Ask a Caltech Expert: David Prober on the Brain and Sleep

What is sleep and how do scientists define it?

The electroencephalogram, or EEG, recordings that are used to monitor sleep and wake states in humans were first made about 100 years ago. But we've known for much longer than 100 years that humans and other animals sleep. That's because sleep is a behavior, so it can be defined by behavioral criteria: In general, a sleeping animal doesn't move very much, although there are some interesting exceptions to this rule. Second, an animal that's sleeping can be distinguished from one that's sitting quietly but awake because there's an increased arousal threshold during sleep, meaning it takes a stronger stimulus to arouse the sleeping animal. The third key criterion for sleep is that if you deprive someone of sleep, they'll experience rebound sleep to make up for the sleep that was lost.

What is happening in our brains when we sleep?

I think it's fair to say that this is one of the big remaining mysteries in biology. When EEG recordings are performed in mammals, they only measure the activity of neurons in the dorsal [upper] part of the brain, so you're basically seeing what's happening in the cortex [the wrinkled, outermost portion of the brain] with very low resolution. It's unclear what's happening in other regions of the brain. In general, studies using animal models show there's overall less activity in the brain during sleep, but there is still plenty of activity in the brain, including some parts of the brain that are specifically active during sleep and some patterns of brain activity that are specific to sleep.

Is there a specific region of the brain that is involved in controlling sleep?

It's thought that a lot of the neurons and neuropeptides that regulate sleep are in the hypothalamus and brain stem. These are parts of the brain that are ancient and well conserved in vertebrates but are not as well studied in mammals because they are located deep within the brain, and, as a result, they are challenging to access for experiments. We have easier access to these brain regions using the animal model I study, the zebrafish, because zebrafish are transparent at the stage when we study them.

What is dreaming, and are humans the only creatures that dream?

We still don't know what dreams are. In mammals, sleep is divided into two categories: rapid eye movement (REM) sleep and non-REM (NREM) sleep. One possibility is that during dreams, your brain replays events that happened while you were awake. Another idea is that different parts of the brain are active during REM and NREM sleep, and specific patterns of activity during these states may result in what we experience as dreams. In this context, dreams could be the result of your brain trying to interpret confusing patterns of activity in a manner that makes sense to you. This could explain why dreams have some basis in reality. They may refer to something that happened to you or something you've been thinking about, but they're often also strange and do not perfectly reflect what happened in reality. But that's pure speculation.

The problem with answering this question is you need to be able to communicate with an animal to know if they're dreaming. When you watch your dog sleep, often it looks like it's having a dream because it moves its limbs and makes sounds, so we assume it's dreaming, but we don't actually know if that's the case. In humans, there's lots of ideas for what might be involved in dreaming, but there's no conclusive scientific evidence for any of these ideas.

What are some of the major open questions about the neuroscience of sleep?

There are two major open questions. One is: How is sleep regulated? The second is: What's the function of sleep?

My lab focuses on the first question, how sleep is regulated. Several genetic mechanisms and neural circuits have been identified that either promote wakefulness or promote sleep. There are some examples of mechanisms that are conserved from nematodes all the way to humans. The hope is that by studying animals that have relatively simple brains, like zebrafish, where we can image every neuron in the brain when the fish is awake or asleep, we can answer some basic questions. For example, is there a single key neural circuit that regulates both sleep and wakefulness, or are there many different pathways that act in parallel? If there are many pathways, how do they talk to each other to ensure that an animal is either asleep or awake?

In terms of the function of sleep, there are several ideas with different levels of experimental support for different hypotheses.

One popular idea is known as the synaptic homeostasis hypothesis. The idea is that during the day when you're awake, you're always taking in information. As a result, you build up synapses (connections between neurons, or brain cells) within your brain while you are awake. If this were to go on indefinitely, eventually you'd run out of space in your brain, and you wouldn't be able to learn anymore. To prevent this from happening, when you sleep, you eliminate many of the synapses that were built when you were awake. This hypothesis assumes that the most important things that you learned while awake form strong synapses that are not eliminated during sleep, while synapses that formed in response to relatively unimportant experiences are weaker and are eliminated during sleep.

