​​​​​Science Review of "Collector's Edition"｜Memory regulation of the sleeping brain

Written | Edited by Jiang Hong | Enzymatic sleep is the basic physiological activity of human and animal kingdoms
.

Our bodies can be static and immobile, but our brains can be very active
.

Philosophers divide the state of the brain into four stages: Awakening, Sleep, Slumber, and Supreme
.

How important is sleep? William Shakespeare, the famous Renaissance playwright and writer, has a classic and beautiful line about sleep: Why do we need to lie down and lose consciousness for several hours every day to rest? What exactly is the role of sleep? What happens if you don't sleep? In the past ten years, thanks to the advancement of technological means, there have been many breakthroughs in understanding sleep and solving sleep disorders and many other issues related to the sleep brain, including from the microscopic molecular mechanism to the macroscopic network level, making us more and more Get closer to understanding the truth about sleep
.

The importance of sleep for physical and mental health in modern society is often underestimated (996 Dangerous question of how it works?): studies in laboratory zoology also show that sleep deprivation leads to cognitive decline; in a US community In the large number of cases of sleep deprivation documented in the study, the average sleep duration was only 6.
1 hours, while experts recommend 7 to 9 hours for adults
.

Average sleep duration and sleep quality are also low among poor people in both developing and developed countries
.

This shows that sleep problems are not only related to physical health, but also related to economic development and social livelihood
.

At the end of 2021, the Science column will focus on sleep research, the latest review of sleep and learning and memory, sedation, and the triggers and outcomes of sleep written by experts in the field
.

This article summarizes a review on sleep and learning and memory of Brain neural patterns and the memory function of sleep, in order to give readers an intuitive understanding and review
.

Sleep is essential for good cognition, including memory
.

There are two main methods commonly used to record sleep: electrophysiological recordings with percutaneous and intracranial electrode implantation
.

According to its electrophysiological characteristics, sleep is divided into two phases: rapid eye movement (REM) and non-rapid eye movement (NREM)
.

The main features used to distinguish the two are: non-rapid eye movement sleep shows Sharp-wave ripples (hippocampal sharp wave ripples), Cortical slow oscillations (cortical slow-wave oscillations), delta waves and Spindles and other characteristics; REM sleep is mainly characterized by theta The wave is mainly oscillating
.

The electrical signatures of sleep with different characteristics accurately reflect the temporal activity of the underlying neural circuits
.

In this review, the author (Gabrielle Girardeau, independent research group leader at the French National Institutes of Health and Sorbonne University) mainly reviews how these electrical features guide our understanding and understanding of the role of sleep in the circuit mechanism of memory consolidation.
and mechanism of occurrence
.

Particular attention was paid to hippocampal theta-wave oscillations and Sharp-wave ripples and how they coordinate cortical patterns
.

Finally, the role of these EEG activity patterns in maintaining sleep homeostasis is emphasized, and future research issues and challenges for sleep involvement in memory-related learning consolidation are prospected
.

The process of memory formation is actually a challenging process for the brain to decide which new experiences can be stored and integrated into existing memories so that they can be sustained and corrected.

.

Memories are formed in the waking state as a series of new sensory experiences
.

After sleep provides the brain with a window of opportunity to sort and enhance newly encoded memories without the interference of external stimuli
.

This process, called memory consolidation, facilitates the creation of long-term memory or the formation of memory engrams (the process used to support the re-retrieval of information during waking periods)
.

During sleep, neural network pyramids involved in memory processing are endogenously activated
.

These activities can be recorded by atraumatic surface electrodes (EEG) or by intracranial electrodes to record local field potentials (LFPs) as well as spike potentials
.

Numerous studies have shown that these recorded electrical oscillation patterns during sleep are useful for understanding how the brain works.
These electrical signals include: oscillations (such as theta waves), transient potentials with recognizable waveforms (such as dendritic potentials), and Type of spike discharge activity (up and down states)
.

Several aspects of the role these electrical signals play in how the sleeping brain is involved in learning and memory are discussed below
.

