mjcurzi, a fellow attendee of the Rationality Mini-Camp, is organizing an experiment among the minicampers. If we coordinated, we could get a larger n than either group alone.
UPDATE: If you are a DC lesswronger interested in participating in the joint experiment, you can contact mjcurzi directly.
For the record, one suggestion at the meeting* was simple arithmetic problems a la Seth Roberts.
* Yes, I know who made the suggestion; I don't remember his name. I apologize.
Edit: I have been informed that his name is Jeff - thanks, atucker!
Is there any good test for sleep deprivation? I've been wondering if polyphasic sleep really works as well as the proponents tend to believe.
A Psychomotor Vigilance Task is what I've been using for my sleep dep experiments. It maps pretty well to my subjective experience of tiredness and reveals my worst moments in all their quantitative horror. (Once, at my worst, it took me about 15 seconds to respond to a stimulus.)
If you're trying polyphasic sleep, I would recommend everyman over uberman from my experiences.
One of our members tried polyphasic sleep and quit it after 5 days when he realized that he could no longer adequately understand physics papers.
This highlights one thing that many polyphasic experiments get wrong. As far as I know (and I should know, having tried repeatedly to get through adaptation for 6 months, then stopped when I got bad flu) polyphasic makes everybody very sleep-deprived for the first week, and usually also for the second week. Only after that do you start getting used to it. It will take even longer to adapt the more you break the polyphasic routine.
Not that I'm recommending polyphasic. Humans have been fine-tuned by evolution for biphasic and monophasic.
I have been too lazy to check if there are dedicated tests for sleep deprivation, but a timed test of something that you should be able to do quickly and easily would work, right? Sudoku, an easy crossword in your daily paper, any of the online IQ tests. IIRC somewhere in gwern's comment history here he's replying to a guy who did polyphasic sleep for a while who found that after about three days he couldn't learn anything (or so he interpreted it based on being unable to learn any new Anki cards). Apparently one proper deep long sleep reverted that. Check brain workshop for a suite of tests anyway.
That would be http://lesswrong.com/lw/5n0/optimizing_sleep/44zu
As before, I'll suggest testing yourself against Gbrainy and dual n-back (as well as Anki, I suppose). I recently recovered my old polyphasic sleep logs off a backup CD I didn't know I had, and unsurprisingly, my Gbrainy scores were consistently bad; not that I ever felt I 'adapted', anyway.
On a side note, I just finished up a few months of on-and-off melatonin use, recording nights with my Zeo. I haven't updated my site yet, but the summary is: total sleep did fall by ~1 hour, REM fell by ~25 minutes, fewer awakenings, and morning feel was the same. So, I can definitely recommend Zeo for tracking things to do with sleep.
I think I read somewhere that researchers measure the time someone needs to fall asleep for that.
During our latest meetup, the DC Less Wrong group has decided that we are interested in experimentally testing various lifehacks on ourselves (on an opt-in volunteer basis of course).
We need two things:
Do any other groups have any measurements that they take to track their various attributes? Anything that they'd be interested in testing?