Learning drives differential clustering of axodendritic contacts in the barn owl auditory system

Clustering of synaptic engram: Functional and structural basis of memory

Studies on memory engram have demonstrated how experience and learning can be allocated at a neuronal level for centuries. Recently emerging evidence narrowed down further to the synaptic connections and their patterned allocation on dendrites. Notably, groups of synapses within a specific range within dendrites known as 'synaptic clusters' have been revealed in association with learning and memory. Previous investigations have shown that a variety of factors mediated by both presynaptic inputs and postsynaptic dendrites contribute to clustering. Here, we review the neural mechanism of synaptic clustering and its correlation with memory. We highlight the recent findings about the clustering of synaptic engrams and memory formation and discuss future directions.

Clustered synapses develop in distinct dendritic domains in visual cortex before eye opening

Synaptic inputs to cortical neurons are highly structured in adult sensory systems, such that neighboring synapses along dendrites are activated by similar stimuli. This organization of synaptic inputs, called synaptic clustering, is required for high-fidelity signal processing, and clustered synapses can already be observed before eye opening. However, how clustered inputs emerge during development is unknown. Here, we employed concurrent in vivo whole-cell patch-clamp and dendritic calcium imaging to map spontaneous synaptic inputs to dendrites of layer 2/3 neurons in the mouse primary visual cortex during the second postnatal week until eye opening. We found that the number of functional synapses and the frequency of transmission events increase several fold during this developmental period. At the beginning of the second postnatal week, synapses assemble specifically in confined dendritic segments, whereas other segments are devoid of synapses. By the end of the second postnatal week, just before eye opening, dendrites are almost entirely covered by domains of co-active synapses. Finally, co-activity with their neighbor synapses correlates with synaptic stabilization and potentiation. Thus, clustered synapses form in distinct functional domains presumably to equip dendrites with computational modules for high-capacity sensory processing when the eyes open.

Engram mechanisms of memory linking and identity

Memories are thought to be stored in neuronal ensembles referred to as engrams. Studies have suggested that when two memories occur in quick succession, a proportion of their engrams overlap and the memories become linked (in a process known as prospective linking) while maintaining their individual identities. In this Review, we summarize the key principles of memory linking through engram overlap, as revealed by experimental and modelling studies. We describe evidence of the involvement of synaptic memory substrates, spine clustering and non-linear neuronal capacities in prospective linking, and suggest a dynamic somato-synaptic model, in which memories are shared between neurons yet remain separable through distinct dendritic and synaptic allocation patterns. We also bring into focus retrospective linking, in which memories become associated after encoding via offline reactivation, and discuss key temporal and mechanistic differences between prospective and retrospective linking, as well as the potential differences in their cognitive outcomes.

VAP spatially stabilizes dendritic mitochondria to locally support synaptic plasticity

Synapses are pivotal sites of plasticity and memory formation. Consequently, synapses are energy consumption hotspots susceptible to dysfunction when their energy supplies are perturbed. Mitochondria are stabilized near synapses via the cytoskeleton and provide the local energy required for synaptic plasticity. However, the mechanisms that tether and stabilize mitochondria to support synaptic plasticity are unknown. We identified proteins exclusively tethering mitochondria to actin near postsynaptic spines. We find that VAP, the vesicle-associated membrane protein-associated protein implicated in amyotrophic lateral sclerosis, stabilizes mitochondria via actin near the spines. To test if the VAP-dependent stable mitochondrial compartments can locally support synaptic plasticity, we used two-photon glutamate uncaging for spine plasticity induction and investigated the induced and adjacent uninduced spines. We find VAP functions as a spatial stabilizer of mitochondrial compartments for up to ~60 min and as a spatial ruler determining the ~30 μm dendritic segment supported during synaptic plasticity.

Generalized extinction of fear memory depends on co-allocation of synaptic plasticity in dendrites

Memories can be modified by new experience in a specific or generalized manner. Changes in synaptic connections are crucial for memory storage, but it remains unknown how synaptic changes associated with different memories are distributed within neuronal circuits and how such distributions affect specific or generalized modification by novel experience. Here we show that fear conditioning with two different auditory stimuli (CS) and footshocks (US) induces dendritic spine elimination mainly on different dendritic branches of layer 5 pyramidal neurons in the mouse motor cortex. Subsequent fear extinction causes CS-specific spine formation and extinction of freezing behavior. In contrast, spine elimination induced by fear conditioning with >2 different CS-USs often co-exists on the same dendritic branches. Fear extinction induces CS-nonspecific spine formation and generalized fear extinction. Moreover, activation of somatostatin-expressing interneurons increases the occurrence of spine elimination induced by different CS-USs on the same dendritic branches and facilitates the generalization of fear extinction. These findings suggest that specific or generalized modification of existing memories by new experience depends on whether synaptic changes induced by previous experiences are segregated or co-exist at the level of individual dendritic branches.

Emergence of synaptic organization and computation in dendrites

Single neurons in the brain exhibit astounding computational capabilities, which gradually emerge throughout development and enable them to become integrated into complex neural circuits. These capabilities derive in part from the precise arrangement of synaptic inputs on the neurons' dendrites. While the full computational benefits of this arrangement are still unknown, a picture emerges in which synapses organize according to their functional properties across multiple spatial scales. In particular, on the local scale (tens of microns), excitatory synaptic inputs tend to form clusters according to their functional similarity, whereas on the scale of individual dendrites or the entire tree, synaptic inputs exhibit dendritic maps where excitatory synapse function varies smoothly with location on the tree. The development of this organization is supported by inhibitory synapses, which are carefully interleaved with excitatory synapses and can flexibly modulate activity and plasticity of excitatory synapses. Here, we summarize recent experimental and theoretical research on the developmental emergence of this synaptic organization and its impact on neural computations.

Developmental emergence of two-stage nonlinear synaptic integration in cerebellar interneurons

Synaptic transmission, connectivity, and dendritic morphology mature in parallel during brain development and are often disrupted in neurodevelopmental disorders. Yet how these changes influence the neuronal computations necessary for normal brain function are not well understood. To identify cellular mechanisms underlying the maturation of synaptic integration in interneurons, we combined patch-clamp recordings of excitatory inputs in mouse cerebellar stellate cells (SCs), three-dimensional reconstruction of SC morphology with excitatory synapse location, and biophysical modeling. We found that postnatal maturation of postsynaptic strength was homogeneously reduced along the somatodendritic axis, but dendritic integration was always sublinear. However, dendritic branching increased without changes in synapse density, leading to a substantial gain in distal inputs. Thus, changes in synapse distribution, rather than dendrite cable properties, are the dominant mechanism underlying the maturation of neuronal computation. These mechanisms favor the emergence of a spatially compartmentalized two-stage integration model promoting location-dependent integration within dendritic subunits.

