For years, scientists had the idea that during sleep, the connections between brain cells – synapses – kept firing like mad through the night.

Why? So the brain could replay the information acquired during waking hours.

This was thought to consolidate memory and learning. And it served as an explanation as to why we need sleep.

Sounds reasonable, right?

Underpinning this idea was the mistaken belief that consolidating memory took place during a brief window of a few hours.

Memory consolidation is a complex process that can take weeks, months, even years. Certainly, it takes longer than the seven to nine hours we spend sleeping.

So what’s going on ?

In 2008, researchers from the University of Wisconsin School of Medicine and Public Health came up with a new theory. Sleep served to reset the brain.

When we’re awake, the synapses are firing with countless messages, stimulated by our thoughts and conversations, the work we do, the effect of the environment and just staying upright.

When we become tired, the brain is saying: Hey, I’ve had enough, I’m struggling to keep up.

As study author Chiara Cirelli, associate professor of psychiatry, explained at the time: “The human brain expends up to 80 per cent of its energy on synaptic activity, constantly adding and strengthening connections in response to all kinds of stimulation.”

There are millions of neurons (cells) in the human brain. Each neuron contains thousands of synapses.

Not surprisingly, this energy expenditure “is huge and can’t be sustained”.

Therefore, instead of the synapses firing intensely as we sleep, their activity is weaker. This allows the brain to prepare for a new day.

“We need an off-line period, when we are not exposed to the environment, to take synapses down,” Cirelli said.

Experiments with rodents supported the idea: When the rats were awake, synaptic activity was strong. When they slept it was weak.

Cirella and company called their theory ‘the synaptic homeostasis hypothesis’.

See-through fish support the theory

A clever new study, using zebrafish, has refined the theory.

Researchers from the University College London (UCL) found that during sleep, the brain weakens the newest connections between neurons that had been forged while awake.

But this only occurs during the first half of a night’s sleep.

According to a statement from UCL, the scientists used optically translucent zebrafish. In other words, especially engineered see-through fish.

The fish were doctored with genes that enabled synapses (those brain connections) to be easily seen and imaged.

The research team monitored the fish over several sleep-wake cycles.

The researchers, as per the theory, found that brain cells “gain more connections during waking hours, and then lose them during sleep”.

The researchers also found that “this was dependent on how much sleep pressure (need for sleep) the animal had built up before being allowed to rest”.

When the fish were deprived of sleep for a few extra hours, “the connections continued to increase until the animal was able to sleep”.

Lead author Professor Jason Rihel (UCL Cell and Developmental Biology) said: “If the patterns we observed hold true in humans, our findings suggest that this remodelling of synapses might be less effective during a midday nap, when sleep pressure is still low, rather than at night when we really need the sleep.”

The researchers also found that “these rearrangements of connections between neurons mostly happened in the first half of the animal’s nightly sleep”.

This, they say, “mirrors the pattern of slow-wave activity, which is part of the sleep cycle that is strongest at the beginning of the night”.

If the brain resets itself in three to four hours, what’s the point of sleeping twice as long as that? The researchers admit this remains an “open question”.

The mystery continues.