Impermanence of dendritic spines in live adult CA1 hippocampus

Autoimmune hypothesis of Alzheimer's disease: unanswered question

Alzheimer's disease (AD) was described more than a century ago. However, there are still no effective approaches to its treatment, which may suggest that the search for the cure is not being conducted in the most productive direction. AD begins as selective impairments of declarative memory with no deficits in other cognitive functions. Therefore, understanding of the AD pathogenesis has to include the understanding of this selectivity. Currently, the main efforts aimed at prevention and treatment of AD are based on the dominating hypothesis for the AD pathogenesis: the amyloid hypothesis. But this hypothesis does not explain selective memory impairments since β-amyloid accumulates extracellularly and should be toxic to all types of cerebral neurons, not only to "memory engram neurons." To explain selective memory impairment, I propose the autoimmune hypothesis of AD, based on the analysis of risk factors for AD and molecular mechanisms of memory formation. Memory formation is associated with epigenetic modifications of chromatin in memory engram neurons and, therefore, might be accompanied by the expression of memory-specific proteins recognized by the adaptive immune system as "non-self" antigens. Normally, the brain is protected by the blood-brain barrier (BBB). All risk factors for AD provoke BBB disruptions, possibly leading to an autoimmune reaction against memory engram neurons. This reaction would make them selectively sensitive to tauopathy. If this hypothesis is confirmed, the strategies for AD prevention and treatment would be radically changed.

Fiber-based in vivo imaging: unveiling avenues for exploring mechanisms of synaptic plasticity and neuronal adaptations underlying behavior

In recent decades, various subfields within neuroscience, spanning molecular, cellular, and systemic dimensions, have significantly advanced our understanding of the elaborate molecular and cellular mechanisms that underpin learning, memory, and adaptive behaviors. There have been notable advancements in imaging techniques, particularly in reaching superficial brain structures. This progress has led to their widespread adoption in numerous laboratories. However, essential physiological and cognitive processes, including sensory integration, emotional modulation of motivated behavior, motor regulation, learning, and memory consolidation, are intricately encoded within deeper brain structures. Hence, visualization techniques such as calcium imaging using miniscopes have gained popularity for studying brain activity in unrestrained animals. Despite its utility, miniscope technology is associated with substantial brain tissue damage caused by gradient refractive index lens implantation. Furthermore, its imaging capabilities are primarily confined to the neuronal somata level, thus constraining a comprehensive exploration of subcellular processes underlying adaptive behaviors. Consequently, the trajectory of neuroscience's future hinges on the development of minimally invasive optical fiber-based endo-microscopes optimized for cellular, subcellular, and molecular imaging within the intricate depths of the brain. In pursuit of this goal, select research groups have invested significant efforts in advancing this technology. In this review, we present a perspective on the potential impact of this innovation on various aspects of neuroscience, enabling the functional exploration of in vivo cellular and subcellular processes that underlie synaptic plasticity and the neuronal adaptations that govern behavior.

"There's plenty of room at the bottom": deep brain imaging with holographic endo-microscopy

Significance

Over more than 300 years, microscopic imaging keeps providing fundamental insights into the mechanisms of living organisms. Seeing microscopic structures beyond the reach of free-space light-based microscopy, however, requires dissection of the tissue-an intervention seriously disturbing its physiological functions. The hunt for low-invasiveness tools has led a growing community of physicists and engineers into the realm of complex media photonics. One of its activities represents exploiting multimode optical fibers (MMFs) as ultra-thin endoscopic probes. Employing wavefront shaping, these tools only recently facilitated the first peeks at cells and their sub-cellular compartments at the bottom of the mouse brain with the impact of micro-scale tissue damage.

Aim

Here, we aim to highlight advances in MMF-based holographic endo-microscopy facilitating microscopic imaging throughout the whole depth of the mouse brain.

Approach

We summarize the important technical and methodological prerequisites for stabile high-resolution imaging in vivo.

Results

We showcase images of the microscopic building blocks of brain tissue, including neurons, neuronal processes, vessels, intracellular calcium signaling, and red blood cell velocity in individual vessels.

Conclusions

This perspective article helps to understand the complexity behind the technology of holographic endo-microscopy, summarizes its recent advances and challenges, and stimulates the mind of the reader for further exploitation of this tool in the neuroscience research.

Non-spatial hippocampal behavioral timescale synaptic plasticity during working memory is gated by entorhinal inputs

Behavioral timescale synaptic plasticity (BTSP) is a form of synaptic potentiation where the occurrence of a single large plateau potential in CA1 hippocampal neurons leads to the formation of reliable place fields during spatial learning tasks. We asked whether BTSP could also be a plasticity mechanism for generation of non-spatial responses in the hippocampus and what roles the medial and lateral entorhinal cortex (MEC and LEC) play in driving non-spatial BTSP. By performing simultaneous calcium imaging of dorsal CA1 neurons and chemogenetic inhibition of LEC or MEC while mice performed an olfactory working memory task, we discovered BTSP-like events which formed stable odor-specific fields. Critically, the success rate of calcium events generating a significant odor-field increased with event amplitude, and large events exhibited asymmetrical formation with the newly formed odor-fields preceding the timepoint of their induction event. We found that MEC and LEC play distinct roles in modulating BTSP: MEC inhibition reduced the frequency of large calcium events, while LEC inhibition reduced the success rate of odor-field generation. Using two-photon calcium imaging of LEC and MEC temporammonic axons projecting to CA1, we found that LEC projections to CA1 were strongly odor selective even early in task learning, while MEC projection odor-selectivity increased with task learning but remained weaker than LEC. Finally, we found that LEC and MEC inhibition both slowed representational drift of odor representations in CA1 across 48 hours. Altogether, odor-specific information from LEC and strong odor-timed activity from MEC are crucial for driving BTSP in CA1, which is a synaptic plasticity mechanism for generation of both spatial and non-spatial responses in the hippocampus that may play a role in explaining representational drift and one-shot learning of non-spatial information.

A dendritic mechanism for balancing synaptic flexibility and stability

Biological and artificial neural networks learn by modifying synaptic weights, but it is unclear how these systems retain previous knowledge and also acquire new information. Here, we show that cortical pyramidal neurons can solve this plasticity-versus-stability dilemma by differentially regulating synaptic plasticity at distinct dendritic compartments. Oblique dendrites of adult mouse layer 5 cortical pyramidal neurons selectively receive monosynaptic thalamic input, integrate linearly, and lack burst-timing synaptic potentiation. In contrast, basal dendrites, which do not receive thalamic input, exhibit conventional NMDA receptor (NMDAR)-mediated supralinear integration and synaptic potentiation. Congruently, spiny synapses on oblique branches show decreased structural plasticity in vivo. The selective decline in NMDAR activity and expression at synapses on oblique dendrites is controlled by a critical period of visual experience. Our results demonstrate a biological mechanism for how single neurons can safeguard a set of inputs from ongoing plasticity by altering synaptic properties at distinct dendritic domains.

