Dreamland: Adventures in the Strange Science of Sleep (9 page)

What the researchers didn’t tell the subjects in the study was that there was actually a much easier way to get to the answer. In each instance, the second three digits in the answer were the mirror images of the first. That meant that if the first part of the answer was 4-9-1, the second part would be 1-9-4. It was a subtle pattern that no subject recognized during the training session, even after completing a block of thirty trial runs.

After everyone knew how to solve the puzzle the long way, the researchers broke them up into groups based on how many hours they would get to sleep. One group was allowed to sleep normally for eight hours. Another was kept up all night. The third group, subjects who were trained to solve the puzzle during a morning session, was asked to come back eight hours later without taking a nap in between. Through this setup, researchers ensured that each group stepped away from the problem for the same number of hours. If the groups more or less improved equally, it would suggest that solutions to problems come after the brain has a long enough time to reflect. But if the improvement rates between the groups were different, it would suggest that something happened during sleep and dreaming that made a difference in their ability to interact with new challenges.

When the results came in, it was clear that sleep was key. Subjects who did not get to sleep before their second shot at the puzzle showed little improvement. Those who slept eight hours, meanwhile, solved the task 17 percent faster. But that wasn’t all. The subjects who figured out the hidden, easy solution to the puzzle completed each set approximately 70 percent faster than their peers because they had to solve only the first three digits in the six-digit answer. Only one out of every four subjects in the groups that did not get to sleep caught on to the pattern by the end of the study. But almost everyone who slept eventually discovered the quick solution. Sometime in the night, their minds were able to construct a novel approach to a problem they had faced while awake. Subjects who didn’t get to sleep continued to conceive of each puzzle literally, following the by-the-book instructions handed to them by the research team. Sleeping, meanwhile, allowed the brains to develop a cognitive flexibility that led them to consider the situation in a new way.

It was as if sleep stretched the muscles of the brain, and it responded by bending its conception of facts and reality in a way that let it arrive at a new vision. While this study confirmed that sleep did in fact enhance problem solving, the question remained of whether dreaming played a role in the process. Were dreams just a part of sleep that occurred at the same time that the brain was consolidating its memories and honing its new skills, or did dreams help the brain reach its goal?

Back across the Atlantic one researcher at Harvard University turned to video games in his investigation into how the brain tags new information that later reappears in dreams. Robert Stickgold, a professor of psychiatry who was then in his early sixties, became interested in dream studies because of an experience he had had while hiking with his family in Vermont. One night, as he started to drift off to sleep, he felt like he was still on the mountain. Even though he was comfortably in bed, he had the very real sensation that he was grabbing rocks and pulling himself up. When he woke up two hours later, the feeling was gone.

A few days later, he mentioned to his colleagues that he had the strange sense that his mind was replaying its day just as he fell asleep. He then learned he wasn’t alone. His friends told him that they had had the same experience after completing intense, focused activities like whitewater rafting, or—this being a group of Harvard professors—studying organic chemistry all day. Stickgold wanted to conduct a study to see whether this was a common occurrence, but he was stuck trying to design an experiment that didn’t require his bussing subjects to Vermont and leading them up a mountain.

That’s when a colleague suggested Tetris. One of the most popular video games in history, Tetris requires players to sort falling pieces of assorted shapes into straight lines while listening to the soundtrack of a Russian folk song. As anyone who has played the game knows, there is something about it that sticks with you when you are sleeping. Stickgold assembled a group of college students, taking care to include those who had never played the game before and those who had spent more than fifty hours on the game. As part of the study, Stickgold let the subjects fall asleep normally in rooms in his sleep lab. He woke them up not long after and asked what they were dreaming about. Approximately three out of every five replied that they saw falling Tetris pieces. The challenges that the brain had grappled with during the daytime replayed in the mind as the subject went to sleep, just like with Stickgold’s sensation of climbing over rocks after his day in Vermont.

