Blog

K-complexes are patterns of activity in which the voltage across the scalp rises then drops. K-complexes seem to kick off a period of reduced firing between neurons, perhaps because of the massive changes in the neuron’s polarity.

To explain why those large large fluctuations lead to a period of quiet, we need to dig into the way that signals propagate through the brain. EEGs measure the combined effect of action potentials in the brain. These action potentials are the electrical signals that travel from the cell nucleus, down the axons. However, the electrical potential is not what is transmitted between neurons. When an axon terminal reaches a sufficient voltage (positive or negative with respect to its surroundings) this opens ‘gates’ in those axons. What is released are neurotransmitters: chemical signals that bind with the receptor sites on other neurons. Those neurotransmitters cause a voltage change in the receiving neuron. If enough neurotransmitters from enough axons reach the dendrites of a given cell, gates open in the neuron, allowing either positively or negatively charged ions to enter the cell. If enough channels open to let in enough positive (or negative) ions then an action potential builds up again. For a K-complex to occur, there must be a large, coordinated flow of neurotransmitters that allow positive ions into the neurons, followed by a large, coordinated flow of neurotransmitters that allow negative ions to accumulate in the neurons. After that, the reserves of neurotransmitters are pretty much exhausted, which is why the rates of neuronal firing fall throughout the brain just after the K-complex. 

It would seem from this that K-complexes are designed to keep us asleep, to exhaust the neurotransmitters rapidly and leave our brains in what is called a ‘down state.’(Cash et al. 2009) However, recent research suggests that there may be more to it than that. K-complexes may serve as a kind of sentry, telling us whether to stay asleep, or to wake up. 

Combined EEG and fMRI recordings show that when K-complexes are picked up at the scalp, the brainstem, thalamus, and sensory areas of the brain are all doing…something.  Thirty-seven incredibly resilient volunteers managed to fall asleep inside a banking, clanking fMRI tube, and their activation levels during K-complexes were recorded and analysed by Professor Kolja Jahnke and colleagues at Goethe University in Frankfurt. For some time it had been thought that K-complexes were largely spontaneous, but Prof Jahnke’s team discovered that sensory processing areas of the brain are active during K-complexes. They particularly noticed that auditory processing areas of the cortex were active. As they pointed out: ‘acoustical scanner noise is unavoidable during [fMRI].’  In other words, between the pumps driving helium around the coils, to the rapid vibrations of the coils themselves, the inside of an fMRI machine is loud. It seems as though, during the up-state of K-complexes, we’re open to processing sensory information, although the parts of the brain involved in conscious awareness (the frontal-parietal area) stays fast asleep. 

That would explain why K-complexes can be useful ‘wake points’ if we hear a worrying noise in the night. But how do they keep us asleep? Activity along the midline of the brain in particular seemed to preserve sleep by triggering the downstate mentioned above.

K-complexes are then usually followed by sleep spindles. Sleep spindles are patterns of activity in the brain that are believed to help us go to sleep and stay there, although this might not be entirely true in very young children: babies whose brains produce more sleep spindles are no more difficult to wake up than those babies who produce fewer. (Horne et al. 2003) But in older children and adults, these spindles seem to be a necessary part of stage two NREM sleep. They might be part of the mechanism by which sleep helps us build memories and retain skills. People are more likely to remember more words, shapes and faces the next morning if they experienced more sleep spindles during the night. (Schabus et al. 2004; Clemens et al. 2005) Sleep can improve our physical skills too. People were asked to type the sequence 4-1-3-2-4 as quickly and accurately as possible with their non dominant hand. Those who napped between the practice and test phases did better than those who spend the intervening time awake. (Nishida and Walker 2007)

Babies and young children show a distinct pattern in their development of sleep spindles. Babies in their first year have rapid sleep spindles, which occur roughly once every 20 seconds, and they last for about 1-1.5 seconds. Children between the ages of four and 16 have slightly faster frequencies and similar durations with sleep spindles lasting around 1.5 seconds and occurring every five to six seconds. But between the ages of one and three, something unusual happens. Despite the density of learning that toddlers are undergoing, sleep spindles get further apart – once every 40 seconds at one year of age, to 110 seconds at 19 months, to every 83 seconds at two years and four months.  The durations of spindles drop too, roughly halving from the length that babies and older children experience. (Scholle et al. 2007) No one is quite sure what is happening during those toddler years, or what other mechanisms might take over to aid learning, but we can be certain that there’s something that is just different about sleep in the toddler years. The development of K-complexes, too, is rapid for the first two years, variable until five, but doesn’t reach mature rates until at least early adolescence. (Metcalf et al. 1971) Perhaps some of this wakefulness is not so much disordered sleep as developmental stages.

