Alexei,
Probably the most important thing is to establish a good baseline,
probably at least a month of daily data for productivity (how much,
when, how much at a time), sleep (how much, when, how good
subjectively), and if you can, a test of brain ability (Seth Roberts
uses the timing of one digit additions, averaged over 20+ trials.
Another option is the various *al n-backs).
Because there's going to be a lot of variation in the normal rhythm,
but you want to be able to identify things that actually improve how
well the person is doing.
Now, once you have that, it sounds like the things to improve first
are productivity and sleep.
For sleep, you probably want to try out Vitamin D (with blood level
testing to prevent poisoning) (Seth Roberts argues that time of day is
important, and that morning is the right time of day -- you may want
to experiment), melatonin, and many dietary and behavioral changes,
including but not limited to:
* different kinds of fats
* eggs, cooked, and, if possible, raw yolk
* exercise, and especially pay attention to core exercise because it
doesn't sound like leg exercise or cardio was helpful
* massage, if possible
* eating at different times of day relative to sleep -- some people
need to eat right before, others nowhere near right before. some
breakfast, some no breakfast.
* relation to people -- some people wake up to talk to people, some
people stay asleep to avoid people.
* alarm clocks. There are good accelerometer based alarm clocks on
iphone and android (Sleep Cycle is one), other people need no alarm
clock.
* light in the bedroom. Some people no light, some people lots of
morning light, some people both (does anyone make autoblinds yet?)
Productivity is a lot harder. It's worth it, since you'll be trying
out diet and pills for other reasons, to consider experimenting with
nootropics. The synthetic B vitamin-related molecules are often
reported to have productivity effects: those are sulbutiamine,
pyritinol, and I can't remember the one that's made of thiamine? and a
gaba molecule, but there's a third. They cross the blood brain barrier
better than the vitamin molecules.
Nootropics also worth considering are many. ImmInst forums used to
maintain a spreadsheet that was interesting:
https://docs.google.com/spreadsheet/ccc?key=0Ah6BY-Lk_-wEdEwyMXNyQ29kTVJDT1FDSXhsWFo2VXc&authkey=CMXbgvEF&authkey=CMXbgvEF#gid=0
But I don't know if it's been updated recently. Piracetam + a good
choline source is generally considered the easiest and safest place to
start.
Many people like to look at graphs of their productivity as a way to
encourage productivity. Other people promise themselves a certain
amount of time of fun in exchange for the productive time (a la Unfuck
Your Habitat, which advises 20/10s, or 20 minutes cleaning followed by
10 minutes goofing off). You might be able to combine those to good
effect.
You know about BeeMinder of course? Paul Fenwick (pjf on twitter) has
started writing plugins that connect all kinds of things into
Beeminder so that that process of updating is automatic, if that's
helpful.
I think the big thing is to establish your set of daily trackings, and
look for changes, then attempt to magnify the changes by identifying
the variable that changed, duplicating it and seeing if the effect is
controllable.
Thanks,
Jeff