The estrous cycle modulates hippocampal spine dynamics, dendritic processing, and spatial coding

Histological evidence suggests that the estrous cycle exerts a powerful effect on CA1 neurons in mammalian hippocampus. Decades have passed since this landmark observation, yet how the estrous cycle shapes dendritic spine dynamics and hippocampal spatial coding in vivo remains a mystery. Here, we used a custom hippocampal microperiscope and two-photon calcium imaging to track CA1 pyramidal neurons in female mice over multiple cycles. Estrous cycle stage had a potent effect on spine dynamics, with heightened density during periods of greater estradiol (proestrus). These morphological changes were accompanied by greater somatodendritic coupling and increased infiltration of back-propagating action potentials into the apical dendrite. Finally, tracking CA1 response properties during navigation revealed enhanced place field stability during proestrus, evident at the single-cell and population level. These results establish the estrous cycle as a driver of large-scale structural and functional plasticity in hippocampal circuits essential for learning and memory.

Maintenance of taste receptor cell presynaptic sites requires gustatory nerve fibers

The turnover and re-establishment of peripheral taste synapses is vital to maintain connectivity between the primary taste receptor cells and the gustatory neurons which relay taste information from the tongue to the brain. Despite the importance of neuron-taste cell reconnection, mechanisms governing synapse assembly and the specificity of synaptic connections is largely unknown. Here we use the expression of presynaptic proteins, CALHM1 and Bassoon, to probe whether nerve fiber connectivity is an initiating factor for the recruitment of presynaptic machinery in different populations of taste cells. Under homeostatic conditions, the vast majority (>90%) of presynaptic sites are directly adjacent to nerve fibers. In the days immediately following gustatory nerve transection and complete denervation, Bassoon and CALHM1 puncta are markedly reduced. This suggests that nerve fiber innervation is crucial for the recruitment and maintenance of presynaptic sites. In support of this, we find that expression of Bassoon and Calhm1 mRNA transcripts are significantly reduced after denervation. During nerve fiber regeneration into the taste bud, presynaptic sites begin to replenish, but are not as frequently connected to nerve fibers as intact controls (∼50% compared to >90%). This suggests that gustatory neuron proximity, rather than direct contact, likely drives taste receptor cells to express and aggregate presynaptic proteins at the cell membrane. Together, these data support the idea that trophic factors secreted by gustatory nerve fibers prompt taste receptor cells to produce presynaptic specializations at the cell membrane, which in turn may guide neurons to form mature synapses. These findings provide new insights into the mechanisms driving synaptogenesis and synaptic plasticity within the rapidly changing taste bud environment.

Excitatory and inhibitory synapses show a tight subcellular correlation that weakens over development

Neurons receive correlated levels of excitation and inhibition, a feature that is important for proper brain function. However, how this relationship between excitatory and inhibitory inputs is established during the dynamic period of circuit wiring remains unexplored. Using multiple techniques, including in utero electroporation, electron microscopy, and electrophysiology, we reveal a tight correlation in the distribution of excitatory and inhibitory synapses along the dendrites of developing CA1 hippocampal neurons. This correlation was present within short dendritic stretches (<20 μm) and, surprisingly, was most pronounced during early development, sharply declining with maturity. The tight matching between excitation and inhibition was unexpected, as inhibitory synapses lacked an active zone when formed and exhibited compromised evoked release. We propose that inhibitory synapses form as a stabilizing scaffold to counterbalance growing excitation levels. This relationship diminishes over time, suggesting a critical role for a subcellular balance in early neuronal function and circuit formation.

Structural plasticity of pyramidal cell neurons measured after FLASH and conventional dose-rate irradiation

Evidence shows that ultra-high dose-rate FLASH-radiotherapy (FLASH-RT) protects against normal tissue complications and functional decrements in the irradiated brain. Past work has shown that radiation-induced cognitive impairment, neuroinflammation and reduced structural complexity of granule cell neurons were not observed to the same extent after FLASH-RT (> MGy/s) compared to conventional dose-rate (CONV, 0.1 Gy/s) delivery. To explore the sensitivity of different neuronal populations to cranial irradiation and dose-rate modulation, hippocampal CA1 and medial prefrontal cortex (PFC) pyramidal neurons were analyzed by electron and confocal microscopy. Neuron ultrastructural analyses by electron microscopy after 10 Gy FLASH- or CONV-RT exposures indicated that irradiation had little impact on dendritic complexity and synapse density in the CA1, but did increase length and head diameter of smaller non-perforated synapses. Similarly, irradiation caused no change in PFC prelimbic/infralimbic axospinous synapse density, but reductions in non-perforated synapse diameters. While irradiation resulted in thinner myelin sheaths compared to controls, none of these metrics were dose-rate sensitive. Analysis of fluorescently labeled CA1 neurons revealed no radiation-induced or dose-rate-dependent changes in overall dendritic complexity or spine density, in contrast to our past analysis of granule cell neurons. Super-resolution confocal microscopy following a clinical dosing paradigm (3×10Gy) showed significant reductions in excitatory vesicular glutamate transporter 1 and inhibitory vesicular GABA transporter puncta density within the CA1 that were largely dose-rate independent. Collectively, these data reveal that, compared to granule cell neurons, CA1 and mPFC neurons are more radioresistant irrespective of radiation dose-rate.

The Effects of Sevoflurane and Aβ Interaction on CA1 Dendritic Spine Dynamics and MEGF10-Related Astrocytic Synapse Engulfment

General anesthetics may accelerate the neuropathological changes related to Alzheimer's disease (AD), of which amyloid beta (Aβ)-induced toxicity is one of the main causes. However, the interaction of general anesthetics with different Aβ-isoforms remains unclear. In this study, we investigated the effects of sevoflurane (0.4 and 1.2 maximal alveolar concentration (MAC)) on four Aβ species-induced changes on dendritic spine density (DSD) in hippocampal brain slices of Thy1-eGFP mice and multiple epidermal growth factor-like domains 10 (MEGF10)-related astrocyte-mediated synaptic engulfment in hippocampal brain slices of C57BL/6 mice. We found that both sevoflurane and Aβ downregulated CA1-dendritic spines. Moreover, compared with either sevoflurane or Aβ alone, pre-treatment with Aβ isoforms followed by sevoflurane application in general further enhanced spine loss. This enhancement was related to MEGF10-related astrocyte-dependent synaptic engulfment, only in AβpE3 + 1.2 MAC sevoflurane and 3NTyrAβ + 1.2 MAC sevoflurane condition. In addition, removal of sevoflurane alleviated spine loss in Aβ + sevoflurane. In summary, these results suggest that both synapses and astrocytes are sensitive targets for sevoflurane; in the presence of 3NTyrAβ, 1.2 MAC sevoflurane alleviated astrocyte-mediated synaptic engulfment and exerted a lasting effect on dendritic spine remodeling.