There are a few papers, most using mice, showing that dendritic spines (parts of neurons that receive signals from other neurons) tend to be larger after periods of wakefulness than they are after periods of sleep. There's also evidence that some of the proteins that are known to be involved in building up or breaking down synapses correlate with wakefulness or sleep.

Another idea is that toxins build up in the brain when you're awake, and these toxins get flushed out of the brain during sleep. There are a few papers using rodents and fruit flies suggesting that this happens. One molecule that may be particularly important in this process is amyloid beta, a peptide fragment that's implicated in causing Alzheimer's disease. It's been shown that levels of amyloid beta build up in the brain when you're awake, and then decrease during sleep. So, one function of sleep could be to get rid of this toxic fragment that builds up during wakefulness.

There's another hypothesis that was proposed by Alex Varshavsky at Caltech about 10 years ago that he calls the fragment generation hypothesis. Dr. Varshavsky studies mechanisms that regulate protein cleavage. In order to function properly, many proteins are cleaved by specific enzymes resulting in two protein fragments. Very often, one of the fragments performs a useful function, while the second fragment is essentially a piece of junk that has no function. If this junk keeps building up, eventually it causes problems with the biochemistry of your cells, so these fragments need to be eliminated. Dr. Varshavsky's idea is that these protein fragments build up during wakefulness and are selectively eliminated during sleep.

Why haven't we evolved to need less sleep or none at all?

There are survival costs to sleep. For example, you're more susceptible to predation during sleep, so why do we still sleep even though it has this cost? The synaptic homeostasis hypothesis for why we sleep implies that sleep is an essential feature of neurons, and because all animals have neurons, all animals need to sleep. Similarly, if the fragment generation hypothesis is correct, it's not possible to evolve away the need to sleep because the generation of protein fragments is an essential part of the biochemistry of all cells.

Are there diseases that are known to be associated with a lack of sleep or interrupted sleep?

The most famous but very rare disease associated with sleep loss is called fatal familial insomnia. This is a neurodegenerative disorder where the first symptom is insomnia, but there's lots of degeneration throughout the brain. It's lethal, usually a few years after insomnia onset. More importantly, over the last five to 10 years it has become clear that sleep defects are present in many, perhaps most, neurological disorders, and, in some cases, there is evidence that disrupted sleep may not just be a consequence of the disorder but rather may contribute to the disorder. One example is Alzheimer's disease, where sleep defects often precede memory defects. Most people with autism have sleep defects, and we've used zebrafish to study the functions of genes that increase the risk of autism. When we mutate these genes in zebrafish, we often see striking sleep defects. Determining the extent to which disordered sleep contributes to these disorders, and whether treatments to improve sleep might help to slow or prevent neurological disorders, are active areas of research.

Are there clinically studied ways to help treat insomnia and sleep defects?

Back in 2015, we showed that melatonin is required for circadian regulation of sleep, which had been an open question for a long time. There are studies in humans showing that taking melatonin can shorten the time to fall asleep and also increase the amount of sleep, but the effect is relatively subtle compared to sleep aids that require a prescription, such as Ambien. However, while prescribed sleep drugs have very strong sleep-promoting effects, most don't induce a natural type of sleep. Most of these drugs act by activating GABA receptors, which essentially turns the brain off but also results in all kinds of undesirable consequences. And it's become increasingly clear that taking these drugs for a long time causes problems. A new class of drugs that inhibit a specific neuropeptide signaling pathway became available a few years ago. The neuropeptide is called orexin or hypocretin. This peptide promotes wakefulness, and it's been shown that drugs that inhibit its receptors help people to fall asleep. So far, there seems to be fewer unintended consequences than more traditional sleep drugs such as Ambien.

You can submit your own questions to the Caltech Science Exchange.