In a study at the Walter Reed Army Institute of Research in 2003, for example, scientists examined the cognitive effects of a week of poor sleep, followed by three days of sleeping at least eight hours a night. The scientists found that the "recovery" sleep did not fully reverse declines in performance on a test of reaction times and other psychomotor tasks, especially for subjects who had been forced to sleep only three or five hours a night.

In a similar study in 2008, scientists at the Karolinska Institute in Stockholm found that when subjects slept four hours a night over five days, and then "recovered" with eight hours a night over the following week, they still showed slight residual cognitive impairments a week later, even though they reported no sleepiness.

But in another study, also at the Walter Reed Army Institute of Research, scientists found that people recovered much more quickly from a week of poor sleep when it was preceded by a "banking" week that included nights with 10 hours of shuteye. In other words, if you know you have a week of little sleep ahead of you, try loading up on sleep beforehand, not simply afterward.