Robust variability of grid cell properties within individual grid modules enhances encoding of local space

Although grid cells are one of the most well studied functional classes of neurons in the mammalian brain, the assumption that there is a single grid orientation and spacing per grid module has not been carefully tested. We investigate and analyze a recent large-scale recording of medial entorhinal cortex to characterize the presence and degree of heterogeneity of grid properties within individual modules. We find evidence for small, but robust, variability and hypothesize that this property of the grid code could enhance the ability of encoding local spatial information. Performing analysis on synthetic populations of grid cells, where we have complete control over the amount heterogeneity in grid properties, we demonstrate that variability, of a similar magnitude to the analyzed data, leads to significantly decreased decoding error, even when restricted to activity from a single module. Our results highlight how the heterogeneity of the neural response properties may benefit coding and opens new directions for theoretical and experimental analysis of grid cells.

Memory consolidation affects the interplay of place and response navigation

Navigation through space is based on memory representations of landmarks ('place') or movement sequences ('response'). Over time, memory representations transform through consolidation. However, it is unclear how the transformation affects place and response navigation in humans. In the present study, healthy adults navigated to target locations in a virtual maze. The preference for using place and response strategies and the ability to recall place and response memories were tested after a delay of one hour (n = 31), one day (n = 30), or two weeks (n = 32). The different delays captured early-phase synaptic changes, changes after one night of sleep, and long-delay changes due to the reorganization of navigation networks. Our results show that the relative contributions of place and response navigation changed as a function of time. After a short delay of up to one day, participants preferentially used a place strategy and exhibited a high degree of visual landmark exploration. After a longer delay of two weeks, place strategy use decreased significantly. Participants now equally relied on place and response strategy use and increasingly repeated previously taken paths. Further analyses indicate that response strategy use predominantly occurred as a compensatory strategy in the absence of sufficient place memory. Over time, place memory faded before response memory. We suggest that the observed shift from place to response navigation is context-dependent since detailed landmark information, which strongly relied on hippocampal function, decayed faster than sequence information, which required less detail and depended on extra-hippocampal areas. We conclude that changes in place and response navigation likely reflect the reorganization of navigation networks during systems consolidation.

The purpose and ubiquity of turnover

Turnover-constant component production and destruction-is ubiquitous in biology. Turnover occurs across organisms and scales, including for RNAs, proteins, membranes, macromolecular structures, organelles, cells, hair, feathers, nails, antlers, and teeth. For many systems, turnover might seem wasteful when degraded components are often fully functional. Some components turn over with shockingly high rates and others do not turn over at all, further making this process enigmatic. However, turnover can address fundamental problems by yielding powerful properties, including regeneration, rapid repair onset, clearance of unpredictable damage and errors, maintenance of low constitutive levels of disrepair, prevention of stable hazards, and transitions. I argue that trade-offs between turnover benefits and metabolic costs, combined with constraints on turnover, determine its presence and rates across distinct contexts. I suggest that the limits of turnover help explain aging and that turnover properties and the basis for its levels underlie this fundamental component of life.

Drift of neural ensembles driven by slow fluctuations of intrinsic excitability

Representational drift refers to the dynamic nature of neural representations in the brain despite the behavior being seemingly stable. Although drift has been observed in many different brain regions, the mechanisms underlying it are not known. Since intrinsic neural excitability is suggested to play a key role in regulating memory allocation, fluctuations of excitability could bias the reactivation of previously stored memory ensembles and therefore act as a motor for drift. Here, we propose a rate-based plastic recurrent neural network with slow fluctuations of intrinsic excitability. We first show that subsequent reactivations of a neural ensemble can lead to drift of this ensemble. The model predicts that drift is induced by co-activation of previously active neurons along with neurons with high excitability which leads to remodeling of the recurrent weights. Consistent with previous experimental works, the drifting ensemble is informative about its temporal history. Crucially, we show that the gradual nature of the drift is necessary for decoding temporal information from the activity of the ensemble. Finally, we show that the memory is preserved and can be decoded by an output neuron having plastic synapses with the main region.

Lifelong absence of microglia alters hippocampal glutamatergic networks but not synapse and spine density

Microglia sculpt developing neural circuits by eliminating excess synapses in a process called synaptic pruning, by removing apoptotic neurons, and by promoting neuronal survival. To elucidate the role of microglia during embryonic and postnatal brain development, we used a mouse model deficient in microglia throughout life by deletion of the fms-intronic regulatory element (FIRE) in the Csf1r locus. Surprisingly, young adult Csf1rΔFIRE/ΔFIRE mice display no changes in excitatory and inhibitory synapse number and spine density of CA1 hippocampal neurons compared with Csf1r+/+ littermates. However, CA1 neurons are less excitable, receive less CA3 excitatory input and show altered synaptic properties, but this does not affect novel object recognition. Cytokine profiling indicates an anti-inflammatory state along with increases in ApoE levels and reactive astrocytes containing synaptic markers in Csf1rΔFIRE/ΔFIRE mice. Notably, these changes in Csf1rΔFIRE/ΔFIRE mice closely resemble the effects of acute microglial depletion in adult mice after normal development. Our findings suggest that microglia are not mandatory for synaptic pruning, and that in their absence pruning can be achieved by other mechanisms.

Synaptic architecture of a memory engram in the mouse hippocampus

Memory engrams are formed through experience-dependent remodeling of neural circuits, but their detailed architectures have remained unresolved. Using 3D electron microscopy, we performed nanoscale reconstructions of the hippocampal CA3-CA1 pathway following chemogenetic labeling of cellular ensembles with a remote history of correlated excitation during associative learning. Projection neurons involved in memory acquisition expanded their connectomes via multi-synaptic boutons without altering the numbers and spatial arrangements of individual axonal terminals and dendritic spines. This expansion was driven by presynaptic activity elicited by specific negative valence stimuli, regardless of the co-activation state of postsynaptic partners. The rewiring of initial ensembles representing an engram coincided with local, input-specific changes in the shapes and organelle composition of glutamatergic synapses, reflecting their weights and potential for further modifications. Our findings challenge the view that the connectivity among neuronal substrates of memory traces is governed by Hebbian mechanisms, and offer a structural basis for representational drifts.

Liprin-α proteins are master regulators of human presynapse assembly

The formation of mammalian synapses entails the precise alignment of presynaptic release sites with postsynaptic receptors but how nascent cell-cell contacts translate into assembly of presynaptic specializations remains unclear. Guided by pioneering work in invertebrates, we hypothesized that in mammalian synapses, liprin-α proteins directly link trans-synaptic initial contacts to downstream steps. Here we show that, in human neurons lacking all four liprin-α isoforms, nascent synaptic contacts are formed but recruitment of active zone components and accumulation of synaptic vesicles is blocked, resulting in 'empty' boutons and loss of synaptic transmission. Interactions with presynaptic cell adhesion molecules of either the LAR-RPTP family or neurexins via CASK are required to localize liprin-α to nascent synaptic sites. Liprin-α subsequently recruits presynaptic components via a direct interaction with ELKS proteins. Thus, assembly of human presynaptic terminals is governed by a hierarchical sequence of events in which the recruitment of liprin-α proteins by presynaptic cell adhesion molecules is a critical initial step.

