2007 : WHAT ARE YOU OPTIMISTIC ABOUT?

[ print ]

Professor of Applied Mathematics, Cornell University; Author, The Joy of x
Applied mathematician, Cornell University; Author, Sync

Why Do We Need to Sleep?

In the coming year, we're going to witness a breakthrough in our understanding of what sleep is for.

It surprises most people to learn that this is even a question. Every other basic bodily function—like eating, drinking, breathing, urinating, or defecating—has a pretty clear purpose. Our bodies and brains need food, water, and oxygen to stay alive and to replenish themselves, and if they didn't rid themselves of the byproducts of this metabolism, we'd be awash in our own toxic waste. Likewise, sleep must be for something important. We all spend decades in this strange state, immobilized, unconscious, and vulnerable. But what exactly does sleep do for us?

Parents tell their children they need to sleep because they're tired and need to rest. But of course rest is not good enough. Lying still for eight hours is no substitute for sleep. My own mother had a different theory. She said I needed to sleep because I had too much "sleepy gas." It had been building up all day long, and so I needed to sleep to get rid of it. In fact, scientists observed a long time ago that if you keep a sheep awake continuously for several days and then inject some of its cerebrospinal fluid into another, well-rested sheep, that sheep will fall right asleep, presumably because some naturally-occurring sleep substance had reached a soporific level in the donor. But this line of research never quite solved the puzzle. Although a number of putative sleep substances have now been identified, we're not sure how they might work biochemically, or how sleep (as opposed to mere rest) might break them down.

Other sleep-deprivation studies done in the early 1980's took a more brutal approach, keeping rats awake for weeks until they died from a lack of sleep, and then looking for the precise cause of death. Such studies (now outlawed) could not pinpoint any specific culprits, such as particular organ failures. One striking observation, however, was that the rats ate much more than normal and yet wasted away. Their metabolism seemed to be wrecked. So maybe sleep is for energy regulation, in some unspecified way. Other popular theories are that sleep is for tissue repair, or immune function, or for consolidating learning and memory.

The new development, and the cause for optimism, is an original approach to the problem that makes the first quantitative, testable predictions about the function of sleep. Two physicists, Van Savage (Harvard Medical School) and Geoff West (Santa Fe Institute), have analyzed how sleep varies across mammals of different species. Normally physiological time ticks slower for bigger animals. For example, elephants live much longer than mice and their hearts beat much slower. The interesting thing is that both animals' lifetimes and pulse times scale in the same way with their body mass—in direct proportion to their mass raised to the 1/4 power—with the curious implication that the hearts of mice and elephants will typically beat the same number of times in their lifetime.

What is so strange about sleep in this regard is that it behaves differently from all other physiological times. It's backward. Mice sleep longer than elephants, not shorter—about fourteen hours a day compared to four. Savage and West interpret this as evidence that sleep is related to cellular repair. They note that cells produce destructive byproducts, such as free radicals, in the course of normal metabolism, and they hypothesize that sleep somehow helps repair the damage that ensues. (In this view, the mouse needs to sleep longer to clean up all the byproducts generated by its revved-up metabolism.). Then, using classic laws about how metabolic rate varies across different species, they derive mathematical predictions about how sleep duration should vary with an animal's size. But which size is most relevant—brain size or body size? The key is that they are not proportional. If sleep is for repairing the brain, Savage and West derive one prediction; if it's for repairing the body, they derive a different prediction. When they finally confront their model with the available data, they infer that sleep is mainly for repairing the brain, not the body. So much for beauty sleep.