Sleep to prune / Giulio Tononi and Chiara Chiarli

Every night, when we are asleep, blind, mute and almost paralyzed, our brains are hard at work. Neurons in the sleeping brain fire almost as often as during waking hours, and they consume almost the same amount of energy. What is the purpose of this incessant activity when we are supposed to rest? Why does the mind disconnect so completely from the external environment while the brain continues to chatter?

Brain activity at rest probably plays some vital role. And apparently, all animals sleep, although this state of unconsciousness and inability to react greatly increases the chance of becoming someone else's lunch. Old birds, old bees, old iguanas and old cockroaches, even old fruit flies, as we and others proved more than ten years ago.

More than that, during evolution some wonderful adaptations have developed that allow sleep even in special situations. For example, dolphins and some other marine animals that must come to the surface of the sea to breathe alternately turn off one half of the brain to sleep, while the other half remains awake.

Like many scientists and many other non-scientists, we have long wondered what the benefits are that make sleep so essential to living things. More than 20 years ago, when we were working together at the St. Anna School of Advanced Studies in Pisa, Italy, we began to suspect that brain activity during naps somehow resets the billions of neural connections that change daily due to the events that occur while awake. According to this approach, sleep preserves the ability of the neural circuits in the brain to create new memories during the individual's lifetime, without reaching oversaturation and without erasing old memories.

We also have a hypothesis as to why awareness of the external environment should be turned off during sleep. It seems to us that the mind must stop the conscious experience of the here and now in order to have a chance to merge old and new memories. Sleep provides this opportunity.

Our theory arouses some controversy among our neurobiological colleagues who study the role of sleep in learning and memory, because we claim that the resetting of neural connections results from the weakening of the connections between nerve cells that fire during sleep. According to the popular opinion, however, brain activity during sleep Reinforces the neural connections related to storing new memories. However, years of research in various creatures, starting with flies and ending with humans, support our hypothesis.

Memorizing in sleep

Already about a century ago, scientists hypothesized that sleep is important for memory function, and many experiments conducted since then have shown that after a night of rest, and sometimes even after a light nap, new memories are retained better compared to being awake for the same period of time. This phenomenon is true both for declarative memories, such as lists of words and the connections between places and images, and for procedural (procedural) memories, which are the basis of sensory and movement skills, such as playing a musical instrument.

The evidence that sleep improves memory has scientists looking for signs that the brain is processing new material learned during the day at night. And indeed they found them: studies conducted over the past 20 years, first in rodents and then in humans, show that patterns of neural activity during sleep sometimes resemble patterns of activity recorded while the subjects were awake. For example, when a rat learns to navigate a maze, nerve cells (neurons) in an area of ​​the brain called the hippocampus fire in certain patterns. During the subsequent sleep, rats "repeat" these patterns with a higher frequency than would be expected at random.

Because of such findings, many researchers assumed that the "repetitive gear" during sleep fixes the memories through an additional strengthening of the synapses - those points of contact between nerve cells - that began to strengthen during wakefulness. The idea is that if interconnected neurons fire repeatedly, the synapses connecting them will transmit signals from one neuron to another more easily, helping the neural circuits to encode memories. This process of selective strengthening is called synaptic potentiation (strengthening), and it is commonly thought that this is the mechanism underlying learning and memory in the brain.

However, while the processes of repetition and reinforcement occur while awake, scientists have not yet found direct evidence that the synapses participating in the repetition circuits are indeed strengthened during sleep. We were not surprised by the lack of evidence, because it is consistent with our explanation that while we are asleep and unconscious, all that activity of the brain, the "repetitive gear" and apparently random activity of nerve cells, basically Weakening neural connections, and does not strengthen them.

The price of plasticity

There are many good reasons to think that in order for the brain to function properly, synapses must be weakened, not just strengthened. First, strong synapses consume more energy than weak synapses, and the brain does not have inexhaustible stores of energy. In humans, the brain is responsible for almost 20% of the body's energy budget, more than any other organ relative to its weight. And at least two-thirds of the energy consumption supports synaptic activity. Building and strengthening synapses is also a major source of cellular stress, because they require the cells to synthesize and send to the synapses various components, from mitochondria (the power plants of the cells) and synaptic vesicles (which carry signaling molecules), to proteins and lipids necessary for synaptic communication.

