The Physiology of Sleep
New developments challenge our understanding of non-wakefulness


by Kayt Sukel

April 28, 2008

Most animals sleep—mammals, reptiles, even fruit flies. And while plenty of behavioral studies show the restorative effects of sleep and the detrimental cognitive effects to vigilance and short-term memory tasks when it is withheld, it has not been clear exactly why we must periodically lose consciousness in this way. But some recent studies examining its neurobiological mechanisms have led to new hypotheses about sleep.

The stages of sleep

“No matter what physiological or behavioral measure you pick, it will change when you sleep: Heart rate, respiratory rate, the body’s response to carbon dioxide, metabolism, immune response, even your posture is different in sleep,” says James M. Krueger, a sleep researcher at Washington State University. “Historically, medicine has not studied these changes. They’ve dealt with the physiology of waking.”

But with advancements in medicine and medical technology, researchers have been able to examine sleep in more detail. We now know more about its two main stages: rapid eye movement (REM) sleep, characterized by the eye movements and an inert body, and non-rapid eye movement (NREM) sleep, which can be divided into four successive stages, the first two considered “light” sleep, or drowsy sleep, and the following two slow wave sleep (SWS), or deep sleep.

“During SWS, electrophysiological activity in the brain changes dramatically,” says Steffen Gais, a researcher in Jan Born’s laboratory at the University of Lübeck in Germany. “During wakefulness, activity is fast and chaotic. But during SWS, you have these electrophysiological waves of activity where all the neurons fire together in a very slow rhythm.”

The mechanisms of sleep

What causes those electrophysiological changes in the brain? How, exactly, does the brain know that it’s time to sleep? Sleep researchers used to believe that a network in the brain, a circuit of neurons, imposed sleep on the brain.

“The problem with that paradigm is that there is a philosophical inconsistency,” says Krueger. “Who is telling that network that the brain should sleep?”

Krueger views sleep regulation as happening at the level of neuronal assemblies, or collections of neurons that work together in the waking brain. Krueger has found that cytokines, signaling proteins released along with neurotransmitters by the neurons, interact with the neuronal receptors. As the neuron fires over time, the cytokines build up, changing the electrical properties of the cell.

“What you see is that the input to the neuronal assembly then induces a different output, which is by definition a state change,” Krueger says. That state change explains the electrophysiological differences between sleep and waking, as well as for the different stages of sleep.

 “Changes in neuronal firing change the state of the network independent of any exterior thing. There’s no little demon telling you it’s time to sleep, just how often that neuron is used,” he says.

Sleep as a memory consolidator

The state changes of slow wave sleep are of particular interest to learning and memory researchers. Jan Born, director of the department of neuroendocrinology at the University of Lübeck, believes sleep’s main function is to consolidate memories, or to encode the important parts of your daily learning so it is stored for the long term.

“The acquisition of information and retrieval of information—this is what is done in the waking brain,” Born says. “But if you want to retain information over a long time, then you need an offline mode in the brain to reprocess what you’ve learned in any given day.”

During slow wave sleep, the slower oscillations of activity generated by the brain’s neocortex trigger the hippocampus, reactivating the memories of what you’ve learned that day. The neurons start firing at the same spatio-temporal pattern as they did during learning, which transfers the information back to the neocortex for long term storage.

Being “offline” is critical to that reactivation process. If those memories were reactivated during wake times, it would disrupt normal information processing. “With pharmacological treatment, we can establish conditions in humans where the brain starts this consolidation process while the person is still awake,” says Born. “But when you do this, the brain cannot discriminate between the reactivated memories and external inputs and the results are hallucinations.”

Synaptic homeostasis

Giulio Tononi and colleagues at the University of Wisconsin–Madison’s Center for Sleep and Consciousness believe that sleep has a different purpose. They argue that synaptic potentiation, or the strengthening of neuronal connections they think underlie learning and memory, is regulated during deep sleep.

“During waking, we are always learning, even when we do not realize it,” says Chiara Cirelli, a researcher at the center. “This learning results in synaptic potentiation and connections between neurons are greater at the end of the day. This synaptic activity is expensive in terms of energy in the brain and we just can’t afford to keep growing synapses larger and larger.” The group postulates that the fundamental function of sleep is to downscale the size of synapses after a long day’s learning to get them ready for the next day’s activity. The idea seems contrary to both the memory reactivation data and other behavioral studies that shown that sleep improves memory.

But in a paper published in the February 2008 issue of Nature Neuroscience, the group offered molecular and electrophysiological evidence supporting their hypothesis. The researchers measured the amount of proteins concentrated in synapses in rats both after waking and sleeping and found that almost 50 percent more of the proteins were present after waking. In the companion electrophysiological study, the group induced electrophysiology activity in one brain region and that recorded follow-on activity in other regions.

“If synapses potentiated during sleep, then for the same amount of starting stimulation, the response should be bigger,” Cirelli says. “When awake, when we induced a response, it was very big. But when participants were asleep, the response was smaller. It’s a complementary indication that there is an increase in synaptic size during waking.”

Potential for multiple functions

Both Born and Cirelli caution that their hypotheses are not necessarily mutually exclusive.

“These slow oscillations may have a two-fold function where they act globally across the whole neocortex to downscale synapses but act locally to create these memory reactivations in the brain,” says Born.

Cirelli agrees. “It is difficult to tease apart. It would require a technique that could look at single synapses and track which potentiated during a particular waking test and then what happens to that same synapse during a sleep period. We’re just not there yet, technically speaking.”

But whether future research shows that sleep has one function or more working in concert, James Krueger looks forward to seeing more neurobiological results come to light.

“This is a very exciting time in the sleep world, both clinically and in terms of basic science,” says Krueger. “It’s very different from when I started 30 years ago.”