Local and global predictors of synapse elimination during motor learning

During learning, synaptic connections between excitatory neurons in the brain display considerable dynamism, with new connections being added and old connections eliminated. Synapse elimination offers an opportunity to understand the features of synapses that the brain deems dispensable. However, with limited observations of synaptic activity and plasticity in vivo, the features of synapses subjected to elimination remain poorly understood. Here, we examined the functional basis of synapse elimination in the apical dendrites of L2/3 neurons in the primary motor cortex throughout motor learning. We found no evidence that synapse elimination is facilitated by a lack of activity or other local forms of plasticity. Instead, eliminated synapses display asynchronous activity with nearby synapses, suggesting that functional synaptic clustering is a critical component of synapse survival. In addition, eliminated synapses show delayed activity timing with respect to postsynaptic output. Thus, synaptic inputs that fail to be co-active with their neighboring synapses or are mistimed with neuronal output are targeted for elimination.

DeepD3, an open framework for automated quantification of dendritic spines

Dendritic spines are the seat of most excitatory synapses in the brain, and a cellular structure considered central to learning, memory, and activity-dependent plasticity. The quantification of dendritic spines from light microscopy data is usually performed by humans in a painstaking and error-prone process. We found that human-to-human variability is substantial (inter-rater reliability 82.2±6.4%), raising concerns about the reproducibility of experiments and the validity of using human-annotated 'ground truth' as an evaluation method for computational approaches of spine identification. To address this, we present DeepD3, an open deep learning-based framework to robustly quantify dendritic spines in microscopy data in a fully automated fashion. DeepD3's neural networks have been trained on data from different sources and experimental conditions, annotated and segmented by multiple experts and they offer precise quantification of dendrites and dendritic spines. Importantly, these networks were validated in a number of datasets on varying acquisition modalities, species, anatomical locations and fluorescent indicators. The entire DeepD3 open framework, including the fully segmented training data, a benchmark that multiple experts have annotated, and the DeepD3 model zoo is fully available, addressing the lack of openly available datasets of dendritic spines while offering a ready-to-use, flexible, transparent, and reproducible spine quantification method.

Deep-brain optical recording of neural dynamics during behavior

In vivo fluorescence recording techniques have produced landmark discoveries in neuroscience, providing insight into how single cell and circuit-level computations mediate sensory processing and generate complex behaviors. While much attention has been given to recording from cortical brain regions, deep-brain fluorescence recording is more complex because it requires additional measures to gain optical access to harder to reach brain nuclei. Here we discuss detailed considerations and tradeoffs regarding deep-brain fluorescence recording techniques and provide a comprehensive guide for all major steps involved, from project planning to data analysis. The goal is to impart guidance for new and experienced investigators seeking to use in vivo deep fluorescence optical recordings in awake, behaving rodent models.

An optical clearing imaging window: Realization of mouse brain imaging and manipulation through scalp and skull

Cortical visualization is essential to understand the dynamic changes in brain microenvironment under physiopathological conditions. However, the turbid scalp and skull severely limit the imaging depth and resolution. Existing cranial windows require invasive scalp excision and various subsequent skull treatments. Non-invasive in vivo imaging of skull bone marrow, meninges, and cortex through scalp and skull with high resolution yet remains a challenge. In this work, a non-invasive trans-scalp/skull optical clearing imaging window is proposed for cortical and calvarial imaging, which is achieved by applying a novel skin optical clearing reagent. The imaging depth and resolution are greatly enhanced in near infrared imaging and optical coherence tomography imaging. Combining this imaging window with adaptive optics, we achieve the visualization and manipulation of the calvarial and cortical microenvironment through the scalp and skull using two-photon imaging for the first time. Our method provides a well-performed imaging window and paves the way for intravital brain studies with the advantages of easy-operation, convenience and non-invasiveness.

Cortical reactivation of spatial and non-spatial features coordinates with hippocampus to form a memory dialogue

Episodic memories comprise diverse attributes of experience distributed across neocortical areas. The hippocampus is integral to rapidly binding these diffuse representations, as they occur, to be later reinstated. However, the nature of the information exchanged during this hippocampal-cortical dialogue remains poorly understood. A recent study has shown that the secondary motor cortex carries two types of representations: place cell-like activity, which were impaired by hippocampal lesions, and responses tied to visuo-tactile cues, which became more pronounced following hippocampal lesions. Using two-photon Ca2+ imaging to record neuronal activities in the secondary motor cortex of male Thy1-GCaMP6s mice, we assessed the cortical retrieval of spatial and non-spatial attributes from previous explorations in a virtual environment. We show that, following navigation, spontaneous resting state reactivations convey varying degrees of spatial (trajectory sequences) and non-spatial (visuo-tactile attributes) information, while reactivations of non-spatial attributes tend to precede reactivations of spatial representations surrounding hippocampal sharp-wave ripples.

Memory: Synaptic or Cellular, That Is the Question

According to the commonly accepted opinion, memory engrams are formed and stored at the level of neural networks due to a change in the strength of synaptic connections between neurons. This hypothesis of synaptic plasticity (HSP), formulated by Donald Hebb in the 1940s, continues to dominate the directions of experimental studies and the interpretations of experimental results in the field. The universal acceptance of the HSP has transformed it from a hypothesis into an incontrovertible theory. In this article, I show that the entire body of experimental and clinical data obtained in studies of long-term memory in mammals and humans is inconsistent with the HSP. Instead, these data suggest that long-term memory is formed and stored at the intracellular level where it is reliably protected from ongoing synaptic activity, including pathological epileptic activity. It seems that the generally accepted HSP became a serious obstacle to understanding the mechanisms of memory and that progress in this field requires rethinking this doctrine and shifting experimental efforts toward exploring the intracellular mechanisms.

Synaptic memory and CaMKII

Ca2+/calmodulin-dependent protein kinase II (CaMKII) and long-term potentiation (LTP) were discovered within a decade of each other and have been inextricably intertwined ever since. However, like many marriages, it has had its up and downs. Based on the unique biochemical properties of CaMKII, it was proposed as a memory molecule before any physiological linkage was made to LTP. However, as reviewed here, the convincing linkage of CaMKII to synaptic physiology and behavior took many decades. New technologies were critical in this journey, including in vitro brain slices, mouse genetics, single-cell molecular genetics, pharmacological reagents, protein structure, and two-photon microscopy, as were new investigators attracted by the exciting challenge. This review tracks this journey and assesses the state of this marriage 40 years on. The collective literature impels us to propose a relatively simple model for synaptic memory involving the following steps that drive the process: 1) Ca2+ entry through N-methyl-d-aspartate (NMDA) receptors activates CaMKII. 2) CaMKII undergoes autophosphorylation resulting in constitutive, Ca2+-independent activity and exposure of a binding site for the NMDA receptor subunit GluN2B. 3) Active CaMKII translocates to the postsynaptic density (PSD) and binds to the cytoplasmic C-tail of GluN2B. 4) The CaMKII-GluN2B complex initiates a structural rearrangement of the PSD that may involve liquid-liquid phase separation. 5) This rearrangement involves the PSD-95 scaffolding protein, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs), and their transmembrane AMPAR-regulatory protein (TARP) auxiliary subunits, resulting in an accumulation of AMPARs in the PSD that underlies synaptic potentiation. 6) The stability of the modified PSD is maintained by the stability of the CaMKII-GluN2B complex. 7) By a process of subunit exchange or interholoenzyme phosphorylation CaMKII maintains synaptic potentiation in the face of CaMKII protein turnover. There are many other important proteins that participate in enlargement of the synaptic spine or modulation of the steps that drive and maintain the potentiation. In this review we critically discuss the data underlying each of the steps. As will become clear, some of these steps are more firmly grounded than others, and we provide suggestions as to how the evidence supporting these steps can be strengthened or, based on the new data, be replaced. Although the journey has been a long one, the prospect of having a detailed cellular and molecular understanding of learning and memory is at hand.