Synapses from the Motor Cortex and a High-Order Thalamic Nucleus are Spatially Clustered in Proximity to Each Other in the Distal Tuft Dendrites of Mouse Somatosensory Cortex

The posterior medial nucleus of the thalamus (POm) and vibrissal primary motor cortex (vM1) convey essential information to the barrel cortex (S1BF) regarding whisker position and movement. Therefore, understanding the relative spatial relationship of these two inputs is a critical prerequisite for acquiring insights into how S1BF synthesizes information to interpret the location of an object. Using array tomography, we identified the locations of synapses from vM1 and POm on distal tuft dendrites of L5 pyramidal neurons where the two inputs are combined. Synapses from vM1 and POm did not show a significant branchlet preference and impinged on the same set of dendritic branchlets. Within dendritic branches, on the other hand, the two inputs formed robust spatial clusters of their own type. Furthermore, we also observed POm clusters in proximity to vM1 clusters. This work constitutes the first detailed description of the relative distribution of synapses from POm and vM1, which is crucial to elucidate the synaptic integration of whisker-based sensory information.

Transcriptional Control of Parallel-Acting Pathways That Remove Specific Presynaptic Proteins in Remodeling Neurons

Synapses are actively dismantled to mediate circuit refinement, but the developmental pathways that regulate synaptic disassembly are largely unknown. We have previously shown that the epithelial sodium channel ENaC/UNC-8 triggers an activity-dependent mechanism that drives the removal of presynaptic proteins liprin-α/SYD-2, Synaptobrevin/SNB-1, RAB-3, and Endophilin/UNC-57 in remodeling GABAergic neurons in Caenorhabditis elegans (Miller-Fleming et al., 2016). Here, we report that the conserved transcription factor Iroquois/IRX-1 regulates UNC-8 expression as well as an additional pathway, independent of UNC-8, that functions in parallel to dismantle functional presynaptic terminals. We show that the additional IRX-1-regulated pathway is selectively required for the removal of the presynaptic proteins, Munc13/UNC-13 and ELKS, which normally mediate synaptic vesicle (SV) fusion and neurotransmitter release. Our findings are notable because they highlight the key role of transcriptional regulation in synapse elimination during development and reveal parallel-acting pathways that coordinate synaptic disassembly by removing specific active zone proteins.SIGNIFICANCE STATEMENT Synaptic pruning is a conserved feature of developing neural circuits but the mechanisms that dismantle the presynaptic apparatus are largely unknown. We have determined that synaptic disassembly is orchestrated by parallel-acting mechanisms that target distinct components of the active zone. Thus, our finding suggests that synaptic disassembly is not accomplished by en masse destruction but depends on mechanisms that dismantle the structure in an organized process.

Emergence of local and global synaptic organization on cortical dendrites

Synaptic inputs on cortical dendrites are organized with remarkable subcellular precision at the micron level. This organization emerges during early postnatal development through patterned spontaneous activity and manifests both locally where nearby synapses are significantly correlated, and globally with distance to the soma. We propose a biophysically motivated synaptic plasticity model to dissect the mechanistic origins of this organization during development and elucidate synaptic clustering of different stimulus features in the adult. Our model captures local clustering of orientation in ferret and receptive field overlap in mouse visual cortex based on the receptive field diameter and the cortical magnification of visual space. Including action potential back-propagation explains branch clustering heterogeneity in the ferret and produces a global retinotopy gradient from soma to dendrite in the mouse. Therefore, by combining activity-dependent synaptic competition and species-specific receptive fields, our framework explains different aspects of synaptic organization regarding stimulus features and spatial scales.

Dendritic Spine Density and Dynamics of Layer 5 Pyramidal Neurons of the Primary Motor Cortex Are Elevated With Aging

It is well established that motor impairment often occurs alongside healthy aging, leading to problems with fine motor skills and coordination. Although previously thought to be caused by neuronal death accumulating across the lifespan, it is now believed that the source of this impairment instead stems from more subtle changes in neural connectivity. The dendritic spine is a prime target for exploration of this problem because it is the postsynaptic partner of most excitatory synapses received by the pyramidal neuron, a cortical cell that carries much of the information processing load in the cerebral cortex. We repeatedly imaged the same dendrites in young adult and aged mouse motor cortex over the course of 1 month to look for differences in the baseline state of the dendritic spine population. These experiments reveal increased dendritic spine density, without obvious changes in spine clustering, occurring at the aged dendrite. Additionally, aged dendrites exhibit elevated spine turnover and stabilization alongside decreased long-term spine survival. These results suggest that at baseline the aged motor cortex may exist in a perpetual state of relative instability and attempts at compensation. This phenotype of aging may provide clues for future targets of aging-related motor impairment remediation.

The gradient clusteron: A model neuron that learns to solve classification tasks via dendritic nonlinearities, structural plasticity, and gradient descent

Synaptic clustering on neuronal dendrites has been hypothesized to play an important role in implementing pattern recognition. Neighboring synapses on a dendritic branch can interact in a synergistic, cooperative manner via nonlinear voltage-dependent mechanisms, such as NMDA receptors. Inspired by the NMDA receptor, the single-branch clusteron learning algorithm takes advantage of location-dependent multiplicative nonlinearities to solve classification tasks by randomly shuffling the locations of "under-performing" synapses on a model dendrite during learning ("structural plasticity"), eventually resulting in synapses with correlated activity being placed next to each other on the dendrite. We propose an alternative model, the gradient clusteron, or G-clusteron, which uses an analytically-derived gradient descent rule where synapses are "attracted to" or "repelled from" each other in an input- and location-dependent manner. We demonstrate the classification ability of this algorithm by testing it on the MNIST handwritten digit dataset and show that, when using a softmax activation function, the accuracy of the G-clusteron on the all-versus-all MNIST task (~85%) approaches that of logistic regression (~93%). In addition to the location update rule, we also derive a learning rule for the synaptic weights of the G-clusteron ("functional plasticity") and show that a G-clusteron that utilizes the weight update rule can achieve ~89% accuracy on the MNIST task. We also show that a G-clusteron with both the weight and location update rules can learn to solve the XOR problem from arbitrary initial conditions.

