A neural circuit architecture for rapid learning in goal-directed navigation

Anchoring goals to spatial representations enables flexible navigation but is challenging in novel environments when both representations must be acquired simultaneously. We propose a framework for how Drosophila uses internal representations of head direction (HD) to build goal representations upon selective thermal reinforcement. We show that flies use stochastically generated fixations and directed saccades to express heading preferences in an operant visual learning paradigm and that HD neurons are required to modify these preferences based on reinforcement. We used a symmetric visual setting to expose how flies' HD and goal representations co-evolve and how the reliability of these interacting representations impacts behavior. Finally, we describe how rapid learning of new goal headings may rest on a behavioral policy whose parameters are flexible but whose form is genetically encoded in circuit architecture. Such evolutionarily structured architectures, which enable rapidly adaptive behavior driven by internal representations, may be relevant across species.

Synaptic homeostasis transiently leverages Hebbian mechanisms for a multiphasic response to inactivity

Homeostatic regulation of synapses is vital for nervous system function and key to understanding a range of neurological conditions. Synaptic homeostasis is proposed to operate over hours to counteract the destabilizing influence of long-term potentiation (LTP) and long-term depression (LTD). The prevailing view holds that synaptic scaling is a slow first-order process that regulates postsynaptic glutamate receptors and fundamentally differs from LTP or LTD. Surprisingly, we find that the dynamics of scaling induced by neuronal inactivity are not exponential or monotonic, and the mechanism requires calcineurin and CaMKII, molecules dominant in LTD and LTP. Our quantitative model of these enzymes reconstructs the unexpected dynamics of homeostatic scaling and reveals how synapses can efficiently safeguard future capacity for synaptic plasticity. This mechanism of synaptic adaptation supports a broader set of homeostatic changes, including action potential autoregulation, and invites further inquiry into how such a mechanism varies in health and disease.

Disentangling neuroplasticity mechanisms in post-stroke language recovery

A major objective in post-stroke aphasia research is to gain a deeper understanding of neuroplastic mechanisms that drive language recovery, with the ultimate goal of enhancing treatment outcomes. Subsequent to recent advances in neuroimaging techniques, we now have the ability to examine more closely how neural activity patterns change after a stroke. However, the way these neural activity changes relate to language impairments and language recovery is still debated. The aim of this review is to provide a theoretical framework to better investigate and interpret neuroplasticity mechanisms underlying language recovery in post-stroke aphasia. We detail two sets of neuroplasticity mechanisms observed at the synaptic level that may explain functional neuroimaging findings in post-stroke aphasia recovery at the network level: feedback-based homeostatic plasticity and associative Hebbian plasticity. In conjunction with these plasticity mechanisms, higher-order cognitive control processes dynamically modulate neural activity in other regions to meet communication demands, despite reduced neural resources. This work provides a network-level neurobiological framework for understanding neural changes observed in post-stroke aphasia and can be used to define guidelines for personalized treatment development.

Heterosynaptic plasticity of the visuo-auditory projection requires cholecystokinin released from entorhinal cortex afferents

The entorhinal cortex is involved in establishing enduring visuo-auditory associative memory in the neocortex. Here we explored the mechanisms underlying this synaptic plasticity related to projections from the visual and entorhinal cortices to the auditory cortex in mice using optogenetics of dual pathways. High-frequency laser stimulation (HFS laser) of the visuo-auditory projection did not induce long-term potentiation. However, after pairing with sound stimulus, the visuo-auditory inputs were potentiated following either infusion of cholecystokinin (CCK) or HFS laser of the entorhino-auditory CCK-expressing projection. Combining retrograde tracing and RNAscope in situ hybridization, we show that Cck expression is higher in entorhinal cortex neurons projecting to the auditory cortex than in those originating from the visual cortex. In the presence of CCK, potentiation in the neocortex occurred when the presynaptic input arrived 200 ms before postsynaptic firing, even after just five trials of pairing. Behaviorally, inactivation of the CCK+ projection from the entorhinal cortex to the auditory cortex blocked the formation of visuo-auditory associative memory. Our results indicate that neocortical visuo-auditory association is formed through heterosynaptic plasticity, which depends on release of CCK in the neocortex mostly from entorhinal afferents.

Can we manipulate brain connectivity? A systematic review of cortico-cortical paired associative stimulation effects

OBJECTIVE

Cortico-cortical paired associative stimulation (ccPAS) is a form of dual-site transcranial magnetic stimulation (TMS) entailing a series of single-TMS pulses paired at specific interstimulus intervals (ISI) delivered to distant cortical areas. The goal of this article is to systematically review its efficacy in inducing plasticity in humans focusing on stimulation parameters and hypotheses of underlying neurophysiology.

METHODS

A systematic review of the literature from 2009-2023 was undertaken to identify all articles utilizing ccPAS to study brain plasticity and connectivity. Six electronic databases were searched and included.

RESULTS

32 studies were identified. The studies targeted connections within the same hemisphere or between hemispheres. 28 ccPAS studies were in healthy participants, 1 study in schizophrenia, and 1 in Alzheimer's disease (AD) patients. 2 additional studies used cortico-cortical repetitive paired associative stimulation (cc-rPAS) in generalized anxiety disorder (GAD) patients. Outcome measures include electromyography (EMG), behavioral measures, electroencephalography (EEG), and functional magnetic resonance imaging (fMRI). ccPAS seems to be able to modulate brain connectivity depending on the ISI.

CONCLUSIONS

ccPAS can be used to modulate corticospinal excitability, brain activity, and behavior. Although the stimulation parameters used across studies reviewed in this paper are varied, ccPAS is a promising approach for basic research and potential clinical applications.

SIGNIFICANCE

Recent advances in neuroscience have caused a shift of interest from the study of single areas to a more complex approach focusing on networks of areas that orchestrate brain activity. Consequently, the TMS community is also witnessing a change, with a growing interest in targeting multiple brain areas rather than a single locus, as evidenced by an increasing number of papers using ccPAS. In light of this new enthusiasm for brain connectivity, this review summarizes existing literature and stimulation parameters that have proven effective in changing electrophysiological, behavioral, or neuroimaging-derived measures.

New biophysical rate-based modeling of long-term plasticity in mean-field neuronal population models

The formation of customized neural networks as the basis of brain functions such as receptive field selectivity, learning or memory depends heavily on the long-term plasticity of synaptic connections. However, the current mean-field population models commonly used to simulate large-scale neural network dynamics lack explicit links to the underlying cellular mechanisms of long-term plasticity. In this study, we developed a new mean-field population model, the plastic density-based neural mass model (pdNMM), by incorporating a newly developed rate-based plasticity model based on the calcium control hypothesis into an existing density-based neural mass model. Derivation of the plasticity model was carried out using population density methods. Our results showed that the synaptic plasticity represented by the resulting rate-based plasticity model exhibited Bienenstock-Cooper-Munro-like learning rules. Furthermore, we demonstrated that the pdNMM accurately reproduced previous experimental observations of long-term plasticity, including characteristics of Hebbian plasticity such as longevity, associativity and input specificity, on hippocampal slices, and the formation of receptive field selectivity in the visual cortex. In conclusion, the pdNMM is a novel approach that can confer long-term plasticity to conventional mean-field neuronal population models.