It therefore seems clear to us that such exploitation of resources is not sustainable. The brain cannot continue to increase synapses day and night during the entire life of the individual. We have no doubt that learning occurs primarily through synaptic strengthening. We simply doubt that the strengthening of synapses continues to occur during sleep.

On the other hand, weakening synapses during sleep will return the various brain circuits to the basic level of strength and prevent excess energy consumption and cellular stiffness. We consider this process of sleep to be the preservation of synaptic homeostasis, and call the overall theory of the role of sleep the "synaptic homeostasis theory" SHY)). Basically, the SHY theory explains the essential and general need for sleep for all creatures that sleep: sleep returns the brain to a state where it can learn and adapt when we are awake. The risk we take during prolonged disconnection from the environment is the price we pay for the plasticity of the brain - its ability to change the wiring in response to external events.

But how does the SHY theory explain the beneficial effect of sleep on learning and memory? How can weakened synapses improve the general ability to acquire skills and facts? Consider that during a normal day, almost everything you experience leaves neural traces in the brain and that significant events, such as meeting someone new or learning to play a piece on the guitar, are only a tiny part of the neural coding. To improve memory, the sleeping brain needs to distinguish, in some way, between the "noise" of irrelevant information and the "signal" of significant events.

We propose that during sleep, the spontaneous firing of neurons in the brain activates many neural circuits in different combinations, including both new memory records and old networks of already learned connections. (Dreams are a glimpse of this free neural competition.) The spontaneous activity allows the brain to check which new memories match better with significant past memories and to weaken the synapses that do not match the general memory plan. We and other researchers are investigating possible mechanisms by which brain activity can selectively weaken synapses encoding "noise" while preserving synapses corresponding to the "signal."

And while the brain examines imaginary scenarios and weakens connections in the right places, it is better not to be aware of the environment and not to act within it. In other words: we better sleep. Similarly, the restoration of synaptic homeostasis should not occur when we are awake because the events of the day will take over the process and stand out at the expense of all the knowledge the brain has accumulated over the course of life. The deep disconnection during sleep frees the brain from the tyranny of the present, and creates optimal conditions for integrating memories and consolidating them.

weak link

Our theory that the brain uses neural firing during sleep to weaken synapses instead of strengthening them is partially supported by an in-depth analysis of data gathered from a common workhorse of sleep research: the electrical electroencephalogram (EEG). An EAG device records patterns of electrical activity in the cerebral cortex using electrodes attached to the scalp. A few decades ago, EAG recordings of the old brain identified two main categories of sleep: rapid eye movement (REM) sleep and non-REM (NREM) sleep, which alternate throughout the night. Each category has a unique pattern of brain waves. More on jitters The eyeballs under the huge eyelids, which give REM sleep its name, this stage is characterized by relatively fast brain waves: rapid ups and downs in the chart The EAG, which are similar to the EAG charts in the awake state. Conversely, slow oscillations, with a frequency of about one cycle per second, are the most prominent characteristic of NREM sleep.

Ten years ago, the late Mircea Steriade of the University of Lowell in Quebec discovered that the slow oscillations of NREM sleep are created when a group of neurons fire together for a short period of time (ON), become silent for a fraction of a second (OFF) and then fire together again. This was one of the most fundamental discoveries in sleep research. Since then, scientists have discovered that even in birds and mammals the intensity of the slow waves is greater if they were preceded by a long period of wakefulness, and their intensity decreases as sleep continues.

We hypothesized that if the synapses are strong, neurons will be able to coordinate their firing to a greater extent, and the result will be slow waves of great intensity. If the synapses are weak, there will be less coordination between the nerve cells and the intensity of the slow waves that will be received will be lower. From the results of computer simulations and experiments on humans and animals, we concluded that the powerful and steep slow waves at the beginning of the night indicate that synapses were strengthened during the waking period that preceded it, while the weak and shallow waves of activity measured early in the morning indicate that the synapses weakened during sleep.

Direct support that synapses weaken during sleep, and perhaps even thin, comes from animal studies. In flies, for example, we see that sleep reverses the increase in the number of synapses and their size, which occurs during the day, especially when the flies are exposed to environmental stimuli. Synaptic spines are unique protrusions on the extensions responsible for detecting signals in nerve cells. As fruit flies communicate during the day with other flies, by evening new synaptic spines are sprouting on nerve cells throughout the brain.