Imaging dendritic spines in the hippocampus of a living mouse by 3D-stimulated emission depletion microscopy

Significance

Stimulated emission depletion (STED) microscopy has been used to address a wide range of neurobiological questions in optically well-accessible samples, such as cell culture or brain slices. However, the application of STED to deeply embedded structures in the brain of living animals remains technically challenging.

Aim

In previous work, we established chronic STED imaging in the hippocampus in vivo but the gain in spatial resolution was restricted to the lateral plane. In our study, we report on extending the gain in STED resolution into the optical axis to visualize dendritic spines in the hippocampus in vivo.

Approach

Our approach is based on a spatial light modulator to shape the focal STED light intensity in all three dimensions and a conically shaped window that is compatible with an objective that has a long working distance and a high numerical aperture. We corrected distortions of the laser wavefront to optimize the shape of the bottle beam of the STED laser.

Results

We show how the new window design improves the STED point spread function and the spatial resolution using nanobeads. We then demonstrate the beneficial effects for 3D-STED microscopy of dendritic spines, visualized with an unprecedented level of detail in the hippocampus of a living mouse.

Conclusions

We present a methodology to improve the axial resolution for STED microscopy in the deeply embedded hippocampus in vivo, facilitating longitudinal studies of neuroanatomical plasticity at the nanoscale in a wide range of (patho-)physiological contexts.

Homosynaptic plasticity induction causes heterosynaptic changes at the unstimulated neighbors in an induction pattern and location-specific manner

Dendritic spines are highly dynamic structures whose structural and functional fluctuations depend on multiple factors. Changes in synaptic strength are not limited to synapses directly involved in specific activity patterns. Unstimulated clusters of neighboring spines in and around the site of stimulation can also undergo alterations in strength. Usually, when plasticity is induced at single dendritic spines with glutamate uncaging, neighboring spines do not show any significant structural fluctuations. Here, using two-photon imaging and glutamate uncaging at single dendritic spines of hippocampal pyramidal neurons, we show that structural modifications at unstimulated neighboring spines occur and are a function of the temporal pattern of the plasticity-inducing stimulus. Further, the relative location of the unstimulated neighbors within the local dendritic segment correlates with the extent of heterosynaptic plasticity that is observed. These findings indicate that naturalistic patterns of activity at single spines can shape plasticity at nearby clusters of synapses, and may play a role in priming local inputs for further modifications.

Attention Deficit-Hyperactivity Disorder (ADHD): From Abnormal Behavior to Impairment in Synaptic Plasticity

Attention deficit-hyperactivity disorder (ADHD) is a neurodevelopmental disorder with high incidence in children and adolescents characterized by motor hyperactivity, impulsivity, and inattention. Magnetic resonance imaging (MRI) has revealed that neuroanatomical abnormalities such as the volume reduction in the neocortex and hippocampus are shared by several neuropsychiatric diseases such as schizophrenia, autism spectrum disorder and ADHD. Furthermore, the abnormal development and postnatal pruning of dendritic spines of neocortical neurons in schizophrenia, autism spectrum disorder and intellectual disability are well documented. Dendritic spines are dynamic structures exhibiting Hebbian and homeostatic plasticity that triggers intracellular cascades involving glutamate receptors, calcium influx and remodeling of the F-actin network. The long-term potentiation (LTP)-induced insertion of postsynaptic glutamate receptors is associated with the enlargement of spine heads and long-term depression (LTD) with spine shrinkage. Using a murine model of ADHD, a delay in dendritic spines' maturation in CA1 hippocampal neurons correlated with impaired working memory and hippocampal LTP has recently reported. The aim of this review is to summarize recent evidence that has emerged from studies focused on the neuroanatomical and genetic features found in ADHD patients as well as reports from animal models describing the molecular structure and remodeling of dendritic spines.

Probabilistic Neural Computing with Stochastic Devices

The brain has effectively proven a powerful inspiration for the development of computing architectures in which processing is tightly integrated with memory, communication is event-driven, and analog computation can be performed at scale. These neuromorphic systems increasingly show an ability to improve the efficiency and speed of scientific computing and artificial intelligence applications. Herein, it is proposed that the brain's ubiquitous stochasticity represents an additional source of inspiration for expanding the reach of neuromorphic computing to probabilistic applications. To date, many efforts exploring probabilistic computing have focused primarily on one scale of the microelectronics stack, such as implementing probabilistic algorithms on deterministic hardware or developing probabilistic devices and circuits with the expectation that they will be leveraged by eventual probabilistic architectures. A co-design vision is described by which large numbers of devices, such as magnetic tunnel junctions and tunnel diodes, can be operated in a stochastic regime and incorporated into a scalable neuromorphic architecture that can impact a number of probabilistic computing applications, such as Monte Carlo simulations and Bayesian neural networks. Finally, a framework is presented to categorize increasingly advanced hardware-based probabilistic computing technologies.

Representational drift as a window into neural and behavioural plasticity

Large-scale recordings of neural activity over days and weeks have revealed that neural representations of familiar tasks, precepts and actions continually evolve without obvious changes in behaviour. We hypothesise that this steady drift in neural activity and accompanying physiological changes is due in part to the continuous application of a learning rule at the cellular and population level. Explicit predictions of this drift can be found in neural network models that use iterative learning to optimise weights. Drift therefore provides a measurable signal that can reveal systems-level properties of biological plasticity mechanisms, such as their precision and effective learning rates.

A multilevel account of hippocampal function in spatial and concept learning: Bridging models of behavior and neural assemblies

A complete neuroscience requires multilevel theories that address phenomena ranging from higher-level cognitive behaviors to activities within a cell. We propose an extension to the level of mechanism approach where a computational model of cognition sits in between behavior and brain: It explains the higher-level behavior and can be decomposed into lower-level component mechanisms to provide a richer understanding of the system than any level alone. Toward this end, we decomposed a cognitive model into neuron-like units using a neural flocking approach that parallels recurrent hippocampal activity. Neural flocking coordinates units that collectively form higher-level mental constructs. The decomposed model suggested how brain-scale neural populations coordinate to form assemblies encoding concept and spatial representations and why so many neurons are needed for robust performance at the cognitive level. This multilevel explanation provides a way to understand how cognition and symbol-like representations are supported by coordinated neural populations (assemblies) formed through learning.