Cortico-cortical paired associative stimulation (ccPAS) over premotor-motor areas affects local circuitries in the human motor cortex via Hebbian plasticity

Transcranial magnetic stimulation (TMS) studies have shown that cortico-cortical paired associative stimulation (ccPAS) can strengthen connectivity between the ventral premotor cortex (PMv) and the primary motor cortex (M1) by modulating convergent input over M1 via Hebbian spike-timing-dependent plasticity (STDP). However, whether ccPAS locally affects M1 activity remains unclear. We tested 60 right-handed young healthy humans in two studies, using a combination of dual coil TMS and ccPAS over the left PMv and M1 to probe and manipulate PMv-to-M1 connectivity, and single- and paired-pulse TMS to assess neural activity within M1. We provide convergent evidence that ccPAS, relying on repeated activations of excitatory PMv-to-M1 connections, acts locally over M1. During ccPAS, motor-evoked potentials (MEPs) induced by paired PMv-M1 stimulation gradually increased. Following ccPAS, the threshold for inducing MEPs of different amplitudes decreased, and the input-output curve (IO) slope increased, highlighting increased M1 corticospinal excitability. Moreover, ccPAS reduced the magnitude of short-interval intracortical inhibition (SICI), reflecting suppression of GABA-ergic interneuronal mechanisms within M1, without affecting intracortical facilitation (ICF). These changes were specific to ccPAS Hebbian strengthening of PMv-to-M1 connectivity, as no modulations were observed when reversing the order of the PMv-M1 stimulation during a control ccPAS protocol. These findings expand prior ccPAS research that focused on the malleability of cortico-cortical connectivity at the network-level, and highlight local changes in the area of convergent activation (i.e., M1) during plasticity induction. These findings provide new mechanistic insights into the physiological basis of ccPAS that are relevant for protocol optimization.

Spike timing-dependent plasticity alters electrosensory neuron synaptic strength in vitro but does not consistently predict changes in sensory tuning in vivo

How do sensory systems optimize detection of behaviorally relevant stimuli when the sensory environment is constantly changing? We addressed the role of spike timing-dependent plasticity (STDP) in driving changes in synaptic strength in a sensory pathway and whether those changes in synaptic strength could alter sensory tuning. It is challenging to precisely control temporal patterns of synaptic activity in vivo and replicate those patterns in vitro in behaviorally relevant ways. This makes it difficult to make connections between STDP-induced changes in synaptic physiology and plasticity in sensory systems. Using the mormyrid species Brevimyrus niger and Brienomyrus brachyistius, which produce electric organ discharges for electrolocation and communication, we can precisely control the timing of synaptic input in vivo and replicate these same temporal patterns of synaptic input in vitro. In central electrosensory neurons in the electric communication pathway, using whole cell intracellular recordings in vitro, we paired presynaptic input with postsynaptic spiking at different delays. Using whole cell intracellular recordings in awake, behaving fish, we paired sensory stimulation with postsynaptic spiking using the same delays. We found that Hebbian STDP predictably alters sensory tuning in vitro and is mediated by NMDA receptors. However, the change in synaptic responses induced by sensory stimulation in vivo did not adhere to the direction predicted by the STDP observed in vitro. Further analysis suggests that this difference is influenced by polysynaptic activity, including inhibitory interneurons. Our findings suggest that STDP rules operating at identified synapses may not drive predictable changes in sensory responses at the circuit level.NEW & NOTEWORTHY We replicated behaviorally relevant temporal patterns of synaptic activity in vitro and used the same patterns during sensory stimulation in vivo. There was a Hebbian spike timing-dependent plasticity (STDP) pattern in vitro, but sensory responses in vivo did not shift according to STDP predictions. Analysis suggests that this disparity is influenced by differences in polysynaptic activity, including inhibitory interneurons. These results suggest that STDP rules at synapses in vitro do not necessarily apply to circuits in vivo.

Neuronal excitability as a regulator of circuit remodeling

Postnatal remodeling of neuronal connectivity shapes mature nervous systems.^1,2,3 The pruning of exuberant connections involves cell-autonomous and non-cell-autonomous mechanisms, such as neuronal activity. Indeed, experience-dependent competition sculpts various excitatory neuronal circuits.^4,5,6,7,8,9 Moreover, activity has been shown to regulate growth cone motility and the stability of neurites and synaptic connections.^{10,11,12,13,14} However, whether inhibitory activity influences the remodeling of neuronal connectivity or how activity influences remodeling in systems in which competition is not clearly apparent is not fully understood. Here, we use the Drosophila mushroom body (MB) as a model to examine the role of neuronal activity in the developmental axon pruning of γ-Kenyon cells. The MB is a neuronal structure in insects, implicated in associative learning and memory,^15,16 which receives mostly olfactory input from the antennal lobe.^17,18 The MB circuit includes intrinsic neurons, called Kenyon cells (KCs), which receive inhibitory input from the GABAergic anterior paired lateral (APL) neuron among other inputs. The γ-KCs undergo stereotypic, steroid-hormone-dependent remodeling^19,20 that involves the pruning of larval neurites followed by regrowth to form adult connections.²¹ We demonstrate that silencing neuronal activity is required for γ-KC pruning. Furthermore, we show that this is mechanistically achieved by cell-autonomous expression of the inward rectifying potassium channel 1 (irk1) combined with inhibition by APL neuron activity likely via GABA-B-R1 signaling. These results support the Hebbian-like rule "use it or lose it," where inhibition can destabilize connectivity and promote pruning while excitability stabilizes existing connections.

Paired optogenetic and visual stimulation can change the orientation selectivity of visual cortex neurons

Synaptic plasticity is currently considered the main mechanism underlying the plastic modification of neural networks. The vast majority of studies of synaptic plasticity are carried out on reduced preparations, but the situation in vivo is fundamentally different from that in vitro. In this work, we used the Hebbian paradigm, which is known to induce long-term changes in synaptic strength in vitro, to manipulate the properties of a single pyramidal neuron in the mouse visual cortex. We have shown that optogenetic stimulation of a ChR2-expressing pyramidal neuron in the primary visual cortex of Thy-ChR2 mice paired with the presentation of a visual stimulus of non-optimal orientation induces long-term changes in the properties of the receptive field, manifested in alteration of the orientation selectivity of the cell. Non-paired stimulation did not lead to changes in the properties of the receptive field of the neuron during the experiment. Thus, we have demonstrated the role of associative plasticity in the dynamic organization of the receptive fields of neurons in the visual cortex.

Visual Plasticity in Adulthood: Perspectives from Hebbian and Homeostatic Plasticity

The visual system retains profound plastic potential in adulthood. In the current review, we summarize the evidence of preserved plasticity in the adult visual system during visual perceptual learning as well as both monocular and binocular visual deprivation. In each condition, we discuss how such evidence reflects two major cellular mechanisms of plasticity: Hebbian and homeostatic processes. We focus on how these two mechanisms work together to shape plasticity in the visual system. In addition, we discuss how these two mechanisms could be further revealed in future studies investigating cross-modal plasticity in the visual system.

