Sleep on it

In the weeks before university exams, the 24-hour cafés near Hanoi's big campuses fill with students who have no intention of going to bed. "I'll sleep when it's over," a second-year economics student told me at 2 a.m., three empty cups of iced coffee beside her laptop. The logic seems sound: more hours with your notes should mean more in your head. Research on sleep and memory points the other way. A large body of work suggests that sleep is not a break from learning but one of its working stages, the time when the brain sorts out what to keep.

Some of the clearest evidence comes from rats. In the 1990s, researchers recorded cells in the hippocampus, a part of the brain closely involved in memory, while rats ran along a track. Particular cells fired at particular points on the route. When the animals later slept, the same cells often fired again in roughly the same order, as if the rats were replaying the run. Many scientists think something similar happens in people: during deep sleep, the day's experiences are rehearsed and gradually moved into longer-term storage.

Studies on volunteers point in the same direction. In a typical experiment, people learn word pairs or a sequence of finger movements in the evening and are tested the next morning. Those who sleep normally tend to remember more than those kept awake. The differences are modest in some studies, but they turn up again and again. In some studies, the gap remained even after the sleep-deprived group had caught up on sleep over the following nights. Sleep lost on the night after learning, it seems, cannot simply be made up for later.

Sleep may also do more than file memories away. In a well-known 2004 study at the University of Lübeck in Germany, volunteers worked on number puzzles that contained a hidden shortcut. After a night's sleep, they were more than twice as likely to spot it as people who had spent the same hours awake. Findings like this lend weight to the old advice to sleep on a difficult problem. The sleeping brain appears to reorganise information as well as store it, picking out patterns that are hard to see when we are tired.

None of this makes revision pointless. Sleep can only strengthen what has already been learned, so the brain needs material to work with. What the research does suggest is that study and sleep work best as partners. Spreading revision over several days, with proper nights in between, tends to beat a single all-nighter, and in some studies even a short afternoon nap has helped. The student in the café laughed when I mentioned this. "Tell that to my timetable," she said, and opened another chapter.