WEDNESDAY, 10 DECEMBER 2025
Why do we sleep? How is it that spending a third of our lives unconscious, vulnerable to predators and other threats, has survived the test of evolution? It seems so much more practical to rest while conscious, alert to our surroundings, and ready to react. Yet, despite this apparent disadvantage, sleep is a universal necessity for nearly every living creature. So, what’s really going on during those hours of inactivity?To understand this, we need to look at the brain’s fundamental building blocks: these are neurons, or nerve cells, designed to send and receive signals that power everything from emotions to movement and memory. A neuron would be useless in isolation; to send and receive these signals it must be able to communicate with other neurons. In the human brain, 100 trillion connections form between our ~86 billion neurons. These connections are called synapses, which are structures that form between the output end of one neuron and the input end of another, enabling transfer of information between them.
Neurons act like train tracks, forming long branching structures that carry signals across the brain. The signals that travel down neurons are like trains. When they reach a station (a synapse), passengers disembark (like neurotransmitters), cross the platform, and board a new train that will travel down the second neuron, propagating the signal. An important feature of synapses is that they are plastic — they can change and adapt. If two trains connect frequently, the passengers get used to the transfer and become faster at it over time. Similarly, when signals frequently pass between two neurons the synapse is strengthened allowing for more efficient transfer of information. This process plays a critical role in memory formation.
Imagine you’re learning that “hola” means “hello” in Spanish. The more you practise this association, the more the synapses connecting these ideas in your brain are used, strengthening the connections and making this memory easier to retrieve in the future. This goes both ways — a synapse that is used infrequently will weaken. Imagine dumping passengers at an old, derelict station that hasn’t been used in years — they’d likely get lost, slow down, or even miss their connection entirely. Trying to remember a fact someone told you only once, a long time ago, without ever having thought about it again, requires more effort and might even be impossible.
However, this presents us with a problem: if you have an interesting day and learn a lot, your brain will undergo net synaptic strengthening. If this keeps happening, the brain’s synapses should just keep getting stronger and stronger — it would be a one-way process leading to a hyper-connected brain. Importantly, strong synapses use more of the brain’s energy, so the brain would eventually become overloaded with energy-draining synapses and start to lose its capacity to process new information.
This is where sleep, and the synaptic homeostasis hypothesis, comes in. This hypothesis is considered one of the most widely accepted theories for the function of sleep and builds on the idea of synaptic plasticity, proposing that sleep acts as a balancing force to counteract any synaptic strengthening that occurs during the day. The theory suggests that during specific sleep stages, synapses in our brains are selectively weakened, returning the brain to a stable state and making space for a new day of learning. So, while we are awake, net synaptic strengthening occurs, but during sleep, they get weakened — this creates a balance that maintains synaptic homeostasis (stable conditions).
The 2003 paper “Sleep and Synaptic Homeostasis: A Hypothesis” by Giulio Tononi and Chiara Cirelli first introduced this idea, and since then, numerous research groups have continued to formally test and expand upon the hypothesis. One of the key pieces of evidence supporting the original hypothesis was a positive correlation found in mammalian studies between the time spent in slow-wave sleep (SWS) and the duration of prior wakefulness. A full night’s sleep consists of four stages: three stages of NREM (non-rapid eye movement) and one stage of REM (rapid eye movement) sleep. SWS, often referred to as deep sleep, is the third stage of NREM. Tononi suggested that SWS is when synaptic weakening occurs. This makes sense: the longer you spend awake strengthening your synapses, the longer it will take to weaken them during sleep, which explains why the length of SWS increases with the duration of prior wakefulness.
Since the publication of this hypothesis, numerous studies have uncovered supporting evidence, with a 2017 study by Vivo et al. providing particularly compelling results. This study employed a technique called volume electron microscopy (vEM), which allows for a 3D reconstruction of neurons. The process works by taking a block of tissue, slicing it into thin sections, imaging each section with an electron microscope, and then stacking them to build a 3D image.
The researchers used mice in their experiment, allowing one group to sleep normally while keeping a second group awake and actively interacting with their environment. By applying vEM, they examined how the neurons of these two groups had changed. What they found was striking: synapses physically shrank in the mice that were allowed to sleep, but remained unchanged in those that stayed awake. Since synapse size can be an indicator of strength — stronger synapses tend to be larger, while weaker ones are smaller — this finding provided strong evidence for the synaptic homeostasis hypothesis.
Clearly, when we wake up, we haven’t forgotten everything from the day before. A key insight this study provides how the brain decides which memories are worth preserving. It’s actually a simple answer and comes down to amplitude. A consistent weakening force is applied to all synapses, but the ones that were strengthened a lot – those tied to significant memories, like breaking your arm — remain above the threshold and are preserved. Meanwhile, insignificant things like what you had for breakfast, drop below the threshold and are forgotten.
An aspect of this theory that remained uncertain until recently was whether sleep provides the driving force for synaptic weakening or merely functions as a permissive state. In other words, just because widespread synaptic weakening happens during sleep, doesn’t necessarily mean that sleep actively causes it. It could simply be a state in which the process can occur. A 2024 study on zebrafish by Suppermpool Et al. suggests that sleep does, in fact, just provide conditions that enable synaptic weakening, while the true driving force is ‘sleep pressure’ – the body’s need for sleep. The researchers found that zebrafish who were more sleep deprived prior to falling asleep experienced more synaptic weakening during sleep, linking sleep pressure with this process. Conversely, zebrafish who were not sleep-deprived and were given drugs to induce sleep, experienced little synaptic weakening.
So, this might be the beginning of an explanation for why we’ve evolved to sleep: it gives us time to clear out a daily junk drawer of uninteresting memories, leaving space for tomorrow’s more interesting ones.
Pixar’s 2015 animated movie Inside Out captures the essence of this process, with animated characters inside an 11-year-old’s brain sorting her memories into ‘long-term storage’ and the ‘Memory Dump’ while she sleeps. One thing we might want to take from this, as implied by Suppermpool Et al.’s findings, is that to unlock this brain boost, we need to have enough sleep pressure. So daytime napping might be significantly less useful than a full night’s sleep.
Rebekah is a third-year biological natural sciences student at Emmanuel college with a specialisation in Neuroscience.
(Artwork by Ella McGovern)