Transcranial cortico-cortical paired associative stimulation (ccPAS) over ventral premotor-motor pathways enhances action performance and corticomotor excitability in young adults more than in elderly adults

Transcranial magnetic stimulation (TMS) methods such as cortico-cortical paired associative stimulation (ccPAS) can increase the strength of functional connectivity between ventral premotor cortex (PMv) and primary motor cortex (M1) via spike timing-dependent plasticity (STDP), leading to enhanced motor functions in young adults. However, whether this STDP-inducing protocol is effective in the aging brain remains unclear. In two groups of young and elderly healthy adults, we evaluated manual dexterity with the 9-hole peg task before and after ccPAS of the left PMv-M1 circuit. We observed that ccPAS enhanced dexterity in young adults, and this effect was anticipated by a progressive increase in motor-evoked potentials (MEPs) during ccPAS administration. No similar effects were observed in elderly individuals or in a control task. Across age groups, we observed that the magnitude of MEP changes predicted larger behavioral improvements. These findings demonstrate that left PMv-to-M1 ccPAS induces functionally specific improvements in young adults' manual dexterity and an increase in corticomotor excitability, but altered plasticity prevents the effectiveness of ccPAS in the elderly.

Subcortical glutamatergic inputs exhibit a Hebbian form of long-term potentiation in the dentate gyrus

The hippocampus receives glutamatergic and GABAergic inputs from subcortical regions. Despite the important roles of these subcortical inputs in the regulation of hippocampal circuit, it has not been explored whether associative activation of the subcorticohippocampal pathway induces Hebbian plasticity of subcortical inputs. Here, we demonstrate that the hypothalamic supramammillary nucleus (SuM) to the dentate granule cell (GC) synapses, which co-release glutamate and GABA, undergo associative long-term potentiation (LTP) of glutamatergic, but not GABAergic, co-transmission. This LTP is induced by pairing of SuM inputs with GC spikes. We found that this Hebbian LTP is input-specific, requires NMDA receptors and CaMKII activation, and is expressed postsynaptically. By the net increase in excitatory drive of SuM inputs following LTP induction, associative inputs of SuM and the perforant path effectively discharge GCs. Our results highlight the important role of associative plasticity at SuM-GC synapses in the regulation of dentate gyrus activity and for the encoding of SuM-related information.

Hebbian activity-dependent plasticity in white matter

Synaptic plasticity is required for learning and follows Hebb's rule, the computational principle underpinning associative learning. In recent years, a complementary type of brain plasticity has been identified in myelinated axons, which make up the majority of brain's white matter. Like synaptic plasticity, myelin plasticity is required for learning, but it is unclear whether it is Hebbian or whether it follows different rules. Here, we provide evidence that white matter plasticity operates following Hebb's rule in humans. Across two experiments, we find that co-stimulating cortical areas to induce Hebbian plasticity leads to relative increases in cortical excitability and associated increases in a myelin marker within the stimulated fiber bundle. We conclude that Hebbian plasticity extends beyond synaptic changes and can be observed in human white matter fibers.

Adaptive erasure of spurious sequences in sensory cortical circuits

Sequential activity reflecting previously experienced temporal sequences is considered a hallmark of learning across cortical areas. However, it is unknown how cortical circuits avoid the converse problem: producing spurious sequences that are not reflecting sequences in their inputs. We develop methods to quantify and study sequentiality in neural responses. We show that recurrent circuit responses generally include spurious sequences, which are specifically prevented in circuits that obey two widely known features of cortical microcircuit organization: Dale’s law and Hebbian connectivity. In particular, spike-timing-dependent plasticity in excitation-inhibition networks leads to an adaptive erasure of spurious sequences. We tested our theory in multielectrode recordings from the visual cortex of awake ferrets. Although responses to natural stimuli were largely non-sequential, responses to artificial stimuli initially included spurious sequences, which diminished over extended exposure. These results reveal an unexpected role for Hebbian experience-dependent plasticity and Dale’s law in sensory cortical circuits.

•

Recurrent circuits generate spurious sequences without sequential inputs

•

A principled measure of total sequentiality in population responses is developed

•

Theory predicts that Hebbian plasticity should abolish spurious sequences

•

Spurious sequences in the visual cortex diminish with experience

Self-healing codes: How stable neural populations can track continually reconfiguring neural representations

The brain is capable of adapting while maintaining stable long-term memories and learned skills. Recent experiments show that neural responses are highly plastic in some circuits, while other circuits maintain consistent responses over time, raising the question of how these circuits interact coherently. We show how simple, biologically motivated Hebbian and homeostatic mechanisms in single neurons can allow circuits with fixed responses to continuously track a plastic, changing representation without reference to an external learning signal.

As an adaptive system, the brain must retain a faithful representation of the world while continuously integrating new information. Recent experiments have measured population activity in cortical and hippocampal circuits over many days and found that patterns of neural activity associated with fixed behavioral variables and percepts change dramatically over time. Such “representational drift” raises the question of how malleable population codes can interact coherently with stable long-term representations that are found in other circuits and with relatively rigid topographic mappings of peripheral sensory and motor signals. We explore how known plasticity mechanisms can allow single neurons to reliably read out an evolving population code without external error feedback. We find that interactions between Hebbian learning and single-cell homeostasis can exploit redundancy in a distributed population code to compensate for gradual changes in tuning. Recurrent feedback of partially stabilized readouts could allow a pool of readout cells to further correct inconsistencies introduced by representational drift. This shows how relatively simple, known mechanisms can stabilize neural tuning in the short term and provides a plausible explanation for how plastic neural codes remain integrated with consolidated, long-term representations.

A Nonlinear Hidden Layer Enables Actor-Critic Agents to Learn Multiple Paired Association Navigation

Navigation to multiple cued reward locations has been increasingly used to study rodent learning. Though deep reinforcement learning agents have been shown to be able to learn the task, they are not biologically plausible. Biologically plausible classic actor-critic agents have been shown to learn to navigate to single reward locations, but which biologically plausible agents are able to learn multiple cue-reward location tasks has remained unclear. In this computational study, we show versions of classic agents that learn to navigate to a single reward location, and adapt to reward location displacement, but are not able to learn multiple paired association navigation. The limitation is overcome by an agent in which place cell and cue information are first processed by a feedforward nonlinear hidden layer with synapses to the actor and critic subject to temporal difference error-modulated plasticity. Faster learning is obtained when the feedforward layer is replaced by a recurrent reservoir network.

Active inference leads to Bayesian neurophysiology

The neuronal substrates that implement the free-energy principle and ensuing active inference at the neuron and synapse level have not been fully elucidated. This Review considers possible neuronal substrates underlying the principle. First, the foundations of the free-energy principle are introduced, and then its ability to empirically explain various brain functions and psychological and biological phenomena in terms of Bayesian inference is described. Mathematically, the dynamics of neural activity and plasticity that minimise a cost function can be cast as performing Bayesian inference that minimises variational free energy. This equivalence licenses the adoption of the free-energy principle as a universal characterisation of neural networks. Further, the neural network structure itself represents a generative model under which an agent operates. A virtue of this perspective is that it enables the formal association of neural network properties with prior beliefs that regulate inference and learning. The possible neuronal substrates that implement prior beliefs and how to empirically examine the theory are discussed. This perspective renders brain activity explainable, leading to a deeper understanding of the neuronal mechanisms underlying basic psychology and psychiatric disorders in terms of an implicit generative model.

