This is the third part of my description of my new polyphasic sleep adaptation attempt. I will be detailing exactly what I am doing to measure how this adaptation proceeds and show some of the data I have gathered this far.


First things first though. The last 72 hours have been miserable, especially day 7. I have essentially been too tired to do anything even remotely passive, since I would immediately start to fall asleep. Grogginess from some of the naps would last for several hours. That also explains why this post was slightly delayed, as I had planned on posting it yesterday.

The result was that I have been standing up most of that time, and I have been extremely limited in what activities I could do. In the end, this made for a fairly painful few days. Yesterday morning this seemed to start easing up though, as the grogginess wore off form the 2:30 AM nap.

Instead I have found it increasingly hard to actually wake up from my naps, with my girl friend having to step in several times during the day. One of those times I slept through a loud song 3 times in a row, which comes out as a 12 minute oversleep or so.

As for how the adaptation is going, that is still on track. I have no other oversleeping to this point, and I am consistently getting all my naps, and they seem to be fairly good quality. Generally, I can only go about 2-2½ hours on a nap though, which means there one hour of sleepiness at the end of each normal awake period.

Groggy/sleepy periods started out being only when I used to be sleeping monophasically, so at the start of the adaptation my mornings would be worse than the rest of the time. Now, that seems to be fairly evenly spread around the day.

Right now it is ok. I am pretty sure it will go back to being miserable before going the right way. The microsleep thing would be good to get rid of though, hell if I know how. They are very disturbing, since not only are they microsleeps, they are actually something like microdreams at times. I will just notice how I have closed my eyes without wanting to and my “imagination” is going wild about something that I was not thinking about just then, just like a dream. Nothing to do but force your eyes open when that happens I guess. The only way to avoid them is to both be standing up and be doing something active at the same time… which simply is impossible over any longer stretch of time.

I do feel like it’s slowly getting better though, like it very much depends on the quality of the last nap or someting like that. Also, tonight — right now — when things are working out brilliantly, the thought of failing this and going back to monophasic sleep is intolerable.

Experiment documentation

Ever since that first experiment last year I have wanted to do an experiment where I document much more of it. After each nap, I record on a spreadsheet my subjective evaluation (1-5) of a number of factors as well as the amount of oversleep that happened in connection to that nap, and some freeform notes.

The values I record are:

  • Sleep quality: How easy it was to fall asleep, if I feel refreshed afterwards, if I had a dream, if I woke up by myself, etc.
  • Difficulty waking up: Just how horrible was it to push myself out of bed?
  • Grogginess: How groggy I feel after getting up.
  • Sleepyness: How sleepy I am through the period awake after the nap?

I am unsure what this will give me in the long run, but it is clear to me looking back on last year’s experiment that I defenitely did not record enough information on what went on.

I will also be keeping an exact count of how much I oversleep, which is currently sitting at 12 minutes for the oversleep on my 4th nap yesterday.

I actually bought a video camera as well, and have been recording some things in a video diary like format. I never liked video blogs though, so what I would use the material for is more something like a cut documentary at the end of it. This all depends on what footage I get and so on, it may all amount to nothing in the end as well.


I also decided I wanted some more objective measurement on how well I was doing mentally than “how do I feel?” To that end, I hunted down a set of tests to do for my reaction time, working memory, etc.

There is a big problem with those tests however: it is very hard to see if an improvement is due to something with the adaptation or if it is learning a test better or something unrelated. Still, over the course of time this should give me some nice insight into how I am changing.

The main problem right now though is that they are quickly becoming mind-numbingly boring. I have now done them 27 times, and while my initial thought was to maybe do the tests at a predetermined time in each polyphasic segment of the day I quickly realized that I would not be able to make myself do it so often. If I did that I would simply be doing tests with all my extra spare time — if the tests take around 15 minutes to complete, that means I am already approaching 7 hours of nothing but doing tests.

Instead, I am kind of doing them ad-hoc now, trying to get at least one instance in per day.

I have only used tests which are available online, both as a way of making it simple but also to be able to do the tests from anywhere I happen to end up if I go away for a day or two.

The majority of the tests come from the web site

  • Reaction time tests: Visual reaction time, Auditory reaction time. I simply do these a minimum amount of times and write down the average reaction speed.
  • Reactive decision time: Go/No-go Visual reaction time. Same with this one.
  • Interference test: Stroop. I simply do the test, write down the average correctness percentage (average of normal and interfered displays), as well as the average time.
  • Memory tests: Corsi block task. The way I use this test is as an escalating difficulty test. I start it out low (5 is reasonable for me, the default of 3 works too but is somewhat tedious — there is absolutely no risk that I would fail a 3-4 span), and do two repeats for each number before moving on. The end result is the sequence length where I first make a mistake. I also do a working memory test, and write down the correctness percentage I have for the 2-back and 3-back versions of it.

Finally I use one other test not from cognitive fun, which is simply a typing speed test, which measures my dexterity and precision. I use the one at, set to 1 minute on today’s travel headlines. Using headlines means I can’t learn the text by heart, which is good… but also means that sometimes you get more difficult texts.

Preliminary results

When it comes to reaction times, not much can be said. My reflection about it is that the results seem less clustered when I’m tired or sleep deprived, which is especially true for the Go/No-go test. Sometimes I have just as good reactions as on monophasic sleep, at other times I do significantly worse.

The Stroop test is perhaps not interesting, in that it seems to me that the major effect is that I have learned to do it more efficiently. I may stop doing this test as it seems to not give much, but then again it is also one of the quicker ones and I don’t mind it much.

The memory tests are the most obvious, really. While I would solidly hit 7 every time on the corsi block task before starting the experiment, I now rarely do reach 7, nearly always failing on 6 and sometimes even failing on 5.

And for the working memory tests (2-back and 3-back), it’s even more obvious. When doing these tests, I notice a clear lack of ability to keep focus, sometimes even having microsleeps during the sequences. I have a feeling that the n-back tests will be a good thing to look at as I approach full adaptation, to see if they come back up towards 100%. Note here that it is not the “memory” component which is the problem here, it is the attention component.

Finally, then, the typing speed test doesn’t seem to say much. I do think however that the variation in the graph is a sign of something. When I am not tired, I should be able to keep the wpm on 65+ at all times.

And that is it. When I have more results down the line, I will of course share them with you. I also have a feeling that there is more interesting data analysis to be done, like cross referencing the subjective nap data with the test data, for instance.

That, however, is for another time. In the meantime, if someone wants to tinker, I could provide the raw data.