On Sleep Debt

I am extremely careful to sleep enough. Lack of sleep can make you sick! (Likewise, if I don't exercise at least once every two week I get sick.)

sleep deprivation often triggers erratic behavior in people with certain psychological conditions, such as bipolar disorder (formerly called manic depression),

(http://www.sciam.com/article.cfm?id=lunacy-and-the-full-moon&page=2) Friends have been horrified to hear that I schedule 10 hours for sleeping every day, when they routinely sleep for 6 hours. It bears for an explanation: I schedule so much time available for sleep only so that if I want to think in the evening, while in bed, I can do so and still sleep enough. Likewise, if I wake up early in the morning and feel well, I do not force myself to sleep more and start getting things done.

The (1h) video above seems to give an explanation or justification of why it is a good idea: in an experience they forced young military people to stay in bed for 14 hours every night for a week. The first night they did almost sleep 14 hours, and the sleeping time lowered to an average sligthly above 9 hours per night: the interpretation is that as the subjects recover from their sleep debt, they converge to their ideal time of sleep. The most interesting part of this video was when the speaker showed how sport performance improved as sleep debt was reduced to zero: I would like to see a similar study on the performance of students, or even scientists!

I noticed along the years how the scientists that I admire the most (e.g. Rosie Redfield, David Kirkpatrick, Ian Munro,...) are all careful to exercise regularly, but never heard any mention of special care of sleeping pattern or rates, beside the usual complain of jet lags and the like. How much time does a scientist sleeps on average? Is there a difference between "theoretical" and "applied" scientists?