Mechanisms of Plasticity in Subcortical Visual Areas

Visual plasticity is classically considered to occur essentially in the primary and secondary cortical areas. Subcortical visual areas such as the dorsal lateral geniculate nucleus (dLGN) or the superior colliculus (SC) have long been held as basic structures responsible for a stable and defined function. In this model, the dLGN was considered as a relay of visual information travelling from the retina to cortical areas and the SC as a sensory integrator orienting body movements towards visual targets. However, recent findings suggest that both dLGN and SC neurons express functional plasticity, adding unexplored layers of complexity to their previously attributed functions. The existence of neuronal plasticity at the level of visual subcortical areas redefines our approach of the visual system. The aim of this paper is therefore to review the cellular and molecular mechanisms for activity-dependent plasticity of both synaptic transmission and cellular properties in subcortical visual areas.

Alteration in information flow through a pair of feeding command neurons underlies a form of Pavlovian conditioning in the Drosophila brain

Pavlovian conditioning¹ is a broadly used learning paradigm where defined stimuli are associated to induce behavioral switching. To define a causal relationship between activity change in a single neuron and behavioral switching, we took advantage of a "command neuron" that connects cellular function to behavior.² To examine the cellular and molecular basis of Pavlovian conditioning, we previously identified a pair of feeding command neurons termed "feeding neurons" in the adult Drosophila brain³ using genetic screening⁴ and opto- and thermo-genetic techniques.^5-7 The feeding neuron is activated by sweet signals like sucrose and induces the full complement of feeding behaviors, such as proboscis extension and food pumping. Ablation or inactivation of the pair of feeding neurons abolishes feeding behavior, suggesting that this single pair of neurons is indispensable for natural feeding behaviors.^2,3 Here, we describe a novel conditioning protocol to associate a signal-mediating rod removal from legs (conditioned stimulus [CS]) to feeding behavior induced by sucrose stimulation (unconditioned stimulus [US]). Calcium imaging of the feeding neuron demonstrated it acquires responsiveness to CS during conditioning, with inactivation of the feeding neuron during conditioning suppressing plasticity. These results suggest conditioning alters signals flowing from the CS into the feeding circuit, with the feeding neuron functioning as a key integrative hub for Hebbian plasticity.

Multifrequency Hebbian plasticity in coupled neural oscillators

We study multifrequency Hebbian plasticity by analyzing phenomenological models of weakly connected neural networks. We start with an analysis of a model for single-frequency networks previously shown to learn and memorize phase differences between component oscillators. We then study a model for gradient frequency neural networks (GrFNNs) which extends the single-frequency model by introducing frequency detuning and nonlinear coupling terms for multifrequency interactions. Our analysis focuses on models of two coupled oscillators and examines the dynamics of steady-state behaviors in multiple parameter regimes available to the models. We find that the model for two distinct frequencies shares essential dynamical properties with the single-frequency model and that Hebbian learning results in stronger connections for simple frequency ratios than for complex ratios. We then compare the analysis of the two-frequency model with numerical simulations of the GrFNN model and show that Hebbian plasticity in the latter is locally dominated by a nonlinear resonance captured by the two-frequency model.

The aging mouse brain: cognition, connectivity and calcium

Aging is a complex process that differentially impacts multiple cognitive, sensory, neuronal and molecular processes. Technological innovations now allow for parallel investigation of neuronal circuit function, structure and molecular composition in the brain of awake behaving adult mice. Thus, mice have become a critical tool to better understand how aging impacts the brain. However, a more granular systems-based approach, which considers the impact of age on key features relating to neural processing, is required. Here, we review evidence probing the impact of age on the mouse brain. We focus on a range of processes relating to neuronal function, including cognitive abilities, sensory systems, synaptic plasticity and calcium regulation. Across many systems, we find evidence for prominent age-related dysregulation even before 12 months of age, suggesting that emerging age-related alterations can manifest by late adulthood. However, we also find reports suggesting that some processes are remarkably resilient to aging. The evidence suggests that aging does not drive a parallel, linear dysregulation of all systems, but instead impacts some processes earlier, and more severely, than others. We propose that capturing the more fine-scale emerging features of age-related vulnerability and resilience may provide better opportunities for the rejuvenation of the aged brain.

The role of astrocyte-mediated plasticity in neural circuit development and function

Neuronal networks are capable of undergoing rapid structural and functional changes called plasticity, which are essential for shaping circuit function during nervous system development. These changes range from short-term modifications on the order of milliseconds, to long-term rearrangement of neural architecture that could last for the lifetime of the organism. Neural plasticity is most prominent during development, yet also plays a critical role during memory formation, behavior, and disease. Therefore, it is essential to define and characterize the mechanisms underlying the onset, duration, and form of plasticity. Astrocytes, the most numerous glial cell type in the human nervous system, are integral elements of synapses and are components of a glial network that can coordinate neural activity at a circuit-wide level. Moreover, their arrival to the CNS during late embryogenesis correlates to the onset of sensory-evoked activity, making them an interesting target for circuit plasticity studies. Technological advancements in the last decade have uncovered astrocytes as prominent regulators of circuit assembly and function. Here, we provide a brief historical perspective on our understanding of astrocytes in the nervous system, and review the latest advances on the role of astroglia in regulating circuit plasticity and function during nervous system development and homeostasis.

Characteristics of sequential activity in networks with temporally asymmetric Hebbian learning

Sequential activity is a prominent feature of many neural systems, in multiple behavioral contexts. Here, we investigate how Hebbian rules lead to storage and recall of random sequences of inputs in both rate and spiking recurrent networks. In the case of the simplest (bilinear) rule, we characterize extensively the regions in parameter space that allow sequence retrieval and compute analytically the storage capacity of the network. We show that nonlinearities in the learning rule can lead to sparse sequences and find that sequences maintain robust decoding but display highly labile dynamics to continuous changes in the connectivity matrix, similar to recent observations in hippocampus and parietal cortex.

Sequential activity has been observed in multiple neuronal circuits across species, neural structures, and behaviors. It has been hypothesized that sequences could arise from learning processes. However, it is still unclear whether biologically plausible synaptic plasticity rules can organize neuronal activity to form sequences whose statistics match experimental observations. Here, we investigate temporally asymmetric Hebbian rules in sparsely connected recurrent rate networks and develop a theory of the transient sequential activity observed after learning. These rules transform a sequence of random input patterns into synaptic weight updates. After learning, recalled sequential activity is reflected in the transient correlation of network activity with each of the stored input patterns. Using mean-field theory, we derive a low-dimensional description of the network dynamics and compute the storage capacity of these networks. Multiple temporal characteristics of the recalled sequential activity are consistent with experimental observations. We find that the degree of sparseness of the recalled sequences can be controlled by nonlinearities in the learning rule. Furthermore, sequences maintain robust decoding, but display highly labile dynamics, when synaptic connectivity is continuously modified due to noise or storage of other patterns, similar to recent observations in hippocampus and parietal cortex. Finally, we demonstrate that our results also hold in recurrent networks of spiking neurons with separate excitatory and inhibitory populations.