Heterosynaptic cross-talk of pre- and postsynaptic strengths along segments of dendrites

Dendrites are crucial for integrating incoming synaptic information. Individual dendritic branches are thought to constitute a signal processing unit, yet how neighboring synapses shape the boundaries of functional dendritic units is not well understood. Here, we address the cellular basis underlying the organization of the strengths of neighboring Schaffer collateral-CA1 synapses by optical quantal analysis and spine size measurements. Inducing potentiation at clusters of spines produces NMDA-receptor-dependent heterosynaptic plasticity. The direction of postsynaptic strength change shows distance dependency to the stimulated synapses where proximal synapses predominantly depress, whereas distal synapses potentiate; potentiation and depression are regulated by CaMKII and calcineurin, respectively. In contrast, heterosynaptic presynaptic plasticity is confined to weakening of presynaptic strength of nearby synapses, which requires CaMKII and the retrograde messenger nitric oxide. Our findings highlight the parallel engagement of multiple signaling pathways, each with characteristic spatial dynamics in shaping the local pattern of synaptic strengths.

Synaptic remodeling, lessons from C. elegans

Sydney Brenner's choice of Caenorhabditis elegans as a model organism for understanding the nervous system has accelerated discoveries of gene function in neural circuit development and behavior. In this review, we discuss a striking example of synaptic remodeling in the C. elegans motor circuit in which DD class motor neurons effectively reverse polarity as presynaptic and postsynaptic domains at opposite ends of the DD neurite switch locations. Originally revealed by EM reconstruction conducted over 40 years ago, DD remodeling has since been investigated by live cell imaging methods that exploit the power of C. elegans genetics to reveal key effectors of synaptic plasticity. Although synapses are also extensively rewired in developing mammalian circuits, the underlying remodeling mechanisms are largely unknown. Here, we highlight the possibility that studies in C. elegans can reveal pathways that orchestrate synaptic remodeling in more complex organisms. Specifically, we describe (1) transcription factors that regulate DD remodeling, (2) the cellular and molecular cascades that drive synaptic remodeling and (3) examples of circuit modifications in vertebrate neurons that share some similarities with synaptic remodeling in C. elegans DD neurons.

Illuminating dendritic function with computational models

Dendrites have always fascinated researchers: from the artistic drawings by Ramon y Cajal to the beautiful recordings of today, neuroscientists have been striving to unravel the mysteries of these structures. Theoretical work in the 1960s predicted important dendritic effects on neuronal processing, establishing computational modelling as a powerful technique for their investigation. Since then, modelling of dendrites has been instrumental in driving neuroscience research in a targeted manner, providing experimentally testable predictions that range from the subcellular level to the systems level, and their relevance extends to fields beyond neuroscience, such as machine learning and artificial intelligence. Validation of modelling predictions often requires - and drives - new technological advances, thus closing the loop with theory-driven experimentation that moves the field forward. This Review features the most important, to our understanding, contributions of modelling of dendritic computations, including those pending experimental verification, and highlights studies of successful interactions between the modelling and experimental neuroscience communities.

Astrocytic Ephrin-B1 Controls Synapse Formation in the Hippocampus During Learning and Memory

Astrocytes play a fundamental role in synapse formation, pruning, and plasticity, which are associated with learning and memory. However, the role of astrocytes in learning and memory is still largely unknown. Our previous study showed that astrocyte-specific ephrin-B1 knock-out (KO) enhanced but ephrin-B1 overexpression (OE) in hippocampal astrocytes impaired contextual memory recall following fear conditioning. The goal of this study was to understand the mechanism by which astrocytic ephrin-B1 influences learning; specifically, learning-induced remodeling of synapses and dendritic spines in CA1 hippocampus using fear-conditioning paradigm. While we found a higher dendritic spine density and clustering on c-Fos-positive (+) neurons activated during contextual memory recall in both wild-type (WT) and KO mice, overall spine density and mEPSC amplitude were increased in CA1 neurons of KO compared to WT. In contrast, ephrin-B1 OE in hippocampal astrocytes impaired dendritic spine formation and clustering, specifically on c-Fos(+) neurons, coinciding with an overall decrease in vGlut1/PSD95 co-localization. Although astrocytic ephrin-B1 influenced learning-induced spine formation, the changes in astrocytic ephrin-B1 levels did not affect spine enlargement as no genotype differences in spine volume were observed between trained WT, KO, and OE groups. Our results suggest that a reduced formation of new spines rather than spine maturation in activated CA1 hippocampal neurons is most likely responsible for impaired contextual learning in OE mice due to abundantly high ephrin-B1 levels in astrocytes. The ability of astrocytic ephrin-B1 to negatively influence new spine formation during learning can potentially regulate new synapse formation at specific dendritic domains and underlie memory encoding.

Toric Spines at a Site of Learning

We discovered a new type of dendritic spine. It is found on space-specific neurons in the barn owl inferior colliculus, a site of experience-dependent plasticity. Connectomic analysis revealed dendritic protrusions of unusual morphology including topological holes, hence termed "toric" spines (n = 76). More significantly, presynaptic terminals converging onto individual toric spines displayed numerous active zones (up to 49) derived from multiple axons (up to 11) with incoming trajectories distributed widely throughout 3D space. This arrangement is suited to integrate input sources. Dense reconstruction of two toric spines revealed that they were unconnected with the majority (∼84%) of intertwined axons, implying a high capacity for information storage. We developed an ex vivo slice preparation and provide the first published data on space-specific neuron intrinsic properties, including cellular subtypes with and without toric-like spines. We propose that toric spines are a cellular locus of sensory integration and behavioral learning.

