* * Search the Gazette
Harvard shieldHarvard University Gazette Harvard University Gazette
* Harvard News Office | Photo reprints | Previous issues | Contact us | Circulation
Current Issue:
November 20, 2003

News, events, features

Latest scientific findings

The people behind the university

Harvard and neighbor communities

Scores, highlights, upcoming games

On Campus
Newsmakers, notes, students, police log

Museums, concerts, theater

Two-week listing of upcoming events

Subscribe  xml button
Gazette headlines delivered to your desktop




Walker, Brakefield
Matthew Walker watches over Tiffany Brakefield as she sleeps on a memory. The electrodes pasted on her face record the various stages of sleep and dreams. A light stage of nondreaming sleep is best for memory enhancement. (Staff photo Rose Lincoln/Harvard News Office)

New stage of memory found

Combines with sleep to make better memories

By William J. Cromie
Harvard News Office

Practice, as it turns out, does not make perfect. You need to sleep on it, then destabilize it, according to researchers at Harvard Medical School.

It's been known for a while that sleep helps consolidate certain memories; that's probably a major purpose of sleep. But the latest experiments show that a good night's rest can enhance your performance as much as 20-30 percent. Not all memories, just those of motor skills such as riding a bicycle, throwing a ball, or playing the piano.

Still, it's a big finding, especially when combined with what happens next. Reactivating the memory will cause it to destabilize - become somewhat plastic. At first that sounds bad, but when you think about it, you don't want motor memories that you can't change.

"There have been hints from animal studies that certain emotional memories, such as fear, become vulnerable to change when reactivated," says Matthew Walker, a neuroscientist who led the experiments. "But such reformation had never been shown in humans before. It's a most exciting discovery."

Aside from the satisfaction of better understanding how humans develop motor skills, the new findings could have practical applications like helping stroke victims regain basic motor skills, enhancing sports and music-playing capabilities, and ridding people of phobias and unwanted stressful memories.

Walker quickly makes it clear that the benefits of sleeping after practice do not extend to learning new things by putting a tape player under your pillow or beside your bed. "That doesn't work," he says firmly. "In the one case you are trying to force new information into the brain; in the other, it's already there and you're just trying to organize it."

Discovery by mistake

What happens while you sleep gets lots of press attention, but the reshaping of a motor memory after sleep is something new. And Walker stumbled on it by accident. "I designed one of the experiments incorrectly," he admits.

The experiments began with 100 Harvard students, ages 18 to 27, following instructions about tapping out different sequences of numbers on a keyboard. One sequence might be 4,1,2,3,4; another, 2,3,1,4,2. "It's like learning piano scales," Walker comments.

The researchers, who include Tiffany Brakefield, Allan Hobson, and Robert Stickgold, found that it takes about six hours for a sequence memory to become stabilized in the brain.

After such training, if they asked subjects to play back sequences without having slept, their performances remained the same. But after sleeping on it all night, their skills were 20-30 percent better.

"The brain continues to learn during sleep," Walker maintains. "Memory processing is so absolutely dependent on a full night's sleep that you shortchange your brain if you don't get it."

If you learn one motor memory, then try to learn a second before the six-hour stabilization, the second interferes with the first. The second memory is the one that becomes saved in the brain's computer file.

After a night's sleep, Walker thought that such interference would not occur. He had the students sleep on a memory, then quickly gave them another sequence to work on. He was surprised to find that it interfered with retaining the memory he thought had been stabilized and enhanced.

There was only one explanation. On recall, the previously stabilized memory must have reverted to a destabilized state, which was then scrambled by the input of a new memory. "In this third stage of memory, it is plastic enough to be changed by competing learning, or by memories of common elements from other skills," Walker explains. "It's part of a fine-tuning process required to build a well-rounded memory. This had never been demonstrated before in humans."

Changes in the third memory stage then require about six hours to stabilize. They are enhanced by sleep, then are ready to be tweaked when recalled again. And so the cycle goes on. "It's a very elegant mechanism, and discovering it shows that, in science, things you don't plan on sometimes turn out to be the most exciting."

Details of the experiments and results have been published in the Oct. 9 issue of the British science journal Nature.

Awakening the role of sleep

The researchers also found that a particular stage of sleep apparently is best for skill learning, a lighter stage of nondreaming sleep known as stage 2. Walker sees a connection between this stage and the development of motor skills in young children. "The amount of stage 2 sleep that infants get peaks at around age 12 months," he notes. "That's a time of intensive motor learning when they try to coordinate their arms and legs, and to stand and walk."

People spend about one-third of their lives asleep; it wouldn't make much sense to waste that time. Neither would devoting all that time just to consolidating memories of motor skills. "Evolution is so desperately efficient," Walker notes, "that I can't see 25 years of a typical life span being used only for skill memories. The mind and body must do many different things during sleeping as they do during wakefulness, including restoration of the body and recharging the immune system. There's a reason why we have a drive to sleep a certain number of hours."

A stereotype exists of hard-boiled Olympic coaches who train gymnasts, ice-skaters, and other athletes until late at night, then start them again early in the morning. "That could rob their brains of learning potential," Walker believes. "I would caution such coaches that their athletes might learn more by sleeping more."

What about students who study all night in preparation for a big test? "That behavior primes a different type of memory, declarative versus skill memory," Walker points out. "They're trying to remember things like the capital of South Dakota or the years of Harry Truman's presidency. For that kind of memory, all-nighters may not impair your memory the next day, if you don't fall asleep. But do those memories last as long as those enhanced by sleep?" More experiments are planned to answer that question.

Walker believes this new understanding of sleep and memory plasticity can be put to work to boost the performance of athletes, musicians, and doctors learning new surgical techniques. There's an obvious use, too, for helping stroke victims relearn basic movements, such as eating and walking.

More fundamentally, it can teach us more about how children develop motor skills, from standing and walking to fielding a hot grounder. The newly found plasticity period also might be utilized to fight phobias and to help rid people of unwanted memories of combat, rape, or trauma.

Walker himself has begun to look into the relationship between sleep and learning disorders that occur in depression, schizophrenia, and Alzheimer's disease. "I'd like to find out if abnormal sleep patterns in these patients impair learning or if the illness itself disturbs sleep," he says. "Which comes first, and what can we do about it? Who knows, I might get lucky again and discover something new about the brain, sleep, and behavior."

Copyright 2007 by the President and Fellows of Harvard College