The latest Science confirms: Sleep can "reboot" the brain's learning functions.
All animals need sleep, yet the importance of sleep is often overlooked by modern people.
Some view sleep as a waste of time and try to "save" sleeping hours to do what they want, believing this is a proactive way to make the most of their limited lives. However, while the body may be "lying flat" during sleep, the brain certainly isn’t.
If the brain is compared to a computer, going to sleep is not equivalent to "shutting down" or "hibernating," but more like "rebooting" and "updating."
Numerous studies have shown that the brain's neural networks undergo active changes during sleep—changes that occur only during this state.
These changes enable better cognitive function and learning ability after sleep. This explains why many people experience the phenomenon where they can’t remember a new movement or word after hours of practice, but after a good night's sleep, they suddenly can.
A study published this week in Science provides new evidence for the resetting function of sleep. Scientists at Cornell University discovered that during sleep, a unique neural activity pattern in the hippocampus—a critical brain region for learning and memory—plays an indispensable role in memory consolidation.
Scientists had previously discovered that memory consolidation occurs during sleep, especially in the non-rapid eye movement (NREM) phase. During NREM sleep, we typically experience progressively deeper sleep and become harder to wake. However, if we observe brainwave activity, the hippocampus exhibits a distinctive firing pattern in this phase, known as sharp-wave ripples (SWRs). This involves a group of neurons firing synchronously, followed closely by another group doing the same, resembling ripples spreading in water.
Hippocampal SWRs help the brain replay and reinforce recently learned information or experiences, stabilizing them into long-term memories stored in the cortex. Experiments show that disrupting SWRs during sleep impairs memory in animals, while artificially enhancing SWRs improves it. This demonstrates the crucial role of hippocampal SWRs in memory consolidation.
In a recent Science paper, researchers implanted electrodes in the hippocampus of mice to record neuronal activity patterns during the learning of new tasks and subsequent sleep. Beyond SWRs, the scientists identified a new cluster-firing pattern in NREM sleep that balances SWRs.
Interestingly, this pattern involved CA2 pyramidal neurons, a group that had previously received little attention.
The hippocampus is divided into CA1, CA2, and CA3 regions. Most past studies focused on CA1 and CA3, whose pyramidal neurons are known to encode memories. However, the researchers noticed that while the active CA1 and CA3 regions would sometimes suddenly quiet down, a group of CA2 neurons would emit a barrage of action potentials, leading them to term this firing pattern BARR (barrage of action potentials).
BARR lasts significantly longer than SWRs, averaging 300 milliseconds compared to SWRs' 50 milliseconds. Both patterns increase and occur frequently during post-learning sleep, then gradually decrease over time. Interestingly, the two patterns alternate consistently; when researchers used optogenetics to artificially increase SWRs in CA1, CA2-driven BARR also increased shortly thereafter.
By examining individual neurons, the researchers revealed a more complete picture of memory consolidation during sleep. In early NREM sleep, CA1 neurons involved in learning are reactivated via SWRs. As BARR emerges later in NREM, the activity of these neurons gradually returns to baseline levels. This BARR-induced "reset" allows CA1 neurons to be reactivated again, replaying the neural activity from the learning phase.
Although BARR seems to suppress SWRs, experiments revealed that this suppression is essential for memory consolidation. Disrupting BARR artificially did not eliminate SWRs, but it caused "over-synchronization" of learning-related neurons, extending their activity duration and leading to abnormal neural connections. Consequently, the animals' memory performance suffered.
The researchers concluded that reactivation of learning-related neurons must maintain a balanced state—neither too high nor too low—to consolidate memories effectively. The newly discovered BARR pattern finely tunes this balance, preventing excessive neural activity.
In summary, this new discovery reinforces the importance of a good night's sleep after learning. It allows the brain to reset and prepare for ongoing learning and the absorption of new information.