[an error occurred while processing this directive]
May 07, 1998
Harvard
University Gazette

 

Full contents
Notes
Newsmakers
Police Log
Gazette Home
Gazette Archives
News Office
Feedback

SEARCH THE GAZETTE

 

Sweet Dreams Can Lead to Better Learning

By Cassie Ferguson

Gazette Staff

The students with the wildest dreams during finals may be the ones who do best on tests: Not only will they be well rested when they wake up, but dreaming may, according to Harvard researchers, even help them learn.

Robert Stickgold, assistant professor of psychiatry at Massachusetts Mental Health Center, knew that dreaming allows sleepers to subconsciously solve problems and to use what they've learned when they wake.

"All sorts of wacky things get put together in dreams that don't make sense when you're awake. Sleep on a problem and dissociated pieces will fall together," said Stickgold. "Then when you wake up you have the solution. We wanted to know - why is that?"

In a study presented two weeks ago at the Society for Cognitive Neuroscience's annual meeting in San Francisco, Stickgold had 56 volunteers practice picking out patterns flashed on a computer screen. Then the volunteers spent the night in the lab so he could monitor their sleep; the next morning he retested them.

While not all subjects showed improvement in the morning, Stickgold found that 80 percent of the variation among the subjects could be predicted by their sleep patterns. The improvement was most dramatic for those who had gone through both a period of deep sleep characterized by slow brain waves early in the night and a period of dreaming called rapid eye movement (REM) sleep late in the night.

"We seem to need both early slow wave sleep and late REM sleep," said Stickgold. "Fortunately that's how we're naturally designed to sleep."

Stickgold said learning happens with the formation of new connections and the strengthening of old associations in a part of the brain called the cortex. Somehow, said Stickgold, the mind learns the new task by mixing short- and long-term memories while we sleep, reinforcing what we learned during the day.

A similar connecting of thoughts relates to Stickgold's other research studies, where he found that dreaming allows a person to reorganize their old thoughts and make new connections through the irrational association of unlike elements.

As the brain slips into a dream it loosens its grip on logic and reality, allowing juxtapositions of absurd events like, for instance, showing up late for a chemistry final sans clothing and carrying a volume of Robert Frost.

"When we dream, there are switches that change the way we see reality. If we acted like that when we were awake, we'd be locked up," Stickgold said. "The saving grace is that it only happens when we're asleep."

Stickgold and postdoctoral researcher Ruth Propper hypothesized that if they roused people from their dreams, their minds would still be processing information in a disjointed way.

To test this theory, the researchers monitored the sleep of 12 people, waking them from both dream and nondream sleep to take a timed word test. The subjects saw a word flashed on a computer screen, then a series of letters that replaced it. The person then had to decide whether or not the letters spelled out an actual word.

During regular waking hours, Propper said, people automatically use the first word as a subconcious hint of what the next word might be. They can identify the word "cold" in the pair "hot" and "cold" more quickly than the word "rat" in the unrelated pair "rat" and "cold." If the first word is only weakly related, like "snow" in "snow" and "cold," their reaction time will be somewhere in between.

The dreamers took just as long as people who were awake to identify words preceded by unrelated words. But now the weakly related words helped people more than the normally strongly related words.

"You wake them up from REM [dreaming] sleep and to them the indirect associations make more sense," said Propper.

If Stickgold could prescribe a perfect eight hours of sleep, it would start with two hours of serious slow-wave log-sawing followed by a four-hour interlude of cycling between that and dream sleep and a two-hour finale of just dreams.

Unfortunately, said Stickgold, science has yet to devise ways to prescribe types of sleep. Another worrisome matter, said the 52-year-old, is that as people age, they trade an extended interlude for the opening set of slow wave sleep. That means that older folks are less likely to be able to learn skills like the ones in the study while they snooze, said Stickgold.

"You really might not be able to teach an old dog new tricks -- at least, not in his sleep."

 


Copyright 1998 President and Fellows of Harvard College