Why brain-controlled neuroprosthetics matter: mechanisms underlying electrical stimulation of muscles and nerves in rehabilitation

Delivering short trains of electric pulses to the muscles and nerves can elicit action potentials resulting in muscle contractions. When the stimulations are sequenced to generate functional movements, such as grasping or walking, the application is referred to as functional electrical stimulation (FES). Implications of the motor and sensory recruitment of muscles using FES go beyond simple contraction of muscles. Evidence suggests that FES can induce short- and long-term neurophysiological changes in the central nervous system by varying the stimulation parameters and delivery methods. By taking advantage of this, FES has been used to restore voluntary movement in individuals with neurological injuries with a technique called FES therapy (FEST). However, long-lasting cortical re-organization (neuroplasticity) depends on the ability to synchronize the descending (voluntary) commands and the successful execution of the intended task using a FES. Brain-computer interface (BCI) technologies offer a way to synchronize cortical commands and movements generated by FES, which can be advantageous for inducing neuroplasticity. Therefore, the aim of this review paper is to discuss the neurophysiological mechanisms of electrical stimulation of muscles and nerves and how BCI-controlled FES can be used in rehabilitation to improve motor function.

A Model for Structured Information Representation in Neural Networks of the Brain

Humans can reason at an abstract level and structure information into abstract categories, but the underlying neural processes have remained unknown. Recent experimental data provide the hint that this is likely to involve specific subareas of the brain from which structural information can be decoded. Based on this data, we introduce the concept of assembly projections, a general principle for attaching structural information to content in generic networks of spiking neurons. According to the assembly projections principle, structure-encoding assemblies emerge and are dynamically attached to content representations through Hebbian plasticity mechanisms. This model provides the basis for explaining a number of experimental data and provides a basis for modeling abstract computational operations of the brain.

Paired Associative Stimulation Fails to Induce Plasticity in Freely Behaving Intact Rats

Paired associative stimulation (PAS) has been explored in humans as a noninvasive tool to drive plasticity and promote recovery after neurologic insult. A more thorough understanding of PAS-induced plasticity is needed to fully harness it as a clinical tool. Here, we tested the efficacy of PAS with multiple interstimulus intervals in an awake rat model to study the principles of associative plasticity. Using chronically implanted electrodes in motor cortex and forelimb, we explored PAS parameters to effectively drive plasticity. We assessed changes in corticomotor excitability using a closed-loop, EMG-controlled cortical stimulation paradigm. We tested 11 PAS intervals, chosen to force the coincidence of neuronal activity in the motor cortex and spinal cord of rats with timings relevant to the principles of Hebbian spike timing-dependent plasticity. However, despite a relatively large number of stimulus pairings (300), none of the tested intervals reliably changed corticospinal excitability relative to control conditions. Our results question PAS effectiveness under these conditions.

Cognitive Enhancement via Network-Targeted Cortico-cortical Associative Brain Stimulation

Fluid intelligence (gf) represents a crucial component of human cognition, as it correlates with academic achievement, successful aging, and longevity. However, it has strong resilience against enhancement interventions, making the identification of gf enhancement approaches a key unmet goal of cognitive neuroscience. Here, we applied a spike-timing-dependent plasticity (STDP)-inducing brain stimulation protocol, named cortico-cortical paired associative stimulation (cc-PAS), to modulate gf in 29 healthy young subjects (13 females-mean ± standard deviation, 25.43 years ± 3.69), based on dual-coil transcranial magnetic stimulation (TMS). Pairs of neuronavigated TMS pulses (10-ms interval) were delivered over two frontoparietal nodes of the gf network, based on individual functional magnetic resonance imaging data and in accordance with cognitive models of information processing across the prefrontal and parietal lobe. cc-PAS enhanced accuracy at gf tasks, with parieto-frontal and fronto-parietal stimulation significantly increasing logical and relational reasoning, respectively. Results suggest the possibility of using SPTD-inducing TMS protocols to causally validate cognitive models by selectively engaging relevant networks and manipulating inter-regional temporal dynamics supporting specific cognitive functions.

Unsupervised Learning of Persistent and Sequential Activity

Two strikingly distinct types of activity have been observed in various brain structures during delay periods of delayed response tasks: Persistent activity (PA), in which a sub-population of neurons maintains an elevated firing rate throughout an entire delay period; and Sequential activity (SA), in which sub-populations of neurons are activated sequentially in time. It has been hypothesized that both types of dynamics can be “learned” by the relevant networks from the statistics of their inputs, thanks to mechanisms of synaptic plasticity. However, the necessary conditions for a synaptic plasticity rule and input statistics to learn these two types of dynamics in a stable fashion are still unclear. In particular, it is unclear whether a single learning rule is able to learn both types of activity patterns, depending on the statistics of the inputs driving the network. Here, we first characterize the complete bifurcation diagram of a firing rate model of multiple excitatory populations with an inhibitory mechanism, as a function of the parameters characterizing its connectivity. We then investigate how an unsupervised temporally asymmetric Hebbian plasticity rule shapes the dynamics of the network. Consistent with previous studies, we find that for stable learning of PA and SA, an additional stabilization mechanism is necessary. We show that a generalized version of the standard multiplicative homeostatic plasticity (Renart et al., 2003; Toyoizumi et al., 2014) stabilizes learning by effectively masking excitatory connections during stimulation and unmasking those connections during retrieval. Using the bifurcation diagram derived for fixed connectivity, we study analytically the temporal evolution and the steady state of the learned recurrent architecture as a function of parameters characterizing the external inputs. Slow changing stimuli lead to PA, while fast changing stimuli lead to SA. Our network model shows how a network with plastic synapses can stably and flexibly learn PA and SA in an unsupervised manner.

Introducing double bouquet cells into a modular cortical associative memory model

We present an electrophysiological model of double bouquet cells and integrate them into an established cortical columnar microcircuit model that has previously been used as a spiking attractor model for memory. Learning in that model relies on a Hebbian-Bayesian learning rule to condition recurrent connectivity between pyramidal cells. We here demonstrate that the inclusion of a biophysically plausible double bouquet cell model can solve earlier concerns about learning rules that simultaneously learn excitation and inhibition and might thus violate Dale’s principle. We show that learning ability and resulting effective connectivity between functional columns of previous network models is preserved when pyramidal synapses onto double bouquet cells are plastic under the same Hebbian-Bayesian learning rule. The proposed architecture draws on experimental evidence on double bouquet cells and effectively solves the problem of duplexed learning of inhibition and excitation by replacing recurrent inhibition between pyramidal cells in functional columns of different stimulus selectivity with a plastic disynaptic pathway. We thus show that the resulting change to the microcircuit architecture improves the model’s biological plausibility without otherwise impacting the model’s spiking activity, basic operation, and learning abilities.

Neural mechanisms of attending to items in working memory

Working memory, the ability to keep recently accessed information available for immediate manipulation, has been proposed to rely on two mechanisms that appear difficult to reconcile: self-sustained neural firing, or the opposite-activity-silent synaptic traces. Here we review and contrast models of these two mechanisms, and then show that both phenomena can co-exist within a unified system in which neurons hold information in both activity and synapses. Rapid plasticity in flexibly-coding neurons allows features to be bound together into objects, with an important emergent property being the focus of attention. One memory item is held by persistent activity in an attended or "focused" state, and is thus remembered better than other items. Other, previously attended items can remain in memory but in the background, encoded in activity-silent synaptic traces. This dual functional architecture provides a unified common mechanism accounting for a diversity of perplexing attention and memory effects that have been hitherto difficult to explain in a single theoretical framework.

