In July, researchers led by Robert Stickgold, an assistant professor of psychiatry at Harvard Medical School, reported that a full eight hours’ sleep after learning a motor task boosts performance by 20 percent the next day.
Even a one-hour nap can improve scores on a simple visual task, others reported in May.
Perhaps even more compelling, Belgian researchers, using a brain imaging technique called PET scanning, reported two years ago that when people perform reaction time tests, certain areas of the brain become activated. These same areas “light up” again when the people experience REM (rapid eye movement, or dreaming) sleep, as if their brains were actively rehearsing what had been learned.
Over the years, researchers have found tantalizing evidence that sleep may enhance learning and memory. Some have showed that infants who learn a head-turning response have more REM sleep than those who failed to learn the response. Others, that people given 90 minutes of training in Morse code showed an increase in REM sleep. Still others, that people who did well learning French increased their REM sleep, while poorer learners did not – and that the sooner the good learners started dreaming in French, the higher their scores on French tests. (The poor learners never did dream in French.)
Case clinched, right? Sleep, particularly REM sleep, seems to boost memory, just as many scientists – and mothers – have been saying for years.
Would that it were that simple.
“Everybody knows sleep has something to do with memory – except people who study sleep and memory,” says Stickgold.
Brain researchers would like nothing better than to come up with a neat paradigm of how sleep affects memory.
It would go something like this: Learning creates chemical changes in specific cells in specific parts of the brain. When a person sleeps shortly after learning, and perhaps especially when she dreams that night, the brain takes these fragile, new memories, shuffles them around into a more permanent home, or at least a more permanent set of neural circuits. And – Presto! – the memories would be firmly “consolidated” by morning.
There are two main reasons why confidence in such a nice, simple scenario is impossible, as least for now: At the electrical and biochemical level, sleep itself is devilishly complicated. Memory, arguably, is even more so.
A night’s sleep typically goes from light (Stages 1 and 2 sleep to deeper stages 3 and 4, known as slow wave sleep). All of these are called non-REM sleep. Non-REM sleep alternates with REM, or dreaming sleep, with REM periods getting longer and non-REM periods getting shorter as the night progresses.
One reason for thinking that REM sleep may be involved in memory consolidation is that the brain behaves differently during slow wave and REM sleep. During deep sleep, the brain is relatively inactive, and electrical patterns are slow and synchronized; during REM, it is extremely active, and desynchronized..
Brain chemistry changes, too. During REM, some neurotransmitters, or chemical messengers, especially norepinephrine and serotonin are virtually shut off while others, notably acetylcholine, believed to be a memory booster, go up.
Memory is even trickier. Scientists divide memory into two basic categories, declarative and procedural. That’s the difference between “knowing that” and “knowing how,” notes psychologist Carlyle Smith of Trent University in Peterborough, Ontario.
Declarative memory involves learning facts – knowing that the French Revolution began in 1789. Knowing how is knowing, often without knowing that you know, how to turn the key in the ignition to start the car. Declarative memory is “explicit,” and usually consciously acquired; procedural memory is “implicit,” often unconsciously acquired.
The trouble is, although declarative memory is what most of us mean when we talk about memory, most of what researchers study is procedural memory – the less juicy stuff like learning finger-tapping exercises.
For declarative memory, there is little evidence that sleep, even REM sleep, has any effect, says Smith of Trent University.
“No matter what I have done – I have deprived people of sleep, I have deprived people of REM sleep, I have deprived them of non-REM sleep – and I have never seen any difference [in declarative memory] between people who got a good night’s sleep and those who didn’t,” says Smith of Trent University.
Dr. Jerome Siegel, professor of psychiatry at the David Geffen School of Medicine at UCLA and chief of neurobiology research at the VA Greater Los Angeles Healthcare System Sepulveda, agrees, noting in a paper in Science in 2001 that the evidence for such a link is “weak and contradictory.”
And a link between sleep and procedural memory? That’s stronger.
In 1991, researchers studied people learning trampolining, which required new, complex motor skills. The best learners showed increases in REM sleep (and no differences in non-REM sleep); those in control groups who expended the same number of calories but didn’t learn new motor skills showed no difference in either REM or non-REM sleep.
REM and some non-REM sleep may help with a different type of procedural task – visual learning. In a paper published in 2000, Stickgold and his team found that improvement on a visual task increased with stage 3 and 4 slow wave sleep in the first part of the night, and with REM later in the night. The real payoff, he finds, is with REM sleep in the final two hours of an 8-hour sleep.
In a sequel to Stickgold’s studies, doctoral student Sara Mednick wanted “to see if napping had the same effect as a night’s sleep.” To find out, she used the same visual memory task and tested volunteers on it at four points in the same day. They all got worse as the day went on.
So she let some volunteers take a half-hour nap between the second and third session and others, an hour nap. The half-hour nappers were able to stop the decline in performance. The hour-nappers not only stopped the decline but performed as well as they had first thing in the morning. The long nappers, by the way, exhibited both slow wave sleep and some REM sleep as well.
The bottom line in all this? There is still much that remains a mystery. Someday, maybe they’ll figure it all out. For now, “the simple answer is that we don’t know,” says Stickgold of Harvard. “You push this system just a bit and you stumble onto complete ignorance.