Equally amazingly, the number of bites returns to baseline by the next morning, but this only happens if the flies have been allowed to sleep. We saw a similar phenomenon in the cerebral cortex of adolescent mice: the number of synaptic spikes tended to increase when the animals were awake, and decrease when they were asleep. A similar trend prevails in adult rodents, although it is not the number of stings that changes during wakefulness or sleep, but the frequency of certain molecules, called AMPA receptors, which are found in the stings and determine the strength of the synapse. When we tested AMPA receptors, we found that the number of receptors per synapse increases during wakefulness and decreases after sleep. More receptors make the synapse stronger, while fewer receptors mean the synapse has weakened.

It is also possible to assess the strength of the synapse directly through electrical stimulation of the nerve fibers in the cerebral cortex. The neurons react with a stronger electric current when the synapses are stronger and weaker when the synaptic connections are weak. We have shown that in rats, neurons that have been stimulated fire with greater intensity after a few hours awake, and with a weaker intensity after sleep. Marcello Mascimini from the University of Milan in Italy, and Verto Huber, now at the University of Zurich, performed a similar experiment in humans. Instead of electrical stimulation with an electrode, they used magnetic brain stimulation, a short magnetic pulse through the skull to stimulate the nerves underneath. They then recorded the intensity of the response in the cerebral cortex using an EAG device with a dense array of electrodes. The results were unequivocal: the longer the subjects were awake, the more intense the EAG response. A night of sleep is required for the cortical response to return to baseline.

Less is more

The common conclusion of all these experiments that we conducted over the course of twenty years is that the spontaneous activity in the cerebral cortex during sleep does indeed weaken the synaptic connections in neural circuits, either by reducing their ability to send electrical signals or by completely erasing them.

This process that we call "reduction selection" is designed to ensure the survival of the neural circuits with the highest "adaptation", either because they were activated with the greatest intensity and most consistently during waking hours (for example, playing the right notes on the guitar while practicing a new composition) or because they fit together well more with existing memories (eg in the case of a new word in a familiar language). At the same time, synapses in circuits that were only moderately strengthened during wakefulness (such as incorrect notes on a guitar) or that fit less well with old memories (such as a new word in an unfamiliar language) will be suppressed.

Choosing to reduce can ensure that meaningless events do not leave their mark on our neural circuits, while meaningful memories are preserved. As a bonus, choosing to reduce will also make room for another cycle of synapse strengthening during wakefulness. Indeed, there are findings that indicate that besides the many other benefits of sleep for learning and memory, sleep helps in acquiring new memories (ie information acquired before the next cycle of sleep). Quite a few studies have shown that after a night's sleep, you can learn new material much better than at the end of a day of being awake. (Students, take note.)

Although we do not yet have direct evidence for a mechanism that can cause selective weakening of activated synapses, we do have an idea of ​​how synaptic weakening can occur. We believe that the slow waves of NREM sleep in mammals play some role. In laboratory studies of rat brain tissue, nerve cells became less efficient at transmitting signals to each other when they were stimulated in a way that mimicked the synchronous oscillations of the slow waves of sleep.

The chemistry of the brain also changes during NREM sleep in a way that can lead to a weakening of synapses. In an awake person, the brain is flooded with a concentrated soup of chemical signals, or neuromodulators, including acetylcholine, norepinephrine, dopamine, serotonin, histamine and hypocretin. In the presence of these substances, synapses tend to strengthen when signals pass through them. During sleep, especially in NREM sleep, the concentration of chemical signals in the soup drops considerably. This administration of neuromodulators around the synapses can cause them to weaken when signals pass through them, instead of strengthening. Also, a growth factor called BDNF, which strengthens synapses and is involved in the accumulation of memories, may also be involved in the process. BDNF levels in neurons are high during wakefulness and minimal during sleep.

Local sleep

Regardless of the exact mechanisms of the selection process, there is strong evidence in several biological species that the overall strength of synapses increases during wakefulness and decreases during sleep, as predicted by the SHY theory. We can add to and test the theory by testing other interesting difficulties arising from it.