Synaptic retinoic acid receptor signaling mediates mTOR-dependent metaplasticity that controls hippocampal learning

Homeostatic synaptic plasticity is a stabilizing mechanism engaged by neural circuits in response to prolonged perturbation of network activity. The non-Hebbian nature of homeostatic synaptic plasticity is thought to contribute to network stability by preventing "runaway" Hebbian plasticity at individual synapses. However, whether blocking homeostatic synaptic plasticity indeed induces runaway Hebbian plasticity in an intact neural circuit has not been explored. Furthermore, how compromised homeostatic synaptic plasticity impacts animal learning remains unclear. Here, we show in mice that the experience of an enriched environment (EE) engaged homeostatic synaptic plasticity in hippocampal circuits, thereby reducing excitatory synaptic transmission. This process required RARα, a nuclear retinoic acid receptor that doubles as a cytoplasmic retinoic acid-induced postsynaptic regulator of protein synthesis. Blocking RARα-dependent homeostatic synaptic plasticity during an EE experience by ablating RARα signaling induced runaway Hebbian plasticity, as evidenced by greatly enhanced long-term potentiation (LTP). As a consequence, RARα deletion in hippocampal circuits during an EE experience resulted in enhanced spatial learning but suppressed learning flexibility. In the absence of RARα, moreover, EE experience superactivated mammalian target of rapamycin (mTOR) signaling, causing a shift in protein translation that enhanced the expression levels of AMPA-type glutamate receptors. Treatment of mice with the mTOR inhibitor rapamycin during an EE experience not only restored normal AMPA-receptor expression levels but also reversed the increases in runaway Hebbian plasticity and learning after hippocampal RARα deletion. Thus, our findings reveal an RARα- and mTOR-dependent mechanism by which homeostatic plasticity controls Hebbian plasticity and learning.

Tumor necrosis factor (TNF) modulates synaptic plasticity in a concentration-dependent manner through intracellular calcium stores

The role of inflammatory signaling pathways in synaptic plasticity has long been identified. Yet, it remains unclear how inflammatory cytokines assert their pleiotropic effects on neural plasticity. Moreover, the neuronal targets through which inflammatory cytokines assert their effects on plasticity remain not well-understood. In an attempt to learn more about the plasticity-modulating effects of the pro-inflammatory cytokine tumor necrosis factor (TNF), we used two-pathway long-term potentiation (LTP) experiments at Schaffer collateral-CA1 synapses to test for concentration-dependent effects of TNF on synaptic plasticity. We report that high concentrations of TNF (1 μg/mL) impair the ability of mouse CA1 pyramidal neurons to express synaptic plasticity without affecting baseline synaptic transmission and/or previously established LTP. Interestingly, 100 ng/mL of TNF has no apparent effect on LTP, while low concentrations (1 ng/mL) promote the ability of neurons to express LTP. These dose-dependent metaplastic effects of TNF are modulated by intracellular calcium stores: Pharmacological activation of intracellular calcium stores with ryanodine (10 μM) reverses the negative effects of TNF_[high], and the plasticity-promoting effects of TNF_[low] are blocked when intracellular calcium stores are depleted with thapsigargin (1 μM). Consistent with this result, TNF does not promote plasticity in synaptopodin-deficient preparations, which show deficits in neuronal calcium store-mediated synaptic plasticity. Thus, we propose that TNF mediates its pleiotropic effects on synaptic plasticity in a concentration-dependent manner through signaling pathways that are modulated by intracellular calcium stores and require the presence of synaptopodin. These results demonstrate that TNF can act as mediator of metaplasticity, which is of considerable relevance in the context of brain diseases associated with increased TNF levels and alterations in synaptic plasticity.

KEY MESSAGES

• TNF modulates the ability of neurons to express synaptic plasticity. • High concentrations of TNF impair synaptic plasticity. • Low concentrations of TNF improve synaptic plasticity. • TNF does not affect previously established long-term potentiation. • Plasticity effects of TNF are modulated by intracellular calcium stores.

Comparison of the Efficacy of a Real-Time and Offline Associative Brain-Computer-Interface

An associative brain-computer-interface (BCI) that correlates in time a peripherally generated afferent volley with the peak negativity (PN) of the movement related cortical potential (MRCP) induces plastic changes in the human motor cortex. However, in this associative BCI the movement timed to a cue is not detected in real time. Thus, possible changes in reaction time caused by factors such as attention shifts or fatigue will lead to a decreased accuracy, less pairings, and likely reduced plasticity. The aim of the current study was to compare the effectiveness of this associative BCI intervention on plasticity induction when the MRCP PN time is pre-determined from a training data set (BCI_offline), or detected online (BCI_online). Ten healthy participants completed both interventions in randomized order. The average detection accuracy for the BCI_online intervention was 71 ± 3% with 2.8 ± 0.7 min^-1 false detections. For the BCI_online intervention the PN did not differ significantly between the training set and the actual intervention (t₉ = 0.87, p = 0.41). The peak-to-peak motor evoked potentials (MEPs) were quantified prior to, immediately following, and 30 min after the cessation of each intervention. MEP results revealed a significant main effect of time, F_(2,18) = 4.46, p = 0.027. The mean TA MEP amplitudes were significantly larger 30 min after (277 ± 72 μV) the BCI interventions compared to pre-intervention MEPs (233 ± 64 μV) regardless of intervention type and stimulation intensity (p = 0.029). These results provide further strong support for the associative nature of the associative BCI but also suggest that they likely differ to the associative long-term potentiation protocol they were modeled on in the exact sites of plasticity.

Endocannabinoid-LTP Mediated by CB1 and TRPV1 Receptors Encodes for Limited Occurrences of Coincident Activity in Neocortex

Synaptic efficacy changes, long-term potentiation (LTP) and depression (LTD), underlie various forms of learning and memory. Synaptic plasticity is generally assessed under prolonged activation, whereas learning can emerge from few or even a single trial. Here, we investigated the existence of rapid responsiveness of synaptic plasticity in response to a few number of spikes, in neocortex in a synaptic Hebbian learning rule, the spike-timing-dependent plasticity (STDP). We investigated the effect of lowering the number of pairings from 100 to 50, and 10 on STDP expression, using whole-cell recordings from pyramidal cells in rodent somatosensory cortical brain slices. We found that a low number of paired stimulations induces LTP at neocortical layer 4–2/3 synapses. Besides the asymmetric Hebbian STDP reported in the neocortex induced by 100 pairings, we observed a symmetric anti-Hebbian LTD for 50 pairings and unveiled a unidirectional Hebbian spike-timing-dependent LTP (tLTP) induced by 10–15 pairings. This tLTP was not mediated by NMDA receptor activation but requires CB₁ receptors and transient receptor potential vanilloid type-1 (TRPV1) activated by endocannabinoids (eCBs). eCBs have been widely described as mediating short- and long-term synaptic depression. Here, the eCB-tLTP reported at neocortical synapses could constitute a substrate operating in the online learning of new associative memories or during the initial stages of learning. In addition, these findings should provide useful insight into the mechanisms underlying eCB-plasticity occurring during marijuana intoxication.