Non-rapid eye movement sleep and hippocampal sharp wave ripples (SWR) The hippocampus has a three-level structure.
Information flows from the dentate gyrus (DG) through CA3 to CA1.
An important type in sleep is the sharp wave ripple complex (see Figure 1).
a Ripples)
.

During sleep, pyramidal neurons in CA3 fire spontaneously, firing synchronous bursts
.

This activity activates a large number of pyramidal neurons in the CA1 area
.

In the radiatum, the input received by the dendrites of CA3 produces spikes
.

Meanwhile, in the CA1 area, the interaction of activated pyramidal cells and interneurons produces fast oscillatory events (200Hz): ripples
.

There are currently two-step theoretical hypotheses about the involvement of these electrical signals in memory consolidation: first, subpopulations of CA3 and CA1 have already coordinated through theta oscillations during an experience and formed cell assemblies that encode relevant new information
.

These combined CA3 cells reactivate the CA1-integrated cell population through the SWR phase during the following sleep period, thereby strengthening the connection between CA1 and CA3, ultimately leading to memory consolidation
.

Several phenomena that are more consistent with this theory include that pairs of CA1 pyramidal cells that fire synchronously in the open-field explorations remain associated during sleep SWR phases
.

The persistence of activity associations during wakefulness is generally thought to be replayed in subsequent sleep; a similar phenomenon exists in co-discharge patterns and place cells are activated during wakefulness and during sleep, using a number of methods and means.
Reproduced in SWR
.

More importantly, a similar phenomenon exists in humans
.

Initially to study the link between recall and memory consolidation, researchers developed a closed-loop patterning system (Figure 2) that interfered with sleep ripples and found that spatial memory was significantly impaired
.

Optogenetic inhibition of pyramidal cells in the CA1 area during sleep SWR phase severely impairs the reappearance of these cell clusters following open-field exploration
.

These results suggest that the lack of spatial memory recall is due to the lack of consolidation of spatial maps or the formation of memory traces
.

There are many factors that affect SWR-related recurrence, and novelty stimulus is one of them
.

Figure 1 While most SWR and reactivation studies have focused on the CA1 region
.

However, social memory traces reactivated CA2 during the SWR phase, and CA2 reactivation had a bidirectional modulation of social memory enhancement and weakening
.

These results suggest that CA3-CA1 may be more inclined to the integration of SWR reactivation, which is helpful for the consolidation of spatial memory
.

CA2 may be more predisposed to the SWR phase in terms of social memory
.

Compared with the limited observation of local potential ripples, the development of algorithms in rapid online monitoring, especially in terms of reproduction content, will certainly be of great help to the understanding of sleep reproduction
.

Regarding the application of this, Gridchyn et al.
, trained rats to forage in two different environments and interfered with subsequent sleep and resting-state SWR events, but not in the first environment.
active part
.

It was found that the rats performed better in the uninfected environment than the second, suggesting that the consolidation of spatial memory in the first environment was not interrupted and was preserved
.

Taken together, results accumulated over the past few decades suggest that reactivation of hippocampal information integration is associated with fresh information
.

Learning during sleep SWR is critical for memory consolidation
.

Surprisingly, however, it was unknown whether the reactivation of the hippocampus was also present in the ventral part of the hippocampus
.

Brain connections in the ventral hippocampus differ from those in the dorsal hippocampus and are primarily associated with anxiety and stress
.

In addition, during the non-rapid eye movement period, spike discharges in the hippocampal dentate gyrus, reflecting strong cortical-to-dentate gyrus input, are thought to be involved in the consolidation process of NREM, but remain to be studied
.

Figure 2.
Hippocampus-cortical synergy enables most of the main theories about long-term memory consolidation through non-REM sleep patterns, including the interaction of the hippocampus and the neocortex
.

During NREM sleep, cortical circuits experience alternating phases of high and low cluster firing
.

These changes translate into classical slow oscillations of the local field potential during the NREM period
.

In particular, the descending branch is associated with significant LFP deflection, called delta waves
.

