Cover Story

Sleep can be wonderfully restorative. After a long day of work you drag yourself to bed-and then you wake up seven or eight hours later, alert and recharged.

Now, researchers are finding that one reason we sleep may be that our brains, as well as our bodies, need time to rest and repair themselves. Recent studies have suggested that the brain, so active during the day, may use the downtime of sleep to repair damage caused by our busy metabolism, replenish dwindling energy stores and even grow new neurons.

These studies, along with competing and complementary research on sleep and memory (see "Let's sleep on it") and the evolutionary forces driving sleep (see "Wild findings on animal sleep"), are giving us a fuller picture of the biological imperatives behind this most basic need.

Messy metabolism

One way that sleep may aid the brain is by reducing damage caused by oxidative stress. For years, researchers have known that molecules called free radicals can damage human cells, including brain cells. These molecules form naturally whenever the body metabolizes oxygen, so the damage they cause is called oxidative stress. Because free radicals are missing one electron, they're very unstable and bind to other molecules in nerve cells and elsewhere-and when they do they damage those cells.

The body combats free radicals with neutralizing enzymes, including one called superoxide dismutase (SOD) that can break down the damaging molecules into their component, and harmless, parts. But research has found that when a rat is sleep-deprived the level of SOD in its brain drops, suggesting that sleep may be crucial for minimizing the brain cell damage caused by oxidative stress, says study author Jerome Siegel, PhD, a psychologist at the University of California, Los Angeles.

In the 2002 study, published in the journal Neuroreport (Vol. 13, No. 11, pages 1,387-1,390), Siegel and his colleagues kept rats awake for five to 11 days. At the end of the sleep deprivation period, they found that the level of SOD activity had decreased in the rats' hippocampus and brainstem.

In another 2002 study, this one in the journal Brain Research (Vol. 945, No. 1, pages 1-8), Siegel's team found some evidence of the effect that this reduced SOD activity might have-they discovered damage to cell membranes in the hypothalamus of rats kept awake for 45 hours in a row.

Free radicals can also damage other cells in the body. But other types of tissue-like muscles-can rest any time we sit still, Siegel explains. The brain, on the other hand, is hard at work whenever we're awake.

"Putting these together, the message is that sleep deprivation causes oxidative stress in the brain, and in the regions where the stress is more severe, it causes damage," he says. On the other hand, he points out, the cells in his study were damaged but not dead-so some of the damage might be reversible with sleep. The results also fit with what we know about animal sleep, Siegel adds. In general, smaller animals require more sleep than larger ones. Smaller animals also generally have a higher metabolic rate-which would produce more free radicals, and thus necessitate more sleep to neutralize them.

The brain's battery

All that messy metabolism requires energy to keep running. And over the past decade or so, several researchers have suggested that another purpose of sleep might be to replenish the energy stores that we use up while awake.

Harvard Medical School neuroscientist Robert McCarley, MD, studies the link between sleep and a molecule called adenosine triphosphate (ATP), which provides energy to cells. ATP stores energy in the chemical bonds that hold it together, and when it breaks down into its component parts it releases that energy to the cell. One of those component parts is the neurotransmitter adenosine, which McCarley and others now think the brain may use as a signal to monitor its need for sleep.

In a 1997 study published in Science (Vol. 276, No. 5316, pages 1,265-1,268), McCarley and his colleagues measured the level of adenosine in the brains of cats kept awake for six hours-much longer than a cat would naturally stay awake. The researchers found that the level of adenosine in the cat's basal forebrain-an important area for regulating sleep and wake-increased with each hour of sleep deprivation.

In a 2000 follow-up study, published in Neuroscience (Vol. 99, No. 3, pages 507-517), the researchers found that this was not true in other areas of the brain-such as the cerebral cortex-that were less important for regulating sleep.

The idea, then, is that this is a self-regulating cycle. As the level of ATP in brain cells drops, the level of adenosine rises. That rising level of adenosine sends a signal to the areas of the brain that regulate sleep that it's time to sleep.

"Adenosine is the messenger telling the cell to shut off-that you need some rest," McCarley says.

Other researchers have suggested other possible contenders for the substance linking energy to sleep. Neuroscientists Joel Benington, PhD, of St. Bonaventure University, and Craig Heller, PhD, of Stanford University, were among the first to suggest that energy restoration in the brain might be a function of sleep. They focused on glycogen, a form of glucose that provides energy to localized brain cells when they need a short-term boost.

Studies, though, have provided mixed results. While researchers have found that levels of glycogen dip in the brain after sleep deprivation in some strains of mice, other strains haven't shown that expected pattern.

"It's a mixed bag of results, with nothing conclusive yet," says Benington.

Neurogenesis

Other research suggests that sleep may contribute to neurogenesis, or the formation of new nerve cells in the brain.

For decades, scientists believed that animals and humans were born with all the brain cells they would ever have. Over the past 30 years, though, neuroscientists have chipped away at that theory, finding evidence of new brain cell growth first in adult rats, then in primates, and finally, in the late 1990s, in humans-particularly in an area of the hippocampus called the dentate gyrus.

But recent evidence shows that sleep deprivation can impede these new neurons' growth. In a 2003 study in the Journal of Physiology (Vol. 549, No. 2, pages 563-571), neuroscientist Dennis McGinty, PhD, and his colleagues at the University of California, Los Angeles, found that depriving rats of sleep for four days reduced the number of new cells in the dentate gyrus by more than 50 percent.

In a follow-up study published in October in the European Journal of Neuroscience (Vol. 22, No. 8, pages 2,111-2,116), they found that many cells that formed during sleep-deprivation didn't mature normally.

"The nice thing about this model is that it's very concrete," McGinty says. "People have wanted for years to understand how sleep could play a role in brain development and growth, and this gives you a direct way to look at structural change."

Overall, McGinty points out, it's important to remember that most of the hypothesized functions of sleep-reducing oxidative stress, restoring energy levels, promoting neurogenesis and others-are not mutually exclusive. In fact, many complement each other.

"Even something as simple as breathing has multiple functions," he says. "So it's quite likely that something as complex as sleep is in the same boat."