Synaptic Clustering and Memory Formation

In the study of memory engrams, synaptic memory allocation is a newly emerged theme that focuses on how specific synapses are engaged in the storage of a given memory. Cumulating evidence from imaging and molecular experiments indicates that the recruitment of synapses that participate in the encoding and expression of memory is neither random nor uniform. A hallmark observation is the emergence of groups of synapses that share similar response properties and/or similar input properties and are located within a stretch of a dendritic branch. This grouping of synapses has been termed "synapse clustering" and has been shown to emerge in many different memory-related paradigms, as well as in in vitro studies. The clustering of synapses may emerge from synapses receiving similar input, or via many processes which allow for cross-talk between nearby synapses within a dendritic branch, leading to cooperative plasticity. Clustered synapses can act in concert to maximally exploit the nonlinear integration potential of the dendritic branches in which they reside. Their main contribution is to facilitate the induction of dendritic spikes and dendritic plateau potentials, which provide advanced computational and memory-related capabilities to dendrites and single neurons. This review focuses on recent evidence which investigates the role of synapse clustering in dendritic integration, sensory perception, learning, and memory as well as brain dysfunction. We also discuss recent theoretical work which explores the computational advantages provided by synapse clustering, leading to novel and revised theories of memory. As an eminent phenomenon during memory allocation, synapse clustering both shapes memory engrams and is also shaped by the parallel plasticity mechanisms upon which it relies.

Sustained rescue of prefrontal circuit dysfunction by antidepressant-induced spine formation

The neurobiological mechanisms underlying the induction and remission of depressive episodes over time are not well understood. Through repeated longitudinal imaging of medial prefrontal microcircuits in the living brain, we found that prefrontal spinogenesis plays a critical role in sustaining specific antidepressant behavioral effects and maintaining long-term behavioral remission. Depression-related behavior was associated with targeted, branch-specific elimination of postsynaptic dendritic spines on prefrontal projection neurons. Antidepressant-dose ketamine reversed these effects by selectively rescuing eliminated spines and restoring coordinated activity in multicellular ensembles that predict motivated escape behavior. Prefrontal spinogenesis was required for the long-term maintenance of antidepressant effects on motivated escape behavior but not for their initial induction.

How Dendrites Affect Online Recognition Memory

In order to record the stream of autobiographical information that defines our unique personal history, our brains must form durable memories from single brief exposures to the patterned stimuli that impinge on them continuously throughout life. However, little is known about the computational strategies or neural mechanisms that underlie the brain's ability to perform this type of "online" learning. Based on increasing evidence that dendrites act as both signaling and learning units in the brain, we developed an analytical model that relates online recognition memory capacity to roughly a dozen dendritic, network, pattern, and task-related parameters. We used the model to determine what dendrite size maximizes storage capacity under varying assumptions about pattern density and noise level. We show that over a several-fold range of both of these parameters, and over multiple orders-of-magnitude of memory size, capacity is maximized when dendrites contain a few hundred synapses-roughly the natural number found in memory-related areas of the brain. Thus, in comparison to entire neurons, dendrites increase storage capacity by providing a larger number of better-sized learning units. Our model provides the first normative theory that explains how dendrites increase the brain's capacity for online learning; predicts which combinations of parameter settings we should expect to find in the brain under normal operating conditions; leads to novel interpretations of an array of existing experimental results; and provides a tool for understanding which changes associated with neurological disorders, aging, or stress are most likely to produce memory deficits-knowledge that could eventually help in the design of improved clinical treatments for memory loss.

Synaptic topography - Converging connections and emerging function

Brain circuits are constituted of individual neurons that are interconnected with a vast array of synapses. In order to understand how brain function emerges from this complex synaptic network, immense efforts have been made to trace the synaptic topography, i.e. arrangement of synaptic connections, of the network. In addition to anatomically elaborating the synaptic layout at multiple levels across brain regions, recent studies have attempted to elucidate the fundamental wiring principles that govern neural information processing in the brain, establishing a link between anatomy and function. In this review, I will discuss recent discoveries on the topographical organization of synaptic connections at the cell-to-cell and subcellular levels in the cortex and hippocampus. Accumulating evidence leads us to acknowledge the highly structured, non-random synaptic connectivity that emerges together with sensory feature preferences of neurons and synchronous neuronal activity.

Quantitative Analysis of the Spatial Organization of Synaptic Inputs on the Postsynaptic Dendrite

The spatial organization of synaptic inputs on the dendritic tree of cortical neurons is considered to play an important role in the dendritic integration of synaptic activity. Active electrical properties of dendrites and mechanisms of dendritic integration have been studied for a long time. New technological developments are now enabling the characterization of the spatial organization of synaptic inputs on dendrites. However, quantitative methods for the analysis of such data are lacking. In order to place cluster parameters into the framework of dendritic integration and synaptic summation, these parameters need to be assessed rigorously in a quantitative manner. Here I present an approach for the analysis of synaptic input clusters on the dendritic tree that is based on combinatorial analysis of the likelihoods to observe specific input arrangements. This approach is superior to the commonly applied analysis of nearest neighbor distances between synaptic inputs comparing their distribution to simulations with random reshuffling or bootstrapping. First, the new approach yields exact likelihood values rather than approximate numbers obtained from simulations. Second and more importantly, the new approach identifies individual clusters and thereby allows to quantify and characterize individual cluster properties.

Hotspots of dendritic spine turnover facilitate clustered spine addition and learning and memory

Modeling studies suggest that clustered structural plasticity of dendritic spines is an efficient mechanism of information storage in cortical circuits. However, why new clustered spines occur in specific locations and how their formation relates to learning and memory (L&M) remain unclear. Using in vivo two-photon microscopy, we track spine dynamics in retrosplenial cortex before, during, and after two forms of episodic-like learning and find that spine turnover before learning predicts future L&M performance, as well as the localization and rates of spine clustering. Consistent with the idea that these measures are causally related, a genetic manipulation that enhances spine turnover also enhances both L&M and spine clustering. Biophysically inspired modeling suggests turnover increases clustering, network sparsity, and memory capacity. These results support a hotspot model where spine turnover is the driver for localization of clustered spine formation, which serves to modulate network function, thus influencing storage capacity and L&M.

Structural remodeling of dendritic spines is thought to be a mechanism of memory storage. Here, the authors look at how spine turnover and clustering predict future learning and memory performance, and see that a genetically modified mouse with enhanced spine turnover has enhanced learning.