Cognitive Dynamics Estimation: A whole-brain spatial regression paradigm for extracting the temporal dynamics of cognitive processes

Functional MRI (fMRI) has been instrumental in understanding how cognitive processes are spatially mapped in the brain, yielding insights about brain regions and functions. However, in case the orthogonality of behavioral or stimulus timing is not guaranteed, the estimated brain maps fail to dissociate each cognitive process, and the resultant maps become unstable. Also, the brain mapping exercise can not provide temporal information on the cognitive process. Here we propose a qualitatively different approach to fMRI analysis, named Cognitive Dynamics Estimation (CDE), that estimates how multiple cognitive processes change over time even when behavior or stimulus logs are unavailable. This method transposes the conventional brain mapping; the brain activity pattern at each time point is subject to regression analysis with data-driven maps of cognitive processes as regressors, resulting in the time series of cognitive processes. The estimated time series captured the fluctuation of intensity and timing of cognitive processes on a trial-by-trial basis, which conventional analysis could not capture. Notably, the estimated time series predicted participants' cognitive ability to perform each psychological task. As an addition to our fMRI analytic toolkit, these results suggest the potential for CDE to elucidate underexplored cognitive phenomena, especially in the temporal domain.

Time and experience differentially affect distinct aspects of hippocampal representational drift

Hippocampal activity is critical for spatial memory. Within a fixed, familiar environment, hippocampal codes gradually change over timescales of days to weeks-a phenomenon known as representational drift. The passage of time and the amount of experience are two factors that profoundly affect memory. However, thus far, it has remained unclear to what extent these factors drive hippocampal representational drift. Here, we longitudinally recorded large populations of hippocampal neurons in mice while they repeatedly explored two different familiar environments that they visited at different time intervals over weeks. We found that time and experience differentially affected distinct aspects of representational drift: the passage of time drove changes in neuronal activity rates, whereas experience drove changes in the cells' spatial tuning. Changes in spatial tuning were context specific and largely independent of changes in activity rates. Thus, our results suggest that representational drift is a multi-faceted process governed by distinct neuronal mechanisms.

Multi-synaptic boutons are a feature of CA1 hippocampal connections in the stratum oriens

Excitatory synapses are typically described as single synaptic boutons (SSBs), where one presynaptic bouton contacts a single postsynaptic spine. Using serial section block-face scanning electron microscopy, we found that this textbook definition of the synapse does not fully apply to the CA1 region of the hippocampus. Roughly half of all excitatory synapses in the stratum oriens involved multi-synaptic boutons (MSBs), where a single presynaptic bouton containing multiple active zones contacted many postsynaptic spines (from 2 to 7) on the basal dendrites of different cells. The fraction of MSBs increased during development (from postnatal day 22 [P22] to P100) and decreased with distance from the soma. Curiously, synaptic properties such as active zone (AZ) or postsynaptic density (PSD) size exhibited less within-MSB variation when compared with neighboring SSBs, features that were confirmed by super-resolution light microscopy. Computer simulations suggest that these properties favor synchronous activity in CA1 networks.

Paths to hippocampal damage in neuromyelitis optica spectrum disorders

AIMS

Many patients with neuromyelitis optica spectrum disorders (NMOSD) suffer from cognitive impairment affecting memory, processing speed and attention and suffer from depressive symptoms. Because some of these manifestations could trace back to the hippocampus, several magnetic resonance imaging (MRI) studies have been performed in the past, with a number of groups describing volume loss of the hippocampus in NMOSD patients, whereas others did not observe such changes. Here, we addressed these discrepancies.

METHODS

We performed pathological and MRI studies on the hippocampi of NMOSD patients, combined with detailed immunohistochemical analysis of hippocampi from experimental models of NMOSD.

RESULTS

We identified different pathological scenarios for hippocampal damage in NMOSD and its experimental models. In the first case, the hippocampus was compromised by the initiation of astrocyte injury in this brain region and subsequent local effects of microglial activation and neuronal damage. In the second case, loss of hippocampal volume was seen by MRI in patients with large tissue-destructive lesions in the optic nerves or the spinal cord, and the pathological work-up of tissue derived from a patient with such lesions revealed subsequent retrograde neuronal degeneration affecting different axonal tracts and neuronal networks. It remains to be seen whether remote lesions and associated retrograde neuronal degeneration on their own are sufficient to cause extensive volume loss of the hippocampus, or whether they act in concert with small astrocyte-destructive, microglia-activating lesions in the hippocampus that escape detection by MRI, either due to their small size or due to the chosen time window for examination.

CONCLUSIONS

Different pathological scenarios can culminate in hippocampal volume loss in NMOSD patients.

Neurexin-2 restricts synapse numbers and restrains the presynaptic release probability by an alternative splicing-dependent mechanism

α- and β-neurexins are extensively alternatively spliced, presynaptic cell-adhesion molecules that are thought to organize synapse assembly. However, recent data revealed that, in the hippocampus in vivo, the deletion of one neurexin isoform, Nrxn2, surprisingly increased excitatory synapse numbers and enhanced their presynaptic release probability, suggesting that Nrxn2 restricts, instead of enabling, synapse assembly. To delineate the synaptic function and mechanism of action of Nrxn2, we examined cultured hippocampal neurons as a reduced system. In heterologous synapse formation assays, different alternatively spliced Nrxn2β isoforms robustly promoted synapse assembly similar to Nrxn1β and Nrxn3β, consistent with a general synaptogenic function of neurexins. Deletion of Nrxn2 from cultured hippocampal neurons, however, caused a significant increase in synapse density and release probability, replicating the in vivo data that suggested a synapse-restricting function. Rescue experiments revealed that two of the four Nrxn2β splice variants (Nrxn2β-SS4+/SS5- and Nrxn2β-SS4+/SS5+) reversed the increase in synapse density in Nrxn2-deficient neurons, whereas only one of the four Nrxn2β splice variants (Nrxn2β-SS4+/SS5+) normalized the increase in release probability in Nrxn2-deficient neurons. Thus, a subset of Nrxn2 splice variants restricts synapse numbers and restrains their release probability in cultured neurons.

The Memory Orchestra: Contribution of Astrocytes

For decades, memory research has centered on the role of neurons, which do not function in isolation. However, astrocytes play important roles in regulating neuronal recruitment and function at the local and network levels, forming the basis for information processing as well as memory formation and storage. In this review, we discuss the role of astrocytes in memory functions and their cellular underpinnings at multiple time points. We summarize important breakthroughs and controversies in the field as well as potential avenues to further illuminate the role of astrocytes in memory processes.

Dually innervated dendritic spines develop in the absence of excitatory activity and resist plasticity through tonic inhibitory crosstalk

Dendritic spines can be directly connected to both inhibitory and excitatory presynaptic terminals, resulting in nanometer-scale proximity of opposing synaptic functions. While dually innervated spines (DiSs) are observed throughout the central nervous system, their developmental timeline and functional properties remain uncharacterized. Here we used a combination of serial section electron microscopy, live imaging, and local synapse activity manipulations to investigate DiS development and function in rodent hippocampus. Dual innervation occurred early in development, even on spines where the excitatory input was locally silenced. Synaptic NMDA receptor currents were selectively reduced at DiSs through tonic GABAB receptor signaling. Accordingly, spine enlargement normally associated with long-term potentiation on singly innervated spines (SiSs) was blocked at DiSs. Silencing somatostatin interneurons or pharmacologically blocking GABABRs restored NMDA receptor function and structural plasticity to levels comparable to neighboring SiSs. Thus, hippocampal DiSs are stable structures where function and plasticity are potently regulated by nanometer-scale GABAergic signaling.