Delta waves often accompany spindles, about 10-15 Hz oscillations from the thalamus
.

All of these cortical rhythms, mostly through coordination with other hippocampal and cortical patterns, are involved in memory consolidation
.

Transcranial stimulation in humans can be used to promote slow-wave oscillations during NREM
.

Such manipulations help to regain memory the next day
.

Numerous studies related to EEG have demonstrated the role of slow waves and spindles in memory consolidation
.

In an experiment in rodent brain-computer interfaces, it was interesting to find that animals could be trained to control the reward equipment and self-regulate the firing of pre-specified neurons
.

During this operation, task-related neurons were observed to typically fire in synchrony with the slow-wave rise during subsequent sleep events
.

This experiment further confirmed the relationship between slow-wave oscillations and memory recall
.

The performance of memory reproduction basically conforms to this law, that is, the synchronization is enhanced, and the memory is enhanced
.

By optogenetically disinhibiting synchronization, memory is also impaired
.

In fact, with current approaches to studying the brain, we are still limited to technical, statistical, and conceptual biases
.

We try to study and observe things that are easier for us to observe and interpret, such as periods of high group activity, high-firing neurons, and prominent oscillatory patterns
.

An original study using these methods found that some very rare, normally neglected activity in the prefrontal cortex, in the apparent descending branch of the delta wave, is actually reactivated during learning
.

SWR and cortical NREM sleep patterns are temporally coordinated, a mechanism thought to promote plasticity and long-term consolidation of scene (event) memory
.

The frequency of hippocampal SWR increased during the transition to the up-and-down cortical oscillation and the Spindle's trough
.

Using a closed-loop system to augment the correlative effects of the hippocampus and cortex, generating a spindle complex decline following an SWR, can increase performance in memory tasks
.

The artificial spindle produced by the optogenetic approach coordinated with slow-wave ripples in the hippocampus also enhanced memory, highlighting the importance of the "ripple-delta wave-spindle" complex for memory consolidation
.

Moreover, the spiky part of the hippocampal slow wave ripples can predict the cortical discharge of the following delta waves, suggesting the tendency of the hippocampal slow wave ripples in the reactivation of cortical information
.

In turn, cortical firing can also predict information reactivation in the CA1 area of ​​the hippocampus, and sensory stimuli during sleep can promote the reproduction of hippocampal information and strengthen memory, a phenomenon called target memory reactivation
.

In summary, all evidence suggests that memory consolidation involves circuits that cortically favor the reactivation of memory information in the hippocampus, which, in turn, reactivates multimodally relevant information in the neocortex—a synergy between the hippocampus and the cortex Fundamental ideas in memory consolidation
.

REM Sleep and Theta Oscillations Although the main interest in REM sleep has long been related to vivid dreams in humans, related research is still lacking compared to non-REM sleep
.

The EEG and local field potentials of REM sleep are very similar to the activity during wakefulness, so in the past it was called "paradoxical" sleep
.

In fact, REM sleep is dominated by theta oscillations, dominated by 5-12 Hz
.

The wave is more pronounced in the hippocampus and can also be recorded in cortical and subcortical structures
.

During wakefulness, theta oscillations in the hippocampus organize the orderly firing of place cell firing
.

Fine temporal modulation of theta wave oscillations is critical for the encoding and subsequent consolidation of spatial memory, as these processes require place cells to reproduce during NREM sleep
.

Relatively few studies have focused on REM sleep and theta-related or unrelated neuronal activity
.

The rapid increase in the frequency and amplitude of theta waves during the REM period, known as a phase rapid eye movement period, is associated with increased firing throughout the hippocampus and coordination of cortical areas
.

Phase REM sleep is associated with occipital waves from the brainstem and is thought to coordinate many other structures during REM sleep
.

Currently unknown, behavioral associations between specific changes in REM sleep theta oscillations are poorly studied
.

However, theta oscillations in the hippocampus and the prefrontal and amygdala coordinately enhanced after aversive learning
.

