 4h REM.

...but it didn't work out that way. In fact I had to force myself to stay asleep. I did manage to sleep for nearly 13 hours, but as you can see from the graph it was mainly filled with the "padding" of light sleep (N1 and N2). 

Compared to the hospital-payback night shown above I slept for 3:30h more, but only achieved 4 minutes additional deep sleep and 14 minutes additional REM. No one can say that I didn't give my body a chance to catch up on it's life long debt, and it failed to even catch up on the REM that it lost during the preceding two days.

Where is my missing REM?

Will I ever get it back?

Surely I gave my body enough chance to catch up on some much needed REM? Why did it squander the chance by filling the night with light sleep (N1 & N2)?

William Dement suggests that payback occurs over the course of many nights. That fits with my "hospital debt" being largely paid back over two nights. He recommends a period of a few weeks where you have no obligation to stay up late, or be forced awake by an alarm clock. Then after paying back the debt, your sleep will normalise into a pattern that is right for you.

I am torn between the conventional wisdom and what I experience firsthand. Perhaps the only way to reconcile them is that my awakenings during REM are the problem.

I guess that I am trying both approaches: sorting out the cause of my REM interuptions and trying to schedule my day in a way that is more friendly towards sleep.

Maybe one day I'll have a week of solid REM.

Sleep Debt

Imagine that it's midnight: you are sound asleep and the alarm is set for 5:30am as usual.

A noise wakes you up! You sit bolt-upright in bed trying to work out what it was. Eventually you get up to have a look and can't find anything so you go back to bed.

You are no longer sleepy, in fact you are now lying awake looking at the numbers on the clock as they get ever closer to the time that the alarm is set for. You stare at the clock for a whole hour...

...then eventually drop off to sleep.

The alarm goes off and you get up feeling a little more groggy than usual.

What you didn't realise is that while you were staring at the clock, you were racking up debt - sleep debt.

You don’t need to try very hard to build up a sleep debt.

Imagine losing just an hour of sleep a night for a week, the odd late night here, early morning there etc – that would take nearly a whole night of its own to pay off, not to mention the 8 hours that you’d normally spend asleep on that night.

You've been deprived of seven hours of sleep, and your brain is not going to let you forget it! The debt is going to come looking for you when you least expect it. You might be watching the TV, or sitting at your desk, or worse - driving when it creeps up on you and demands that you pay it back!

…and pay it back you will. It has a long memory as far as sleep debt is concerned... or so the current thinking says.

In his book, “The Promise of Sleep”, Dr William Dement (the sleep pioneer who named the stages of sleep and conducted many groundbreaking experiments into sleep and sleep deprivation), states that as far as his experiments were concerned, the sleep debt was still valid for at least two weeks after it was racked up. That isn’t to say that after two weeks, the debt is forgotten, but just that experiments had only explored sleep-debt for two weeks.

But what does that actually mean in practice? Does it mean that you need an additional seven hours of sleep to rebalance things? To be honest, there is division about how much of the debt needs paying back, and it may well vary from person to person.

I decided to look at my own data to see if I could find any clues.

This is a night of my sleep while I was in hospital recently with my son. I had to wake up every two hours to perform a procedure. 

There are solid chunks of wakefulness throughout the night, along with a significant drop in REM.

I normally get around 1:40h of REM and 40m of deep sleep. It's interesting that although my REM suffered, my deep sleep didn't. This suggests that the brain values deep (slow wave) sleep more than REM, so it makes it a priority before it catches up on REM.

Then the following night we were able to have a lie in on the ward, during which we caught up and paid back some of the debt (resulting in more REM and a higher score). 

The bulk of the debt was paid back on the first night, although the debt continued to be paid back the following night too. 

This is easier to see with the following graph covering the two consecutive nights of sleep debt (the first of minor debt and the next of more significant debt shown above) along with the two nights of payback. 

Besides charting your sleep every night, there are other ways of measuring sleep debt. 

People get used to being tired in the day and consider it to be normal, so simply asking someone how tired they are is a little too subjective. The Epworth Sleepiness Scale was designed to be more objective.

You can take the short test here http://www.predictonline.com/epworth.htm

Most of us are familiar with the fact that if you are tired then you can pretty much drop off to sleep as soon as your head hits the pillow. The technical measure of the time it takes to get to sleep is called "Sleep Latency".

Sleep latency is actually a good marker of sleep debt - the quicker it takes for you to get to sleep, the greater your debt. Eventually the time to sleep onset reaches zero, then you have what is called a "sleep attack" - when the brain's desire for sleep is so great that it just blacks you out, even if just for a few moments. This can have disasterous effects, especially if you are driving at the time!

The brain is choosy about the currency that you use to pay back the debt. If you missed out on REM then it wants REM paid back (this is called REM-rebound). If you missed out on slow-wave sleep (deep) then it wants paying back in slow-wave sleep. 

This screenshot of the week leading up to (and including) the sleep debt shows how an unusually low amount of REM is caught up on:

I imagine that the brain would arrange the payback at the expense of the light sleep (N1 and N2) which largely appears to me to be "padding" to keep us asleep and out of trouble (evolutionary hang-up) or a vehicle necessary for the two types of restorative sleep: REM and SWS. I tend to think of this as a "more concentrated" form of sleep.

This is nicely shown by this graph of my REM for the above period, shown as a percentage of total sleep:

When the need arises, REM can form a larger part of total sleep, proving that it isn't accrued in relation to time spent asleep, that is to say that there isn't a formula that says "For every hour of sleep, 15 minutes of it will be REM".

So it seems unlikely that a total-sleep payback of an hour-for-an-hour of whole sleep is likely, although this may be true for REM and slow-wave sleep.

This all seems very well for small debts but it didn't seem to work for a larger debt...