Coordinated drift of receptive fields in Hebbian/anti-Hebbian network models during noisy representation learning

Recent experiments have revealed that neural population codes in many brain areas continuously change even when animals have fully learned and stably perform their tasks. This representational 'drift' naturally leads to questions about its causes, dynamics and functions. Here we explore the hypothesis that neural representations optimize a representational objective with a degenerate solution space, and noisy synaptic updates drive the network to explore this (near-)optimal space causing representational drift. We illustrate this idea and explore its consequences in simple, biologically plausible Hebbian/anti-Hebbian network models of representation learning. We find that the drifting receptive fields of individual neurons can be characterized by a coordinated random walk, with effective diffusion constants depending on various parameters such as learning rate, noise amplitude and input statistics. Despite such drift, the representational similarity of population codes is stable over time. Our model recapitulates experimental observations in the hippocampus and posterior parietal cortex and makes testable predictions that can be probed in future experiments.

The molecular memory code and synaptic plasticity: A synthesis

The most widely accepted view of memory in the brain holds that synapses are the storage sites of memory, and that memories are formed through associative modification of synapses. This view has been challenged on conceptual and empirical grounds. As an alternative, it has been proposed that molecules within the cell body are the storage sites of memory, and that memories are formed through biochemical operations on these molecules. This paper proposes a synthesis of these two views, grounded in a computational model of memory. Synapses are conceived as storage sites for the parameters of an approximate posterior probability distribution over latent causes. Intracellular molecules are conceived as storage sites for the parameters of a generative model. The model stipulates how these two components work together as part of an integrated algorithm for learning and inference.

Arc-driven mGRASP highlights CA1 to CA3 synaptic engrams

Subpopulations of neurons display increased activity during memory encoding and manipulating the activity of these neurons can induce artificial formation or erasure of memories. Thus, these neurons are thought to be cellular engrams. Moreover, correlated activity between pre- and postsynaptic engram neurons is thought to lead to strengthening of their synaptic connections, thus increasing the probability of neural activity patterns occurring during encoding to reoccur at recall. Therefore, synapses between engram neurons can also be considered as a substrate of memory, or a synaptic engram. One can label synaptic engrams by targeting two complementary, non-fluorescent, synapse-targeted GFP fragments separately to the pre- and postsynaptic compartment of engram neurons; the two GFP fragments reconstitute a fluorescent GFP at the synaptic cleft between the engram neurons, thereby highlighting synaptic engrams. In this work we explored a transsynaptic GFP reconstitution system (mGRASP) to label synaptic engrams between hippocampal CA1 and CA3 engram neurons identified by different Immediate-Early Genes: cFos and Arc. We characterized the expression of the cellular and synaptic labels of the mGRASP system upon exposure to a novel environment or learning of a hippocampal-dependent memory task. We found that mGRASP under the control of transgenic ArcCre^ERT2 labeled synaptic engrams more efficiently than when controlled by viral cFostTA, possibly due to differences in the genetic systems rather than the specific IEG promoters.

A flexible two-photon fiberscope for fast activity imaging and precise optogenetic photostimulation of neurons in freely moving mice

We developed a flexible two-photon microendoscope (2P-FENDO) capable of all-optical brain investigation at near cellular resolution in freely moving mice. The system performs fast two-photon (2P) functional imaging and 2P holographic photostimulation of single and multiple cells using axially confined extended spots. Proof-of-principle experiments were performed in freely moving mice co-expressing jGCaMP7s and the opsin ChRmine in the visual or barrel cortex. On a field of view of 250 μm in diameter, we demonstrated functional imaging at a frame rate of up to 50 Hz and precise photostimulation of selected groups of cells. With the capability to simultaneously image and control defined neuronal networks in freely moving animals, 2P-FENDO will enable a precise investigation of neuronal functions in the brain during naturalistic behaviors.

Neurexin-2: An inhibitory neurexin that restricts excitatory synapse formation in the hippocampus

Neurexins are widely thought to promote synapse formation and to organize synapse properties. Here we found that in contrast to neurexin-1 and neurexin-3, neurexin-2 unexpectedly restricts synapse formation. In the hippocampus, constitutive or neuron-specific deletions of neurexin-2 nearly doubled the strength of excitatory CA3➔CA1 region synaptic connections and markedly increased their release probability. No effect on inhibitory synapses was detected. Stochastic optical reconstruction microscopy (STORM) superresolution microscopy revealed that the neuron-specific neurexin-2 deletion elevated the density of excitatory CA1 region synapses nearly twofold. Moreover, hippocampal neurexin-2 deletions also increased synaptic connectivity in the CA1 region when induced in mature mice and impaired the cognitive flexibility of spatial memory. Thus, neurexin-2 controls the dynamics of hippocampal synaptic circuits by repressing synapse assembly throughout life, a restrictive function that markedly differs from that of neurexin-1 and neurexin-3 and of other synaptic adhesion molecules, suggesting that neurexins evolutionarily diverged into opposing pro- and antisynaptogenic organizers.

Unraveling the mysteries of dendritic spine dynamics: Five key principles shaping memory and cognition

Recent research extends our understanding of brain processes beyond just action potentials and chemical transmissions within neural circuits, emphasizing the mechanical forces generated by excitatory synapses on dendritic spines to modulate presynaptic function. From in vivo and in vitro studies, we outline five central principles of synaptic mechanics in brain function: P1: Stability - Underpinning the integral relationship between the structure and function of the spine synapses. P2: Extrinsic dynamics - Highlighting synapse-selective structural plasticity which plays a crucial role in Hebbian associative learning, distinct from pathway-selective long-term potentiation (LTP) and depression (LTD). P3: Neuromodulation - Analyzing the role of G-protein-coupled receptors, particularly dopamine receptors, in time-sensitive modulation of associative learning frameworks such as Pavlovian classical conditioning and Thorndike's reinforcement learning (RL). P4: Instability - Addressing the intrinsic dynamics crucial to memory management during continual learning, spotlighting their role in "spine dysgenesis" associated with mental disorders. P5: Mechanics - Exploring how synaptic mechanics influence both sides of synapses to establish structural traces of short- and long-term memory, thereby aiding the integration of mental functions. We also delve into the historical background and foresee impending challenges.

