Learn While You Sleep

German researchers have found that by using the right timing and electrical stimulation, they can improve a person's ability to remember facts.

A team of researchers in Germany has found that a certain type of memory improves when a person's brain is stimulated with a mild electric current during a particular phase of the sleep cycle.

Neuroscientist Jan Born, of the University of Lubeck, has been studying the role of sleep in human memory for the past decade. In recent years, there's been evidence to suggest that REM sleep and non-REM sleep serve to strengthen neuron connections for different kinds of memories. While the scientific community is split on just how these phases influence memory, Born and his colleagues have recently focused on non-REM sleep, specifically the initial, drowsy, slow-wave phase. They're interested in its role in strengthening declarative memories, otherwise known as fact-based memories, as opposed to other types of memory such as motor-skill, or procedural learning.

"You remember the things consolidated during sleep better than not during sleep," says Born. "Our research is finding out which stages are more important for memories."

In previous studies, scientists have found that different phases of sleep are characterized by different patterns of brain activity. The most well-known is that of REM sleep, a period of heightened activity within the cortex when dreaming usually occurs. During a full night's rest, REM sleep occupies 20 percent of a person's sleep.

The other 80 percent is devoted to non-REM sleep: a period of relatively little action, mostly consisting of slow, synchronized waves that travel across the surface of the brain. Less is known about this relatively dormant phase, and Born theorizes that, in fact, it's not a dormant phase at all. Rather, this slow-wave sleep may serve as a sort of "covert replay" of what was learned before an individual went to sleep. That is, this synchronized activity may somehow reinforce neuron connections involved in establishing long-term memories, particularly for facts.

To test this hypothesis, Born and his colleagues applied electrodes to the scalps of 13 medical-student volunteers. The whole group was subjected to successive nights of testing through various kinds of stimulation. Before going to sleep, they were asked to learn a list of word pairs (a declarative-learning task). They also learned a finger-tapping exercise normally associated with motor-based learning. As they slept, researchers electrically stimulated their brains with a mild, .75 hertz current at 5-minute intervals during an early, 30-minute period of non-REM sleep. The current applied matched the naturally occurring slow-wave frequency of non-REM sleep, and it was low enough not to disturb volunteers as they slept. On other nights, researchers altered the current, and they also applied currents during both non-REM and REM phases of sleep. Sometimes, as a control, they applied no current at all.

After each night's sleep, volunteers, once fully alert, were asked to recall the tasks they had learned the day before. Of all the trials, they improved most significantly after having been stimulated at .75 hertz during non-REM sleep. What's more, the improvement, although small (8 percent), was only seen on the word-pair task, suggesting a link between non-REM sleep and fact-based, or declarative, memory. "I think it's a very good indicator that makes us think we induced a really physiological pattern in these brain networks," says Born. "This [activity] in turn stimulates the brain to replay these memories."

He says that while the improvement measured was small, the fact that the experiment was tested on "high performers"--medical students who are used to memorizing large amounts of information in a single sitting--suggests that any improvement is significant.