Interfering with theta waves in REM sleep using optogenetic perturbation of the medial septal nucleus impairs hippocampal context-related memory
.

In addition, altered activity of adult newborn neurons in the dentate gyrus of the hippocampus, especially during REM, can impair contextual fear memory
.

Although this manipulation did not affect theta oscillations, both increased and decreased firing impaired memory consolidation, suggesting the importance of the underlying temporal fine-tuning of theta wave rhythm for firing in newborn neurons
.

It has also been reported that inhibition of REM sleep induces mild synaptic structural changes in newborn neurons, suggesting reduced synaptic function
.

These results further support previous findings that REM sleep can enhance the selective enhancement and attenuation of dendritic spines in the neocortex
.

But more work is needed to further correlate the fine-phase modulation of specific firings and theta oscillations in REM sleep with known neuronal dendritic spine synaptic structures, and even behavioral outputs, to truly understand REM The functional significance of sleep in learning and memory or other physiological behaviors
.

Aside from the strong conceptual framework provided by the two-step consolidation approach and the theory of gradual transfer of information from the hippocampus to the cortex, most research on sleep patterns and memory consolidation has focused on the dialogue between the hippocampus and the cortex
.

However, many other structures are also involved in the formation of memories
.

SWR in the hippocampus is a very powerful event that can synchronize other structures outside the neocortex, potentially linking other components such as emotional tone to the consolidation of different forms of memory
.

For example, following a reward experience, reward-localized hippocampal place cells and reward-encoded ventral striatal neurons fire in synchrony with the SWR phase of sleep and hippocampal activity leading to striatal activation
.

In addition, cell populations in the nucleus accumbens, another important nucleus in the dorsal-ventral hippocampus that regulate reward processing, are distinct
.

In the basolateral amygdala, the main center of value encoding, a subpopulation of cells is regulated by SWR in the hippocampus
.

In aversive spatial memory, the neuronal representation of the joint hippocampal amygdala reappears during the SWR phase of NREM
.

These results suggest that the SWR of the hippocampus may be a brain-wide regulator that allows the formation of memory traces across the cortex and in extracortical structures
.

Sleep rhythm and plasticity: Consolidation and homeostatic learning are thought to be related to Hebbian plasticity and synaptic enhancement
.

According to the synaptic homeostasis hypothesis, sleep plays a key role in homeostasis regulation by downregulating synaptic weights to prevent saturation, which in turn allows the formation of new memories in subsequent periods of wakefulness
.

In particular, this model predicts that overall synaptic weight increases during wakefulness; decreases during sleep
.

Despite synaptic structural and molecular evidence supporting this process, measuring synaptic structural changes and strength in vivo and in real time remains challenging
.

Although cortical slow-wave activity originates from highly synchronized up and down states, their amplitudes are thought to reflect the synaptic strength of cortical neurons
.

Correspondingly, slow-wave oscillations were strongest after prolonged wakefulness and then diminished with increasing sleep events
.

This phenomenon is consistent with the synaptic homeostasis hypothesis
.

Moreover, changes in the slope of cortical evoked potentials, a marker of synaptic efficiency, correlated with changes in slow-wave activity, suggesting that slow waves may continue to degrade synapses
.

At the same time, the dynamics of discharges during periods of clarity and sleep were used for the representation of neural excitability
.

Also in line with the synaptic homeostasis hypothesis are hippocampal cells, which, as a population, have a gradual increase in firing during wakefulness
.

During sleep, the activity of the entire neuronal network decreased, but there was an opposite trend in different sleep phases, such as spike activity significantly increased during NREM and decreased during REM
.

Notably, the downregulation of REM phase discharge can be predicted by the frequency of spindle and SWR occurrence
.

Finally, the SWR of classical NREM sleep, long thought to help promote memory consolidation through long-term potentiation, also triggers long-term depression
.

Their inhibition prevents the normal weakening of sleep evoked potentials, suggesting a potential interplay between sleep and synaptic plasticity in homeostasis
.

Outlook Although simple to describe, the relationship between sleep and memory is a very complex area of ​​research
.