Dendritic Spines in Learning and Memory: From First Discoveries to Current Insights

The central nervous system is composed of neural ensembles, and their activity patterns are neural correlates of cognitive functions. Those ensembles are networks of neurons connected to each other by synapses. Most neurons integrate synaptic signal through a remarkable subcellular structure called spine. Dendritic spines are protrusions whose diverse shapes make them appear as a specific neuronal compartment, and they have been the focus of studies for more than a century. Soon after their first description by Ramón y Cajal, it has been hypothesized that spine morphological changes could modify neuronal connectivity and sustain cognitive abilities. Later studies demonstrated that changes in spine density and morphology occurred in experience-dependent plasticity during development, and in clinical cases of mental retardation. This gave ground for the assumption that dendritic spines are the particular locus of cerebral plasticity. With the discovery of synaptic long-term potentiation, a research program emerged with the aim to establish whether dendritic spine plasticity could explain learning and memory. The development of live imaging methods revealed on the one hand that dendritic spine remodeling is compatible with learning process and, on the other hand, that their long-term stability is compatible with lifelong memories. Furthermore, the study of the mechanisms of spine growth and maintenance shed new light on the rules of plasticity. In behavioral paradigms of memory, spine formation or elimination and morphological changes were found to correlate with learning. In a last critical step, recent experiments have provided evidence that dendritic spines play a causal role in learning and memory.

Microglial motility is modulated by neuronal activity and correlates with dendritic spine plasticity in the hippocampus of awake mice

Microglia, the resident immune cells of the brain, play a complex role in health and disease. They actively survey the brain parenchyma by physically interacting with other cells and structurally shaping the brain. Yet, the mechanisms underlying microglial motility and significance for synapse stability, especially in the hippocampus during adulthood, remain widely unresolved. Here, we investigated the effect of neuronal activity on microglial motility and the implications for the formation and survival of dendritic spines on hippocampal CA1 neurons in vivo. We used repetitive two-photon in vivo imaging in the hippocampus of awake and anesthetized mice to simultaneously study the motility of microglia and their interaction with dendritic spines. We found that CA3 to CA1 input is sufficient to modulate microglial process motility. Simultaneously, more dendritic spines emerged in mice after awake compared to anesthetized imaging. Interestingly, the rate of microglial contacts with individual dendritic spines and dendrites was associated with the stability, removal, and emergence of dendritic spines. These results suggest that microglia might sense neuronal activity via neurotransmitter release and actively participate in synaptic rewiring of the hippocampal neural network during adulthood. Further, this study has profound relevance for hippocampal learning and memory processes.

The geometry of representational drift in natural and artificial neural networks

Neurons in sensory areas encode/represent stimuli. Surprisingly, recent studies have suggested that, even during persistent performance, these representations are not stable and change over the course of days and weeks. We examine stimulus representations from fluorescence recordings across hundreds of neurons in the visual cortex using in vivo two-photon calcium imaging and we corroborate previous studies finding that such representations change as experimental trials are repeated across days. This phenomenon has been termed "representational drift". In this study we geometrically characterize the properties of representational drift in the primary visual cortex of mice in two open datasets from the Allen Institute and propose a potential mechanism behind such drift. We observe representational drift both for passively presented stimuli, as well as for stimuli which are behaviorally relevant. Across experiments, the drift differs from in-session variance and most often occurs along directions that have the most in-class variance, leading to a significant turnover in the neurons used for a given representation. Interestingly, despite this significant change due to drift, linear classifiers trained to distinguish neuronal representations show little to no degradation in performance across days. The features we observe in the neural data are similar to properties of artificial neural networks where representations are updated by continual learning in the presence of dropout, i.e. a random masking of nodes/weights, but not other types of noise. Therefore, we conclude that a potential reason for the representational drift in biological networks is driven by an underlying dropout-like noise while continuously learning and that such a mechanism may be computational advantageous for the brain in the same way it is for artificial neural networks, e.g. preventing overfitting.

Large-volume and deep brain imaging in rabbits and monkeys using COMPACT two-photon microscopy

In vivo imaging has been widely used for investigating the structure and function of neurons typically located within ~ 800 μm below the cortical surface. Due to light scattering and absorption, it has been difficult to perform in-vivo imaging of neurons in deep cortical and subcortical regions of large animals with two-photon microscopy. Here, we combined a thin-wall quartz capillary with a GRIN lens attached to a prism for large-volume structural and calcium imaging of neurons located 2 mm below the surface of rabbit and monkey brains. The field of view was greatly expanded by rotating and changing the depth of the imaging probe inside a quartz capillary. Calcium imaging of layer 5/6 neurons in the rabbit motor cortex revealed differential activity of these neurons between quiet wakefulness and slow wave sleep. The method described here provides an important tool for studying the structure and function of neurons located deep in the brains of large animals.

Repeated stress exposure leads to structural synaptic instability prior to disorganization of hippocampal coding and impairments in learning

Stress exposure impairs brain structure and function, resulting in cognitive deficits and increased risk for psychiatric disorders such as depression, schizophrenia, anxiety and post-traumatic stress disorder. In particular, stress exposure affects function and structure of hippocampal CA1 leading to impairments in episodic memory. Here, we applied longitudinal deep-brain optical imaging to investigate the link between changes in activity patterns and structural plasticity of dorsal CA1 pyramidal neurons and hippocampal-dependent learning and memory in mice exposed to stress. We found that several days of repeated stress led to a substantial increase in neuronal activity followed by disruption of the temporal structure of this activity and spatial coding. We then tracked dynamics of structural excitatory connectivity as a potential underlying cause of the changes in activity induced by repeated stress. We thus discovered that exposure to repeated stress leads to an immediate decrease in spinogenesis followed by decrease in spine stability. By comparison, acute stress led to stabilization of the spines born in temporal proximity to the stressful event. Importantly, the temporal relationship between changes in activity levels, structural connectivity and activity patterns, suggests that loss of structural connectivity mediates the transition between increased activity and impairment of temporal organization and spatial information content in dorsal CA1 upon repeated stress exposure.

CaMKII: a central molecular organizer of synaptic plasticity, learning and memory

Calcium-calmodulin (CaM)-dependent protein kinase II (CaMKII) is the most abundant protein in excitatory synapses and is central to synaptic plasticity, learning and memory. It is activated by intracellular increases in calcium ion levels and triggers molecular processes necessary for synaptic plasticity. CaMKII phosphorylates numerous synaptic proteins, thereby regulating their structure and functions. This leads to molecular events crucial for synaptic plasticity, such as receptor trafficking, localization and activity; actin cytoskeletal dynamics; translation; and even transcription through synapse-nucleus shuttling. Several new tools affording increasingly greater spatiotemporal resolution have revealed the link between CaMKII activity and downstream signalling processes in dendritic spines during synaptic and behavioural plasticity. These technologies have provided insights into the function of CaMKII in learning and memory.

Representational drift: Emerging theories for continual learning and experimental future directions

Recent work has revealed that the neural activity patterns correlated with sensation, cognition, and action often are not stable and instead undergo large scale changes over days and weeks-a phenomenon called representational drift. Here, we highlight recent observations of drift, how drift is unlikely to be explained by experimental confounds, and how the brain can likely compensate for drift to allow stable computation. We propose that drift might have important roles in neural computation to allow continual learning, both for separating and relating memories that occur at distinct times. Finally, we present an outlook on future experimental directions that are needed to further characterize drift and to test emerging theories for drift's role in computation.