Modulating STDP Balance Impacts the Dendritic Mosaic

The ability for cortical neurons to adapt their input/output characteristics and information processing capabilities ultimately relies on the interplay between synaptic plasticity, synapse location, and the nonlinear properties of the dendrite. Collectively, they shape both the strengths and spatial arrangements of convergent afferent inputs to neuronal dendrites. Recent experimental and theoretical studies support a clustered plasticity model, a view that synaptic plasticity promotes the formation of clusters or hotspots of synapses sharing similar properties. We have previously shown that spike timing-dependent plasticity (STDP) can lead to synaptic efficacies being arranged into spatially segregated clusters. This effectively partitions the dendritic tree into a tessellated imprint which we have called a dendritic mosaic. Here, using a biophysically detailed neuron model of a reconstructed layer 2/3 pyramidal cell and STDP learning, we investigated the impact of altered STDP balance on forming such a spatial organization. We show that cluster formation and extend depend on several factors, including the balance between potentiation and depression, the afferents' mean firing rate and crucially on the dendritic morphology. We find that STDP balance has an important role to play for this emergent mode of spatial organization since any imbalances lead to severe degradation- and in some case even destruction- of the mosaic. Our model suggests that, over a broad range of of STDP parameters, synaptic plasticity shapes the spatial arrangement of synapses, favoring the formation of clustered efficacy engrams.

Local translation in neuronal compartments: how local is local?

Efficient neuronal function depends on the continued modulation of the local neuronal proteome. Local protein synthesis plays a central role in tuning the neuronal proteome at specific neuronal regions. Various aspects of translation such as the localization of translational machinery, spatial spread of the newly translated proteins, and their site of action are carried out in specialized neuronal subcompartments to result in a localized functional outcome. In this review, we focus on the various aspects of these local translation compartments such as size, biochemical and organelle composition, structural boundaries, and temporal dynamics. We also discuss the apparent absence of definitive components of translation in these local compartments and the emerging state-of-the-art tools that could help dissecting these conundrums in greater detail in the future.

Linking Memories across Time via Neuronal and Dendritic Overlaps in Model Neurons with Active Dendrites

Memories are believed to be stored in distributed neuronal assemblies through activity-induced changes in synaptic and intrinsic properties. However, the specific mechanisms by which different memories become associated or linked remain a mystery. Here, we develop a simplified, biophysically inspired network model that incorporates multiple plasticity processes and explains linking of information at three different levels: (1) learning of a single associative memory, (2) rescuing of a weak memory when paired with a strong one, and (3) linking of multiple memories across time. By dissecting synaptic from intrinsic plasticity and neuron-wide from dendritically restricted protein capture, the model reveals a simple, unifying principle: linked memories share synaptic clusters within the dendrites of overlapping populations of neurons. The model generates numerous experimentally testable predictions regarding the cellular and sub-cellular properties of memory engrams as well as their spatiotemporal interactions.

Talking to the neighbours: The molecular and physiological mechanisms of clustered synaptic plasticity

Synaptic connectivity forms the basis for neuronal communication and the storage of information. Experiences and learning of new abilities can drive remodelling of this connectivity and promotes the formation of spine clusters; dendritic segments with a higher spine density. Spines located within these segments are frequently co-activated, undergo different dynamics than synapses located outside of this dendritic compartment and have, in general, a longer lifetime. Several lines of evidence have shown that chemical synapses located close to each other share or compete for intracellular signalling molecules and structural resources. This sharing and competition directly influences spine dynamics. Spines can grow, shrink, increase or decrease the surface expression of receptors, channels and adhesion molecules or remain stable and unchanged over extended periods of time. Here we summarize recent discoveries and provide a closer look at spine clustering, dendritic segment-specific signalling and potential molecular mechanisms underlying associative and heterosynaptic plasticity.

Is realistic neuronal modeling realistic?

Scientific models are abstractions that aim to explain natural phenomena. A successful model shows how a complex phenomenon arises from relatively simple principles while preserving major physical or biological rules and predicting novel experiments. A model should not be a facsimile of reality; it is an aid for understanding it. Contrary to this basic premise, with the 21st century has come a surge in computational efforts to model biological processes in great detail. Here we discuss the oxymoronic, realistic modeling of single neurons. This rapidly advancing field is driven by the discovery that some neurons don't merely sum their inputs and fire if the sum exceeds some threshold. Thus researchers have asked what are the computational abilities of single neurons and attempted to give answers using realistic models. We briefly review the state of the art of compartmental modeling highlighting recent progress and intrinsic flaws. We then attempt to address two fundamental questions. Practically, can we realistically model single neurons? Philosophically, should we realistically model single neurons? We use layer 5 neocortical pyramidal neurons as a test case to examine these issues. We subject three publically available models of layer 5 pyramidal neurons to three simple computational challenges. Based on their performance and a partial survey of published models, we conclude that current compartmental models are ad hoc, unrealistic models functioning poorly once they are stretched beyond the specific problems for which they were designed. We then attempt to plot possible paths for generating realistic single neuron models.

Clusters of synaptic inputs on dendrites of layer 5 pyramidal cells in mouse visual cortex

The spatial organization of synaptic inputs on the dendritic tree of cortical neurons plays a major role for dendritic integration and neural computations, yet, remarkably little is known about it. We mapped the spatial organization of glutamatergic synapses between layer 5 pyramidal cells by combining optogenetics and 2-photon calcium imaging in mouse neocortical slices. To mathematically characterize the organization of inputs we developed an approach based on combinatorial analysis of the likelihoods of specific synapse arrangements. We found that the synapses of intralaminar inputs form clusters on the basal dendrites of layer 5 pyramidal cells. These clusters contain 4 to 14 synapses within ≤30 µm of dendrite. According to the spatiotemporal characteristics of synaptic summation, these numbers suggest that there will be non-linear dendritic integration of synaptic inputs during synchronous activation.

DOI:http://dx.doi.org/10.7554/eLife.09222.001

Neurons in the brain exchange information through points of contact called synapses. If electrical activity arriving at a number of synapses exceeds a certain threshold, it can trigger an electrical impulse, which is transmitted to the next neuron. Synapses generally connect with branch-like structures called dendrites on the receiving neuron. However, little is known about how synapses are arranged on dendrites.

