You probably don't need a neuroscientist to tell you that sleep helps your brain absorb new information. But what about dreams? Some researchers have speculated that they, too, might improve memory. Now, a new study provides some of the first experimental evidence: People who dreamed about a virtual reality maze they'd encountered a few hours earlier were quicker to find a way out when tested a second time.
Lots of studies have suggested that our brains are busy while we sleep, consolidating memories of the day's events and putting them in the context of things we already know. In sleeping rodents, for example, neurons in the hippocampus fire in patterns remarkably similar to those recorded during a previous maze-running session—almost as if the animals replay the experience in their sleep. Some researchers wondered whether the rodents were dreaming of the maze, but of course there was no way to ask them.
So neuroscientists Erin Wamsley, Robert Stickgold, and colleagues at Harvard Medical School in Boston turned to a more verbal species: Harvard undergraduates. Participants in the study sat at a computer for 45 minutes and played with a virtual reality maze (see image). During this time, the researchers tested their memory by asking them to remember a particular object in the maze and find their way back to it from various starting points chosen at random. Fifty of the 99 participants then had the opportunity to take a nap while the others watched videos. The researchers used electroencephalography to monitor the brain activity of the napping students and either woke them once to ask about the content of any dreams or asked them at the end of their naps.
Not surprisingly, people who took a nap improved more on the maze—as judged by the speed with which they found requested objects—than did those who stayed awake. But the four students who reported thoughts of the maze just as they were falling asleep or dreams of the maze during their nap improved on their previous performance about 10 times more, on average, than other nappers did, the researchers report online today in Current Biology.
Although dreaming students reported seeing images of the maze or hearing the computer program's background music, Stickgold notes that the dreams didn't seem to be exact replays of their experience. That suggests to him that the dreams don't reflect an attempt by the brain to create an exact memory of what happened, but rather an attempt to put a new memory in context of existing knowledge. One student, for example, reported images of the virtual maze as well as of a trip he'd taken to visit a bat cave several years earlier. "It's as if the brain is trying to figure out, 'What have I done in the past that might be useful when I take this [maze] test again?' " says Stickgold.
"It's very tempting speculation," says Jan Born, a neuroscientist at the University of Lübeck in Germany. The idea that dreams are related to memory processing during sleep dates back to Freud, and the new study provides some of the first solid evidence of such a link, he says. Psychologist Rosalind Cartwright, a professor emeritus at Rush University Medical Center in Chicago, Illinois, says the study is a good start, but she'd like to see more evidence that dreaming per se, as opposed to thinking about the maze just before dozing off, is linked to big improvements on the maze test. "He's got both kinds of reports, and he's lumping them together and calling them dreaming," Cartwright says. "You have to differentiate."