Night Life

What is the purpose of the neural chaos of dreaming? Scientists are still puzzling that out. In 1983, Nobelist Francis Crick, of the Salk Institute in La Jolla, suggested that the brain was actually "reverse learning," that REM sleep allows the neurons to spew out each day's spurious and extra stimuli, cleansing the brain. "We dream to forget," Crick wrote, to enormous outcry. In 1986, he revised the hypothesis, noting that perhaps we dream to reduce fantasy and obsession—that dreams are a way of forgetting material that might otherwise needlessly intrude on everyday life.

Then, in 1994, two researchers showed that our ability to learn seems dependent on REM sleep. Scientists Avi Karni and Dov Sagi, at Israel's Weitzman Institute, found that if someone is trained in a task and allowed a normal night's sleep, they will show improvement the next day. But if sleep is interrupted in each REM cycle, they show no improvement at all.

And the particular cycle of REM that gets interrupted is crucial. It's REM sleep in the last quarter of the night that counts. Bob Stickgold trained 57 individuals in a task and then tested them 3, 6, 9, or 12 hours later the same day, or overnight after an interval of 13, 16, or 22 hours. The task involved visual discrimination: a subject looks for diagonal lines against a background of horizontal lines.

The time interval had no influence on performance; the amount of sleep did. "If they had less than six hours of sleep, they did not improve," says Stickgold.

One might simply conclude that people need a lot of sleep in order to learn. The truth seems to be: they need certain cycles of sleep, and when awakened before their last REM cycle, the brain is unable to consolidate the memory of the task. But Stickgold and his colleagues found that more than REM cycles were at stake.

"A student of mine did another experiment and found that the amount of slow-wave sleep in the first two hours of the night is highly correlated with the amount of learning as well." How might the two sleep cycles—REM and deep slow-wave sleep—work together? There may be a two-step process of memory enhancement. "We know that levels of acetylcholine are high in REM sleep and low in slow-wave sleep. Perhaps as you cycle from one to the other, you're passing information back and forth between different parts of the brain. It's as if the brain is holding a conversation with itself and identifying exactly what it needs to know."

Stickgold thinks REM sleep may have yet another purpose: to actually alter intrusive experiences and memories from the day. "I was putting my son, who is ten, to bed after a day of skiing together. We were lying there with our eyes closed and I said, 'I feel I'm back on the ski slope.' He said, 'Really? I'm on the ski lift.'" There's a tendency to have an intrusive replay of novel experiences when you go to sleep, says Stickgold, especially ones that involve the vestibular system of the brain, which plays a role in balance. "If I fall asleep, go through one REM cycle, and wake two hours later, the feeling is gone. I can't reproduce it. Something has happened to that memory in those two hours."

Stickgold is looking at this effect in people who play the computer game Tetris, which requires rotating small blocks that float down the screen. "More than one person has told me that the day they first got hooked on the game, they went to bed, closed their eyes, and could see these blocks floating. It's gone the next day. Something in your brain in that first two hours has taken a memory that at sleep onset is incredibly intrusive and altered it so that it no longer behaves that way."

Rosalind Cartwright, Ph.D., the doyenne of dream research, has also found that sleep softens intrusive experiences, especially depressing feelings and moods. Director of the sleep disorders service and head of psychology at Chicago's Rush Presbyterian-St. Luke's Medical Center, she has evidence that dreams help regulate and stabilize mood, defusing negative feelings.

By observing sixty normal and seventy clinically depressed adults, Cartwright found that among those who had a mildly unpleasant day or experience, dreams were negative at the beginning of the night and became pleasant by the end. Among the clinically depressed, dreams were bland at the beginning and negative by night's end. "Normal individuals wake up in a better mood after a depressing day," she says. "Depressed individuals wake up feeling worse."

Cartwright adds a coda: "I'll tell you the kicker: a few of the depressed people showed the opposite pattern. Their dreams got more positive across the night. And those were the ones who got over their depression. You could predict it from a single night of dreaming."