Gökçe et al. have now used a technique called optogenetics to work out the exact arrangement of a type of synapse on neurons in a part of the mouse brain that is devoted to vision. Optogenetics takes advantage of light-activated proteins that can trigger electrical activity. Gökçe et al. used mice that had been genetically engineered to produce these proteins in specific neurons, and then deliberately triggered electrical activity simply by shining light on these neurons. The experiments also used another technique called two-photon calcium imaging to monitor the activity of single synapses in response to the electrical activity triggered by optogenetics.

Gökçe et al. found that these neurons have clusters of four to fourteen synapses within a space of 30 micrometers along a dendrite. Synapses in clusters that are active at the same time can interact and thereby generate electrical signals more effectively than synapses spread across the dendrites. Further experiments are now needed to map the synapses between other kinds of neurons, and to map synapses from two different inputs at the same time.

DOI:http://dx.doi.org/10.7554/eLife.09222.002

Correlated Synaptic Inputs Drive Dendritic Calcium Amplification and Cooperative Plasticity during Clustered Synapse Development

The mechanisms that instruct the assembly of fine-scale features of synaptic connectivity in neural circuits are only beginning to be understood. Using whole-cell electrophysiology, two-photon calcium imaging, and glutamate uncaging in hippocampal slices, we discovered a functional coupling between NMDA receptor activation and ryanodine-sensitive intracellular calcium release that dominates the spatiotemporal dynamics of activity-dependent calcium signals during synaptogenesis. This developmentally regulated calcium amplification mechanism was tuned to detect and bind spatially clustered and temporally correlated synaptic inputs and enacted a local cooperative plasticity rule between coactive neighboring synapses. Consistent with the hypothesis that synapse maturation is spatially regulated, we observed clustering of synaptic weights in developing dendritic arbors. These results reveal developmental features of NMDA receptor-dependent calcium dynamics and local plasticity rules that are suited to spatially guide synaptic connectivity patterns in emerging neural networks.

Hunting increases phosphorylation of calcium/calmodulin-dependent protein kinase type II in adult barn owls

Juvenile barn owls readily adapt to prismatic spectacles, whereas adult owls living under standard aviary conditions do not. We previously demonstrated that phosphorylation of the cyclic-AMP response element-binding protein (CREB) provides a readout of the instructive signals that guide plasticity in juveniles. Here we investigated phosphorylation of calcium/calmodulin-dependent protein kinase II (pCaMKII) in both juveniles and adults. In contrast to CREB, we found no differences in pCaMKII expression between prism-wearing and control juveniles within the external nucleus of the inferior colliculus (ICX), the major site of plasticity. For prism-wearing adults that hunted live mice and are capable of adaptation, expression of pCaMKII was increased relative to prism-wearing adults that fed passively on dead mice and are not capable of adaptation. This effect did not bear the hallmarks of instructive information: it was not localized to rostral ICX and did not exhibit a patchy distribution reflecting discrete bimodal stimuli. These data are consistent with a role for CaMKII as a permissive rather than an instructive factor. In addition, the paucity of pCaMKII expression in passively fed adults suggests that the permissive default setting is "off" in adults.

Reading out a spatiotemporal population code by imaging neighbouring parallel fibre axons in vivo

The spatiotemporal pattern of synaptic inputs to the dendritic tree is crucial for synaptic integration and plasticity. However, it is not known if input patterns driven by sensory stimuli are structured or random. Here we investigate the spatial patterning of synaptic inputs by directly monitoring presynaptic activity in the intact mouse brain on the micron scale. Using in vivo calcium imaging of multiple neighbouring cerebellar parallel fibre axons, we find evidence for clustered patterns of axonal activity during sensory processing. The clustered parallel fibre input we observe is ideally suited for driving dendritic spikes, postsynaptic calcium signalling, and synaptic plasticity in downstream Purkinje cells, and is thus likely to be a major feature of cerebellar function during sensory processing.

The spatiotemporal pattern of synaptic inputs is critical for synaptic integration and plasticity in neurons but whether these inputs are structured or random is not clear. Here the authors use in vivo calcium imaging to monitor the presynaptic activity of cerebellar parallel fibre axons and find clustered patterns of axonal activity during sensory processing.

Hardware-amenable structural learning for spike-based pattern classification using a simple model of active dendrites

This letter presents a spike-based model that employs neurons with functionally distinct dendritic compartments for classifying high-dimensional binary patterns. The synaptic inputs arriving on each dendritic subunit are nonlinearly processed before being linearly integrated at the soma, giving the neuron the capacity to perform a large number of input-output mappings. The model uses sparse synaptic connectivity, where each synapse takes a binary value. The optimal connection pattern of a neuron is learned by using a simple hardware-friendly, margin-enhancing learning algorithm inspired by the mechanism of structural plasticity in biological neurons. The learning algorithm groups correlated synaptic inputs on the same dendritic branch. Since the learning results in modified connection patterns, it can be incorporated into current event-based neuromorphic systems with little overhead. This work also presents a branch-specific spike-based version of this structural plasticity rule. The proposed model is evaluated on benchmark binary classification problems, and its performance is compared against that achieved using support vector machine and extreme learning machine techniques. Our proposed method attains comparable performance while using 10% to 50% less in computational resource than the other reported techniques.

Synaptic clustering within dendrites: an emerging theory of memory formation

It is generally accepted that complex memories are stored in distributed representations throughout the brain, however the mechanisms underlying these representations are not understood. Here, we review recent findings regarding the subcellular mechanisms implicated in memory formation, which provide evidence for a dendrite-centered theory of memory. Plasticity-related phenomena which affect synaptic properties, such as synaptic tagging and capture, synaptic clustering, branch strength potentiation and spinogenesis provide the foundation for a model of memory storage that relies heavily on processes operating at the dendrite level. The emerging picture suggests that clusters of functionally related synapses may serve as key computational and memory storage units in the brain. We discuss both experimental evidence and theoretical models that support this hypothesis and explore its advantages for neuronal function.