More reports of Tetris dreams came on the second night of the study. It seemed that once the mind realized that being asked to sort falling shapes wasn’t a fluke, it decided to devote extra time to figuring out a strategy. All of the subjects who were new to Tetris reported seeing game pieces in their dreams, while only half of the experts did. Intriguingly, Stickgold included in the study several subjects who regularly suffered from amnesia. Among this group, too, he received reports of dreams of falling shapes, even though the subjects could not consciously remember playing the game. Each person’s brain used sleep as a time to rehash what it experienced while awake. When subjects played the game a second time, their Tetris dreams appeared to help them improve more than simply time alone.

Other studies showed the same thing. Researchers in Brazil, using the violent first-person shooter video game Doom instead of Tetris, recruited volunteers to play the game in which they blasted zombies and monsters with shotguns, knives, and chainsaws for at least an hour before they fell asleep. When they were woken up from REM sleep and asked what they were dreaming about, monsters and chainsaws topped the list. Just like with Tetris, subjects who spent more time dreaming about the game demonstrated a greater improvement in their skills the next time they played than those whose brains hadn’t relived its experiences during sleep.

Across town from Harvard, at the Massachusetts Institute of Technology, a neuroscientist named Matthew Wilson found that new information a rat learned during the day was incorporated into its dreams as well. He implanted tiny electrodes into the brains of his test subjects. He then recorded each rat’s brain waves as the rat searched through a maze. Wilson focused on a cluster of neurons in the hippocampus that were responsible for storing memory, including memories indicating that a particular place contained food or was difficult to maneuver around—a job that is very similar to what the hippocampus does in our own brains. While the rats slept, Wilson noticed that the pattern of their brain waves almost perfectly matched what he saw while the rats were awake and moving through the maze. The data were so similar that Wilson was able to tell exactly what part of the maze the rat was dreaming about. The animals were replaying what they went through during the day and committing it to memory.

Just as Crick and Mitchison proposed, sleep appeared to be the time Wilson’s rats focused on new and important information. Stickgold decided to take this line of study further, using the next best thing to electrodes implanted in human heads: more video games. Thanks to the Tetris experiments, Stickgold convinced Harvard to buy an arcade game for him called Alpine Racer 2 and install it in his sleep lab. The game was part of a new line of machines that required players to move their whole bodies, rather than just their thumbs. To play the arcade game, someone steps onto two platforms, each of which represents a ski, and grasps two movable blue handles, which stand in for poles. Players must move their legs and arms simultaneously to dodge trees and slalom through gates in a rough approximation of tackling a black-diamond run in Colorado. The immersive experience was a clear parallel to hiking in Vermont or any other full-body activity that melds decision making with physical movement, a taxing cognitive process in which time and patience lead to skill.

Stickgold designed a study that would test whether humans continue to dream about new information throughout the night. His goal was to determine how novel data interacts with what the mind already knows. Like in the Tetris experiment, Stickgold recruited volunteers, who then played the game for forty-five minutes and slept in his lab that night. But this time, he decided to wait until some subjects had completed one or two sleep cycles, the roughly ninety-minute loops that the brain goes through every night, before he would wake them up and ask what was going on in their dreams. As in the Tetris experiments, almost half of the subjects who were woken up early in the night had dreams that seemed out of the video game, dreams of skiing or hiking in the mountains. But as the night progressed, the dream reports became less straightforward. Subjects began to say they were dreaming of things like moving quickly through a forest as if on a conveyer belt.

The literal replay of new information had started to evolve into analysis. Once an initial phase of dreaming passed, the brain began finding connections and associations with the data embedded on its memory cards. This stage of dreaming, which fused elements of skiing with what the subjects already knew, occurred later in the night, a time when the adult brain spends longer amounts of time in REM sleep. As the subjects slept, their brains conducted free-association sessions, desperately searching for connections. That may explain not only why the dreams we remember upon waking up after a long sleep seem so strange, but also how we craft new ideas from our memory. The open interplay of emotions, facts, and fresh information allows our brains to see things in a new way. A golfer waking up with a better way to grip a club suddenly seemed less like a genius, and instead the natural outcome of what sleep does for the brain.