Modulation of Spike-Timing Dependent Plasticity: Towards the Inclusion of a Third Factor in Computational Models

In spike-timing dependent plasticity (STDP) change in synaptic strength depends on the timing of pre- vs. postsynaptic spiking activity. Since STDP is in compliance with Hebb's postulate, it is considered one of the major mechanisms of memory storage and recall. STDP comprises a system of two coincidence detectors with N-methyl-D-aspartate receptor (NMDAR) activation often posited as one of the main components. Numerous studies have unveiled a third component of this coincidence detection system, namely neuromodulation and glia activity shaping STDP. Even though dopaminergic control of STDP has most often been reported, acetylcholine, noradrenaline, nitric oxide (NO), brain-derived neurotrophic factor (BDNF) or gamma-aminobutyric acid (GABA) also has been shown to effectively modulate STDP. Furthermore, it has been demonstrated that astrocytes, via the release or uptake of glutamate, gate STDP expression. At the most fundamental level, the timing properties of STDP are expected to depend on the spatiotemporal dynamics of the underlying signaling pathways. However in most cases, due to technical limitations experiments grant only indirect access to these pathways. Computational models carefully constrained by experiments, allow for a better qualitative understanding of the molecular basis of STDP and its regulation by neuromodulators. Recently, computational models of calcium dynamics and signaling pathway molecules have started to explore STDP emergence in ex and in vivo-like conditions. These models are expected to reproduce better at least part of the complex modulation of STDP as an emergent property of the underlying molecular pathways. Elucidation of the mechanisms underlying STDP modulation and its consequences on network dynamics is of critical importance and will allow better understanding of the major mechanisms of memory storage and recall both in health and disease.

Learning to Generate Sequences with Combination of Hebbian and Non-hebbian Plasticity in Recurrent Spiking Neural Networks

Synaptic Plasticity, the foundation for learning and memory formation in the human brain, manifests in various forms. Here, we combine the standard spike timing correlation based Hebbian plasticity with a non-Hebbian synaptic decay mechanism for training a recurrent spiking neural model to generate sequences. We show that inclusion of the adaptive decay of synaptic weights with standard STDP helps learn stable contextual dependencies between temporal sequences, while reducing the strong attractor states that emerge in recurrent models due to feedback loops. Furthermore, we show that the combined learning scheme suppresses the chaotic activity in the recurrent model substantially, thereby enhancing its' ability to generate sequences consistently even in the presence of perturbations.

The dialectic of Hebb and homeostasis

It has become widely accepted that homeostatic and Hebbian plasticity mechanisms work hand in glove to refine neural circuit function. Nonetheless, our understanding of how these fundamentally distinct forms of plasticity compliment (and under some circumstances interfere with) each other remains rudimentary. Here, I describe some of the recent progress of the field, as well as some of the deep puzzles that remain. These include unravelling the spatial and temporal scales of different homeostatic and Hebbian mechanisms, determining which aspects of network function are under homeostatic control, and understanding when and how homeostatic and Hebbian mechanisms must be segregated within neural circuits to prevent interference.This article is part of the themed issue 'Integrating Hebbian and homeostatic plasticity'.

Hebbian plasticity requires compensatory processes on multiple timescales

We review a body of theoretical and experimental research on Hebbian and homeostatic plasticity, starting from a puzzling observation: while homeostasis of synapses found in experiments is a slow compensatory process, most mathematical models of synaptic plasticity use rapid compensatory processes (RCPs). Even worse, with the slow homeostatic plasticity reported in experiments, simulations of existing plasticity models cannot maintain network stability unless further control mechanisms are implemented. To solve this paradox, we suggest that in addition to slow forms of homeostatic plasticity there are RCPs which stabilize synaptic plasticity on short timescales. These rapid processes may include heterosynaptic depression triggered by episodes of high postsynaptic firing rate. While slower forms of homeostatic plasticity are not sufficient to stabilize Hebbian plasticity, they are important for fine-tuning neural circuits. Taken together we suggest that learning and memory rely on an intricate interplay of diverse plasticity mechanisms on different timescales which jointly ensure stability and plasticity of neural circuits.This article is part of the themed issue 'Integrating Hebbian and homeostatic plasticity'.

Functional consequences of pre- and postsynaptic expression of synaptic plasticity

Growing experimental evidence shows that both homeostatic and Hebbian synaptic plasticity can be expressed presynaptically as well as postsynaptically. In this review, we start by discussing this evidence and methods used to determine expression loci. Next, we discuss the functional consequences of this diversity in pre- and postsynaptic expression of both homeostatic and Hebbian synaptic plasticity. In particular, we explore the functional consequences of a biologically tuned model of pre- and postsynaptically expressed spike-timing-dependent plasticity complemented with postsynaptic homeostatic control. The pre- and postsynaptic expression in this model predicts (i) more reliable receptive fields and sensory perception, (ii) rapid recovery of forgotten information (memory savings), and (iii) reduced response latencies, compared with a model with postsynaptic expression only. Finally, we discuss open questions that will require a considerable research effort to better elucidate how the specific locus of expression of homeostatic and Hebbian plasticity alters synaptic and network computations.This article is part of the themed issue 'Integrating Hebbian and homeostatic plasticity'.

A metaplasticity view of the interaction between homeostatic and Hebbian plasticity

Hebbian and homeostatic plasticity are two major forms of plasticity in the nervous system: Hebbian plasticity provides a synaptic basis for associative learning, whereas homeostatic plasticity serves to stabilize network activity. While achieving seemingly very different goals, these two types of plasticity interact functionally through overlapping elements in their respective mechanisms. Here, we review studies conducted in the mammalian central nervous system, summarize known circuit and molecular mechanisms of homeostatic plasticity, and compare these mechanisms with those that mediate Hebbian plasticity. We end with a discussion of 'local' homeostatic plasticity and the potential role of local homeostatic plasticity as a form of metaplasticity that modulates a neuron's future capacity for Hebbian plasticity.This article is part of the themed issue 'Integrating Hebbian and homeostatic plasticity'.

An Interactive Simulation Program for Exploring Computational Models of Auto-Associative Memory

While neuroscience students typically learn about activity-dependent plasticity early in their education, they often struggle to conceptually connect modification at the synaptic scale with network-level neuronal dynamics, not to mention with their own everyday experience of recalling a memory. We have developed an interactive simulation program (based on the Hopfield model of auto-associative memory) that enables the user to visualize the connections generated by any pattern of neural activity, as well as to simulate the network dynamics resulting from such connectivity. An accompanying set of student exercises introduces the concepts of pattern completion, pattern separation, and sparse versus distributed neural representations. Results from a conceptual assessment administered before and after students worked through these exercises indicate that the simulation program is a useful pedagogical tool for illustrating fundamental concepts of computational models of memory.

The malleable brain: plasticity of neural circuits and behavior - a review from students to students

One of the most intriguing features of the brain is its ability to be malleable, allowing it to adapt continually to changes in the environment. Specific neuronal activity patterns drive long-lasting increases or decreases in the strength of synaptic connections, referred to as long-term potentiation and long-term depression, respectively. Such phenomena have been described in a variety of model organisms, which are used to study molecular, structural, and functional aspects of synaptic plasticity. This review originated from the first International Society for Neurochemistry (ISN) and Journal of Neurochemistry (JNC) Flagship School held in Alpbach, Austria (Sep 2016), and will use its curriculum and discussions as a framework to review some of the current knowledge in the field of synaptic plasticity. First, we describe the role of plasticity during development and the persistent changes of neural circuitry occurring when sensory input is altered during critical developmental stages. We then outline the signaling cascades resulting in the synthesis of new plasticity-related proteins, which ultimately enable sustained changes in synaptic strength. Going beyond the traditional understanding of synaptic plasticity conceptualized by long-term potentiation and long-term depression, we discuss system-wide modifications and recently unveiled homeostatic mechanisms, such as synaptic scaling. Finally, we describe the neural circuits and synaptic plasticity mechanisms driving associative memory and motor learning. Evidence summarized in this review provides a current view of synaptic plasticity in its various forms, offers new insights into the underlying mechanisms and behavioral relevance, and provides directions for future research in the field of synaptic plasticity. Read the Editorial Highlight for this article on page 788. Cover Image for this issue: doi: 10.1111/jnc.13815.