There are some extraordinary people who can get by on about three-quarters of the amount of sleep that the rest of us need. One particular family has a mutation that allows them to function normally on just six-and-a quarter hours of sleep per 24 hour period. This family was studied by geneticists, who found that a single letter change from C to T on a single gene, ADRB1, was present in all the family members who happily got by on less sleep, and was absent in all the other members of the family. (Shi et al. 2019) Only 4 in 100,000 (or 0.004 percent of the population) is thought to have this mutation, so the team genetically engineered mice so that they would carry the same mutation. Mice with the mutation were more active and slept less than the mice without.

It’s hard to know how the mice felt about this, but according to Professor Ying-Hui Fu of the UCSF Weill Institute for Neurosciences who led the study, the people with the mutation tend to be more optimistic, more energetic, better multitaskers, are more tolerant of pain, and don’t get jet lag.   According to professor Fu. ‘Natural short sleepers experience better sleep quality and sleep efficiency,’ she said. ‘By studying them, we hope to learn what makes for a good night’s sleep, so that all of us can be better sleepers leading happier, healthier lives.’ Forget trying to colonise Mars. This is the research that the billionaires should be funding.

How we fall asleep isn’t widely understood. Of the fifty or so peer-reviewed papers I have sitting on my desk, about 45 conclude that sleep is complicated. The other five conclude that it’s really complicated. That’s because we can only look at the activity of the human brain through snapshots of its metabolism or its electrical charge at the surface.

Genetics, too, is what we non-geneticists like to call ‘fucking loopy.’ For example, among those papers I mentioned, there are several that look at a gene that helps to determine whether or not it’s time to go to sleep. That gene – and the protein that it makes – is called TNFα. So far so acronym-y. But TNFα is short for tumor necrosis factor alpha. That’s what it was first spotted doing somewhere else in the body so the name stuck. But now we know that it also suppresses appetite, increases insulin resistance, and regulates how bitter things taste. But we’re stuck with the name; because history.  

But because I care for you, dear reader, more than I care for my sanity, I’ve autopsied this mangled body of research, and here’s what I’ve found. 

We now know that TNFα expresses a protein inside star-shaped cells in the brain called astrocytes. These cells support neurons both physically and metabolically – they’re a scaffold and a glycogen supply mechanism. But it was recently discovered that glial cells don’t just support neurons, they have signalling duties of their own. Astrocytes accumulate TNFα throughout the day, and then release it into the adjacent neurons, where it binds with receptors, which then contributes to the drive to go to sleep. (Vanderheyden et al. 2018)

But again, neither messing around with human genes, or surveying the entire population to find some very rare mutants, is going to get the science done. Results of previous research had shown that both sleep deprivation and high levels of neuronal activity makes humans and other animals produce more of the TNFα protein. It also showed that injecting TNFα into rabbits makes them go to sleep. So rather than tinkering with humans or searching for mutants, Professor William Vanderheyden and his team at Washington State University bred a population of fruit flies that lacked a very similar gene. 

‘Fruit flies happen to have a molecule that is very similar to TNFα  that is called Eiger, and the receptor to which it binds is called Wengen,’ says Professor Vanderheyden. ‘What we tried to identify through this research were mechanisms by which Eiger and Wengen could be regulating sleep in the fruit fly.’

When the team knocked out the flies’ ability to produce Eiger in their astrocytes, the flies became insomniac, sleeping less and sleeping irregularly. But when they injected the flies with the human TNFα protein, the flies went back to typical levels and patterns of sleep. So far so good:  TNFα sends a ‘sleep’ message. 

Next, they bred flies that lacked the Wengen receptor in their neurons. This time, injecting the flies with TNFα did nothing to fix their sleep. What’s more, the flies with no Wenger receptor proteins were incapable of recovering from sleep deprivation. You may be wondering how you deprive flies of sleep. In the old days you did it by depriving researchers of sleep. They would stay up all night tapping the flies’ jars, or gently moving them around. But this is one of the jobs that has fallen prey to automation: the Sleep Nullifying Apparatus (or SNAP) was invented at the machine shop at Washington University in Saint Louis. It tilts back and forth, moving the flies ten times a minute and preventing them from catching any shut eye for twelve hours or so.

So the next time you’re feeling sleepy, blame your astrocytes. But don;t be too quick to sign up for gene modification… In a future post I’ll explain why lack of sleep is super bad for you!