Input clustering in the normal and learned circuits of adult barn owls

Experience-dependent formation of synaptic input clusters can occur in juvenile brains. Whether this also occurs in adults is largely unknown. We previously reconstructed the normal and learned circuits of prism-adapted barn owls and found that changes in clustering of axo-dendritic contacts (putative synapses) predicted functional circuit strength. Here we asked whether comparable changes occurred in normal and prism-removed adults. Across all anatomical zones, no systematic differences in the primary metrics for within-branch or between-branch clustering were observed: 95-99% of contacts resided within clusters (<10-20 μm from nearest neighbor) regardless of circuit strength. Bouton volumes, a proxy measure of synaptic strength, were on average larger in the functionally strong zones, indicating that changes in synaptic efficacy contributed to the differences in circuit strength. Bootstrap analysis showed that the distribution of inter-contact distances strongly deviated from random not in the functionally strong zones but in those that had been strong during the sensitive period (60-250 d), indicating that clusters formed early in life were preserved regardless of current value. While cluster formation in juveniles appeared to require the production of new synapses, cluster formation in adults did not. In total, these results support a model in which high cluster dynamics in juveniles sculpt a potential connectivity map that is refined in adulthood. We propose that preservation of clusters in functionally weak adult circuits provides a storage mechanism for disused but potentially useful pathways.

Dendritic nonlinearities are tuned for efficient spike-based computations in cortical circuits

Cortical neurons integrate thousands of synaptic inputs in their dendrites in highly nonlinear ways. It is unknown how these dendritic nonlinearities in individual cells contribute to computations at the level of neural circuits. Here, we show that dendritic nonlinearities are critical for the efficient integration of synaptic inputs in circuits performing analog computations with spiking neurons. We developed a theory that formalizes how a neuron's dendritic nonlinearity that is optimal for integrating synaptic inputs depends on the statistics of its presynaptic activity patterns. Based on their in vivo preynaptic population statistics (firing rates, membrane potential fluctuations, and correlations due to ensemble dynamics), our theory accurately predicted the responses of two different types of cortical pyramidal cells to patterned stimulation by two-photon glutamate uncaging. These results reveal a new computational principle underlying dendritic integration in cortical neurons by suggesting a functional link between cellular and systems--level properties of cortical circuits.

Refinement of the retinogeniculate synapse by bouton clustering

Mammalian sensory circuits become refined over development in an activity-dependent manner. Retinal ganglion cell (RGC) axons from each eye first map to their target in the geniculate and then segregate into eye-specific layers by the removal and addition of axon branches. Once segregation is complete, robust functional remodeling continues as the number of afferent inputs to each geniculate neuron decreases from many to a few. It is widely assumed that large-scale axon retraction underlies this later phase of circuit refinement. On the contrary, RGC axons remain stable during functional pruning. Instead, presynaptic boutons grow in size and cluster during this process. Moreover, they exhibit dynamic spatial reorganization in response to sensory experience. Surprisingly, axon complexity decreases only after the completion of the thalamic critical period. Therefore, dynamic bouton redistribution along a broad axon backbone represents an unappreciated form of plasticity underlying developmental wiring and rewiring in the CNS.

Input clustering and the microscale structure of local circuits

The recent development of powerful tools for high-throughput mapping of synaptic networks promises major advances in understanding brain function. One open question is how circuits integrate and store information. Competing models based on random vs. structured connectivity make distinct predictions regarding the dendritic addressing of synaptic inputs. In this article we review recent experimental tests of one of these models, the input clustering hypothesis. Across circuits, brain regions and species, there is growing evidence of a link between synaptic co-activation and dendritic location, although this finding is not universal. The functional implications of input clustering and future challenges are discussed.

An Augmented Two-Layer Model Captures Nonlinear Analog Spatial Integration Effects in Pyramidal Neuron Dendrites

In pursuit of the goal to understand and eventually reproduce the diverse functions of the brain, a key challenge lies in reverse engineering the peculiar biology-based "technology" that underlies the brain's remarkable ability to process and store information. The basic building block of the nervous system is the nerve cell, or "neuron," yet after more than 100 years of neurophysiological study and 60 years of modeling, the information processing functions of individual neurons, and the parameters that allow them to engage in so many different types of computation (sensory, motor, mnemonic, executive, etc.) remain poorly understood. In this paper, we review both historical and recent findings that have led to our current understanding of the analog spatial processing capabilities of dendrites, the major input structures of neurons, with a focus on the principal cell type of the neocortex and hippocampus, the pyramidal neuron (PN). We encapsulate our current understanding of PN dendritic integration in an abstract layered model whose spatially sensitive branch-subunits compute multidimensional sigmoidal functions. Unlike the 1-D sigmoids found in conventional neural network models, multidimensional sigmoids allow the cell to implement a rich spectrum of nonlinear modulation effects directly within their dendritic trees.

Synaptic competition in structural plasticity and cognitive function

Connections between neurons can undergo long-lasting changes in synaptic strength correlating with changes in structure. These events require the synthesis of new proteins, the availability of which can lead to cooperative and competitive interactions between synapses for the expression of plasticity. These processes can occur over limited spatial distances and temporal periods, defining dendritic regions over which activity may be integrated and could lead to the physical rewiring of synapses into functional groups. Such clustering of inputs may increase the computational power of neurons by allowing information to be combined in a greater than additive manner. The availability of new proteins may be a key modulatory step towards activity-dependent, long-term growth or elimination of spines necessary for remodelling of connections. Thus, the aberrant growth or shrinkage of dendritic spines could occur if protein levels are misregulated. Indeed, such perturbations can be seen in several mental retardation disorders, wherein either too much or too little protein translation exists, matching an observed increase or decrease in spine density, respectively. Cellular events which alter protein availability could relieve a constraint on synaptic competition and disturb synaptic clustering mechanisms. These changes may be detrimental to modifications in neural circuitry following activity.

Synaptic clustering during development and learning: the why, when, and how

To contribute to a functional network a neuron must make specific connections and integrate the synaptic inputs that it receives in a meaningful way. Previous modeling and experimental studies have predicted that this specificity could entail a subcellular organization whereby synapses that carry similar information are clustered together on local stretches of dendrite. Recent imaging studies have now, for the first time, demonstrated synaptic clustering during development and learning in different neuronal circuits. Interestingly, this organization is dependent on synaptic activity and most likely involves local plasticity mechanisms. Here we discuss these new insights and give an overview of the candidate plasticity mechanisms that could be involved.