First, sleep is not homogeneous and can be divided into stages and substages classified by different rhythms and patterns
.

Second, there are many types of memory (including eventual, semantic, procedural, skill, and Pavlovian conditioning), and these different types of memory rely on distinct but occasionally overlapping structural networks
.

They exhibit different sleep patterns
.

Furthermore, event memory is not a completely true reflection of actual events
.

The formation of event memory, therefore, involves the initial encoding of information, revision, integration with other memories, and even forgetting
.

Imagining the complexity of sleep, memory and engagement structures, how we design relevant basic research to uncover the role sleep plays in memory is our challenge
.

In rodents, the NREM sleep phase has traditionally been studied as a homogeneous period
.

How to identify NREM-specific substages or microstages, and find potential stages that match the three substages of human NREM, will be an entry into linking them and various aspects of memory processing to various behavioral levels point
.

Both phasic and tonic REM sleep are under study in humans and other species
.

At the same time, studying different patterns of sleep may be more reliable than periods
.

The development of closed-loop systems and the brain-computer interface community have important advantages in real-time pattern monitoring of neuronal firing, as well as EEG or LFP signals for the field to understand the role of sleep patterns in memory formation
.

While place cell sequences with empirical traces are reactivated during subsequent SWR sleep
.

However, there is no evidence of causality that the place cell sequence itself is important for memory consolidation compared to the small activation of place cell clusters within a short time window
.

Testing the theory of the importance of spike timing in patterns will require more precise real-time tools, requiring precision in perturbing time across wider temporal dimensions without perturbing specific neuron firing
.

In turn, the problem of correlations in the validation sequence itself could potentially relocate the domain to nearly 80% of SWR-related neuronal content that current decoding algorithms fail to identify
.

These could be reactivation events that we haven't been able to identify yet
.

According to this hypothesis, major functions of SWR-related hypersynchrony events, including those that we have not decoded, promote the consolidation of various memory replays
.

Another emerging, more comprehensive theory is that during sleep, the cortex and hippocampus enter a default mode due to their physiology and the highly synchronized activity of their hardware connections
.

These modes may be primarily used for homeostatic balance purposes
.

But waking activity and memory encoding may be biased toward the precise timing of discharges that consolidate specific memory traces
.

Activity during waking periods makes this bias stronger and more persistent after learning and novel stimuli, resulting in a stronger recurrence-to-noise ratio of the SWR phase
.

From this perspective, homeostasis and consolidation are in the same dimension and strongly depend on the fine timing of neural activity in classical sleep patterns
.

Finally, reactivation, the primary mechanism of memory consolidation, is not universally applicable in terms of brain structure and sleep duration
.

At present, there are many studies on homeostasis in the neocortex.
Therefore, more work needs to be refined to identify the role of sleep patterns in memory processing involving non-hippocampal cortical structures
.

It also needs to be further investigated whether consolidation and homeostatic processes differ in other structures or in particular in those structures where sleep reactivation or different firing rate distributions are not detected
.

In particular, this direction requires more attention and exploration of how highly complex network structures such as the pons, thalamus, hypothalamus, basket plaque and stromal forebrain are controlled and switched during different periods of sleep
.

Motivated by new recording methods and analysis algorithms that are emerging with each passing day, most researchers are currently in the process of researching the mechanism of multi-dimensional knowledge spaces of different memory types and different sleep stages and sub-stages
.

While researchers may never reach a unified theory of memory function in sleep, expanding and refining these knowledge spaces will allow us to better integrate consolidation and homeostasis, revealing memory function at all stages of memory formation from encoding to Reappearing memories consolidate new connections across all processes and link mnemonics to other aspects of sleep, such as memory control, rhythm, and pathology
.

Link to the original text: http://doi.
org/10.
1126/science.
abi8370 Publisher: 11 Reprint Notice [Original article] BioArt original article, welcome to forward and share it, reprinting is prohibited without permission, and the copyright of all works published is Owned by BioArt
.

BioArt reserves all legal rights and violators will be held accountable
.