Stickgold’s contention that the brain consolidates information during sleep in order to make new connections was supported by research conducted by one of his former students, a red-headed Englishman named Matthew P. Walker, who is a professor at the University of California, Berkeley. Working off of Stickgold’s research, Walker decided to look at how sleep affected what is known in neuroscience as brain plasticity, which is essentially the way the brain remolds and updates itself when it learns a new skill or stores a new memory. At the time, Walker was fresh from a postdoctoral study at Harvard. He had been part of a team that found that subjects who were tested on their ability to type a string of numbers completed the task 20 percent faster when they were given a chance to sleep before approaching it a second time.

In his work at Berkeley, Walker asked right-handed subjects to type a five-number sequence using their left hands. It was an unfamiliar task that lowered the chances that a subject could skew the data because of his or her natural ability. By analyzing the time that it took them to hit the keys, Walker found that almost all of the subjects subconsciously broke the string of digits into smaller, easy-to-manage chunks, much like you might remember your Social Security number by slicing it into a group of three, two, and four digits. You can hear this process at work when you say your number aloud and find yourself breaking into a singsong rhythm. Walker then had his subjects come back after a night of normal sleep. Just like in the studies of Wilson and Stickgold, time spent sleeping improved performance. After eight hours of sleep, nearly every person in the study typed the numbers in one smooth motion.

Not all sleep gives the brain the same benefits, however. Timing matters. The smoothing effect Walker identified depends on the quality of sleep a person gets immediately after learning something new. The most important period of learning occurs in the first six hours of the night. In one study, researchers trained subjects to perform a motor-skill test. One group was awakened after less than six hours of sleep and trained to perform a second, unrelated task. The other group was allowed to sleep normally. Subjects who did not have their sleep interrupted were able to complete the motor-skill test by an average of 21 percent faster the next day. Those who were awakened, however, improved by an average of only 9 percent. Their brains, it appeared, were interrupted at a crucial time.

In another study, researchers trained subjects to complete a typing exercise. Some subjects were deprived of sleep for one night while the rest were allowed to sleep normally. The next night, however, every person in the study was allowed to sleep as long as he or she wished. By the end of the several-day study, all of the subjects had slept roughly the same total number of hours. Yet even when cumulative sleep hours were more or less equal, a subject’s performance clearly reflected how many hours he or she slept the night after learning the new skill. Those who didn’t get to sleep that first night consistently lagged behind those who did. The brain’s initial shot at consolidating new information into memory mattered more than the simple passage of time, or extra sleep on a second night. The adage that practice makes perfect looked to be only half right. Success depended on practice, plus a night of sleep. “Sleep is enhancing that memory so that when you come back the next day you’re even better than where you were the day before,” Walker says.

But what if you aren’t always able to get a full night’s sleep? After all, many of us sleep fewer hours than we would like to after picking up important information—say, what a client expects to spend on our company’s products next year, or how to use an expensive new computer program—not because we want to but because we are forced to. Does knowing that we have deprived ourselves of a vital time for learning and innovation make it all the worse?

Not necessarily. If you can’t get in a full night’s sleep, you can still improve the ability of your brain to synthesize new information by taking a nap. In a study funded by NASA, David Dinges, a professor at the University of Pennsylvania, and a team of researchers found that letting astronauts sleep for as little as fifteen minutes markedly improved their cognitive performance, even when the nap didn’t lead to an increase in alertness or the ability to pay more attention to a boring task. Researchers at the City University of New York, meanwhile, found that naps helped the brain better assess and make connections between objects. Test subjects in their study were shown pairs of objects and told that they would be tested on their ability to remember them later. One group was given a ninety-minute break to take a nap, while the other group spent that time awake watching a movie. Subjects came back to the testing room expecting to complete the simple memory puzzle. Researchers instead asked them to describe the relationships between the objects that made up each pair, rather than recall the pairs. Again, the amount of time that each subject slept determined how he or she performed on the task.