Activity-dependent clustering of functional synaptic inputs on developing hippocampal dendrites

During brain development, before sensory systems become functional, neuronal networks spontaneously generate repetitive bursts of neuronal activity, which are typically synchronized across many neurons. Such activity patterns have been described on the level of networks and cells, but the fine-structure of inputs received by an individual neuron during spontaneous network activity has not been studied. Here, we used calcium imaging to record activity at many synapses of hippocampal pyramidal neurons simultaneously to establish the activity patterns in the majority of synapses of an entire cell. Analysis of the spatiotemporal patterns of synaptic activity revealed a fine-scale connectivity rule: neighboring synapses (<16 μm intersynapse distance) are more likely to be coactive than synapses that are farther away from each other. Blocking spiking activity or NMDA receptor activation revealed that the clustering of synaptic inputs required neuronal activity, demonstrating a role of developmentally expressed spontaneous activity for connecting neurons with subcellular precision.

Multiple clusters of release sites formed by individual thalamic afferents onto cortical interneurons ensure reliable transmission

Thalamic afferents supply the cortex with sensory information by contacting both excitatory neurons and inhibitory interneurons. Interestingly, thalamic contacts with interneurons constitute such a powerful synapse that even one afferent can fire interneurons, thereby driving feedforward inhibition. However, the spatial representation of this potent synapse on interneuron dendrites is poorly understood. Using Ca imaging and electron microscopy we show that an individual thalamic afferent forms multiple contacts with the interneuronal proximal dendritic arbor, preferentially near branch points. More contacts are correlated with larger amplitude synaptic responses. Each contact, consisting of a single bouton, can release up to seven vesicles simultaneously, resulting in graded and reliable Ca transients. Computational modeling indicates that the release of multiple vesicles at each contact minimally reduces the efficiency of the thalamic afferent in exciting the interneuron. This strategy preserves the spatial representation of thalamocortical inputs across the dendritic arbor over a wide range of release conditions.

Delayed appearance of the scaffolding proteins PSD-95 and Homer-1 at the developing rat calyx of Held synapse

The calyx of Held synapse is a giant axosomatic synapse that acts as a fast relay in the sound localization circuit of the brainstem. In rodents it forms within a relatively brief period starting around the second postnatal day (P2). The relative timing of the formation of its pre- and the postsynaptic compartment are not yet known. By means of fluorescent immunohistochemistry in neonatal rats we therefore compared the developmental expression patterns of the vesicular glutamate transporter (VGLUT)-1 and the postsynaptic density scaffolding proteins Homer-1 and PSD-95 in the medial nucleus of the trapezoid body (MNTB). Before its formation, colocalized punctate staining of VGLUT-1 and Homer-1 or PSD-95 was observed on principal neurons, in agreement with earlier work showing that they are already innervated by fibers from the cochlear nucleus before the calyx forms. The expression of VGLUT-1 clusters within the nascent calyx of Held synapse preceded the expression of Homer-1 and PSD-95 clusters, as indicated by the presence of principal neurons with only a large contiguous cluster (LCC) of VGLUT-1 at P2-3, whereas no neurons with only an LCC for Homer-1 or PSD-95 were seen. At P3 the first principal neurons with both a pre- and a postsynaptic LCC were observed, and at P12 all principal neurons had both a pre- and a postsynaptic LCC. The relatively late appearance of Homer-1 and PSD-95 within the developing calyx of Held synapse suggests that they play a role in its stabilization and the recruitment of additional receptors to its postsynaptic density.

Encoding of spatio-temporal input characteristics by a CA1 pyramidal neuron model

The in vivo activity of CA1 pyramidal neurons alternates between regular spiking and bursting, but how these changes affect information processing remains unclear. Using a detailed CA1 pyramidal neuron model, we investigate how timing and spatial arrangement variations in synaptic inputs to the distal and proximal dendritic layers influence the information content of model responses. We find that the temporal delay between activation of the two layers acts as a switch between excitability modes: short delays induce bursting while long delays decrease firing. For long delays, the average firing frequency of the model response discriminates spatially clustered from diffused inputs to the distal dendritic tree. For short delays, the onset latency and inter-spike-interval succession of model responses can accurately classify input signals as temporally close or distant and spatially clustered or diffused across different stimulation protocols. These findings suggest that a CA1 pyramidal neuron may be capable of encoding and transmitting presynaptic spatiotemporal information about the activity of the entorhinal cortex-hippocampal network to higher brain regions via the selective use of either a temporal or a rate code.

Pyramidal neurons in the hippocampus are crucially involved in learning and memory functions, but the ways in which they contribute to the processing of sensory inputs and their internal representation remain mostly unclear. The principal neurons of the CA1 region of the hippocampus are surrounded by at least 21 different types of interneurons. This feature, together with the fact that CA1 pyramidal dendrites associate two major glutamatergic inputs arriving from the entorhinal cortex, makes it laborious to track the ‘how’ and ‘what’ of synaptic integration. The present study tries to shed light on the ‘what’, that is, the information content of the CA1 discharge pattern. Using a detailed biophysical CA1 neuron model, we show that the output of the model neuron contains spatial and temporal features of the incoming synaptic input. This information lies in the temporal pattern of the inter-spike intervals produced during the bursting activity which is induced by the temporal coincidence of the two activated synaptic streams. Our findings suggest that CA1 pyramidal neurons may be capable of capturing features of the ongoing network activity via the use of a temporal code for information transfer.

Auditory processing, plasticity, and learning in the barn owl

The human brain has accumulated many useful building blocks over its evolutionary history, and the best knowledge of these has often derived from experiments performed in animal species that display finely honed abilities. In this article we review a model system at the forefront of investigation into the neural bases of information processing, plasticity, and learning: the barn owl auditory localization pathway. In addition to the broadly applicable principles gleaned from three decades of work in this system, there are good reasons to believe that continued exploration of the owl brain will be invaluable for further advances in understanding of how neuronal networks give rise to behavior.

Sensory Experience and Cortical Rewiring

Adult primary sensory cortex is not hard wired, but adapts to sensory experience. The cellular basis for cortical plasticity involves a combination of functional and structural changes in cortical neurons and the connections between them. Functional changes such as synaptic strengthening have been the focus of many investigations. However, structural modifications to the connections between neurons play an important role in cortical plasticity. In this review, the authors focus on structural remodeling that leads to rewiring of cortical circuits. Recent work has identified axonal remodeling, growth of new dendritic spines, and synapse turnover as important structural mechanisms for experience-dependent plasticity in mature cortex. These findings have begun to unravel how rewiring occurs in adult neocortex and offer new insights into the cellular mechanisms for learning and memory.