For example, if the theory is correct, then the more adaptive processes an area of ​​the brain undergoes, the more sleep it will need. It is possible to estimate the "need for sleep" according to the intensity of slow NREM waves and their duration. To test this trick, we asked volunteers to learn a new task: how to reach a target on a computer screen, with the cursor (controlled by a mouse) rotating in a systematic pattern. The part of the brain involved in this type of learning is the parietal cortex. Indeed, when our subjects slept, the power of the slow waves in their right parietal cortex was greater relative to the power of the waves in the same region the night before learning. The strength of these waves moderated during the night, as such fluctuations do. But the great intensity of the waves we saw at the beginning of the night shows that a certain area of ​​the brain was working hard due to the task we gave it.

Many experiments conducted by both of us and other researchers have confirmed that learning, and more generally the activation of synapses in neural circuits, produces a local need for sleep. Recently we have even discovered that prolonged or intense use of certain neural circuits may cause local groups of nerve cells to "fall asleep" even when the rest of the brain (and the creature itself) is awake. Thus, if a rat stays awake for a longer time than usual, some neurons in the cerebral cortex become silent for short periods that look identical to the OFF periods during slow waves of sleep. At the same time, the rat runs around with its eyes open, going about its business, like any other awake rat.

This phenomenon is called local sleep, and it is attracting the attention of other researchers. Our recent studies indicate that local OFF periods also occur in the brains of sleep-deprived humans and that these periods are more frequent after strenuous learning. It seems that when we're awake too long or overusing certain neural circuits, little bits of the brain take short naps without us knowing. One has to wonder how many errors of judgment, stupid mistakes, nervous reactions and bad moods are due to local sleep in the minds of exhausted people who believe they are wide awake and in complete control.

It also follows from the SHY theory that sleep is especially important in childhood and adolescence. These are times of concentrated learning and intense changes in the structure of synapses, as many studies have shown. During the teenage years, synapses are created, strengthened and cut at an extremely rapid rate that is unmatched during adulthood. If so, it is likely that a reduced selection during sleep will be essential to minimize the energetic cost of the accelerated synaptic remodeling and to clarify the desired neural circuits during these stages of life. One can only wonder what happens when sleep is disturbed or lacking in critical stages of development. Does the lack of sleep impair the normal refinement of the neural circuits? In such a case, the lack of sleep will cause not only random forgetfulness or errors in judgment but long-term changes in the way the brain is wired.

We intend to continue testing the predictions of the SHY theory and examine its implications. For example, we hope to find out if sleep deprivation during neural development actually leads to changes in brain wiring. We would also like to learn more about the effect of sleep on deep areas of the brain, such as the thalamus, cerebellum, hypothalamus, and brainstem, and about the role of REM sleep in synaptic homeostasis. We may be able to find out if sleep is indeed the price for brain plasticity, a price that every brain and every nerve cell must pay.

__________________________________________________________________________________________________________________________________________

About the authors

Giulio Tononi and Chiara Cirelli are professors of psychiatry at the University of Wisconsin-Madison. Their research on the roles of sleep is part of a large-scale research work on the human mind, which is also the subject of Tononi's latest book, Pi: A Journey from the Brain to the Soul (Pantheon, 2012).

in brief

Sleep must play some vital role, because all animals sleep.

Evidence suggests that sleep weakens connections between nerve cells. This is a surprising finding because strengthening these connections during wakefulness supports learning and memory.

But by weakening synapses, sleep may protect brain cells from saturation due to daily experiences and from excess consumption of energy.

And more on the subject

Is Sleep Essential? Chiara Cirelli and Giulio Tononi in PLOS Biology, Vol. 6, no. 8, pages 1605-1611; August 2008.

The Memory Function of Sleep. Susanne Diekelmann and Jan Born in Nature Reviews Neuroscience, Vol. 11, no. 2, pages 114-126; February 2010.

Local Sleep in Awake Rats. Vladyslav V. Vyazovskiy, Umberto Olcese, Erin C. Hanlon, Yuval Nir, Chiara Cirelli and Giulio Tononi in Nature, Vol. 472, pages 443-447; April 28, 2011.

Sleep and Synaptic Homeostasis: Structural Evidence in Drosophila. Daniel Bushey, Giulio Tononi and Chiara Cirelli in Science, Vol. 332, pages 1576-1581; June 24, 2011.

The article was published with the permission of Scientific American Israel