A Spiking Working Memory Model Based on Hebbian Short-Term Potentiation

A dominant theory of working memory (WM), referred to as the persistent activity hypothesis, holds that recurrently connected neural networks, presumably located in the prefrontal cortex, encode and maintain WM memory items through sustained elevated activity. Reexamination of experimental data has shown that prefrontal cortex activity in single units during delay periods is much more variable than predicted by such a theory and associated computational models. Alternative models of WM maintenance based on synaptic plasticity, such as short-term nonassociative (non-Hebbian) synaptic facilitation, have been suggested but cannot account for encoding of novel associations. Here we test the hypothesis that a recently identified fast-expressing form of Hebbian synaptic plasticity (associative short-term potentiation) is a possible mechanism for WM encoding and maintenance. Our simulations using a spiking neural network model of cortex reproduce a range of cognitive memory effects in the classical multi-item WM task of encoding and immediate free recall of word lists. Memory reactivation in the model occurs in discrete oscillatory bursts rather than as sustained activity. We relate dynamic network activity as well as key synaptic characteristics to electrophysiological measurements. Our findings support the hypothesis that fast Hebbian short-term potentiation is a key WM mechanism.

SIGNIFICANCE STATEMENT Working memory (WM) is a key component of cognition. Hypotheses about the neural mechanism behind WM are currently under revision. Reflecting recent findings of fast Hebbian synaptic plasticity in cortex, we test whether a cortical spiking neural network model with such a mechanism can learn a multi-item WM task (word list learning). We show that our model can reproduce human cognitive phenomena and achieve comparable memory performance in both free and cued recall while being simultaneously compatible with experimental data on structure, connectivity, and neurophysiology of the underlying cortical tissue. These findings are directly relevant to the ongoing paradigm shift in the WM field.

Integrating Hebbian and homeostatic plasticity: the current state of the field and future research directions

We summarize here the results presented and subsequent discussion from the meeting on Integrating Hebbian and Homeostatic Plasticity at the Royal Society in April 2016. We first outline the major themes and results presented at the meeting. We next provide a synopsis of the outstanding questions that emerged from the discussion at the end of the meeting and finally suggest potential directions of research that we believe are most promising to develop an understanding of how these two forms of plasticity interact to facilitate functional changes in the brain.This article is part of the themed issue 'Integrating Hebbian and homeostatic plasticity'.

Rules for Shaping Neural Connections in the Developing Brain

It is well established that spontaneous activity in the developing mammalian brain plays a fundamental role in setting up the precise connectivity found in mature sensory circuits. Experiments that produce abnormal activity or that systematically alter neural firing patterns during periods of circuit development strongly suggest that the specific patterns and the degree of correlation in firing may contribute in an instructive manner to circuit refinement. In fish and amphibians, unlike amniotic vertebrates, sensory input directly drives patterned activity during the period of initial projection outgrowth and innervation. Experiments combining sensory stimulation with live imaging, which can be performed non-invasively in these simple vertebrate models, have provided important insights into the mechanisms by which neurons read out and respond to activity patterns. This article reviews the classic and recent literature on spontaneous and evoked activity-dependent circuit refinement in sensory systems and formalizes a set of mechanistic rules for the transformation of patterned activity into accurate neuronal connectivity in the developing brain.

Understanding the role of synaptopodin and the spine apparatus in Hebbian synaptic plasticity - New perspectives and the need for computational modeling

Synaptopodin (SP) is a proline-rich actin-associated protein essential for the formation of a spine apparatus (SA) in dendritic spines. The SA consists of stacks of smooth endoplasmic reticulum (sER) contiguous with the meshwork of somatodendritic ER. Spines of SP-deficient mice contain sER but no SA, demonstrating that SP is necessary for the assembly of ER cisterns into the more complex SA organelle. Although the SA was described decades ago, its function was difficult to investigate and remained elusive, in part because reliable markers for the SA were missing. After SP was identified as an essential component and a reliable marker of the SA, a role of SP/SA in hippocampal synaptic plasticity could be firmly established using loss-of-function approaches. Further studies revealed that SP/SA participate in the regulation of Ca²⁺-dependent spine-specific Hebbian plasticity and in activity-dependent changes in the spine actin cytoskeleton. In this review we are summarizing recent progress made on SP/SA in Hebbian plasticity and discuss open questions such as causality, spatiotemporal dynamics and complementarity of SP/SA-dependent mechanisms. We are proposing that computational modeling of spine Ca²⁺-signaling and actin remodeling pathways could address some of these issues and could indicate future research directions. Moreover, reaction-diffusion simulations could help to identify key feedforward and feedback regulatory motifs regulating the switch between an LTP and an LTD signaling module in SP/SA-containing spines, thus helping to find a unified view of SP/SA action in Hebbian plasticity.

Pencil-and-Paper Neural Networks: An Undergraduate Laboratory Exercise in Computational Neuroscience

Although it has been more than 70 years since McCulloch and Pitts published their seminal work on artificial neural networks, such models remain primarily in the domain of computer science departments in undergraduate education. This is unfortunate, as simple network models offer undergraduate students a much-needed bridge between cellular neurobiology and processes governing thought and behavior. Here, we present a very simple laboratory exercise in which students constructed, trained and tested artificial neural networks by hand on paper. They explored a variety of concepts, including pattern recognition, pattern completion, noise elimination and stimulus ambiguity. Learning gains were evident in changes in the use of language when writing about information processing in the brain.

Reward-Modulated Hebbian Plasticity as Leverage for Partially Embodied Control in Compliant Robotics

In embodied computation (or morphological computation), part of the complexity of motor control is offloaded to the body dynamics. We demonstrate that a simple Hebbian-like learning rule can be used to train systems with (partial) embodiment, and can be extended outside of the scope of traditional neural networks. To this end, we apply the learning rule to optimize the connection weights of recurrent neural networks with different topologies and for various tasks. We then apply this learning rule to a simulated compliant tensegrity robot by optimizing static feedback controllers that directly exploit the dynamics of the robot body. This leads to partially embodied controllers, i.e., hybrid controllers that naturally integrate the computations that are performed by the robot body into a neural network architecture. Our results demonstrate the universal applicability of reward-modulated Hebbian learning. Furthermore, they demonstrate the robustness of systems trained with the learning rule. This study strengthens our belief that compliant robots should or can be seen as computational units, instead of dumb hardware that needs a complex controller. This link between compliant robotics and neural networks is also the main reason for our search for simple universal learning rules for both neural networks and robotics.

Role of the site of synaptic competition and the balance of learning forces for Hebbian encoding of probabilistic Markov sequences

The majority of distinct sensory and motor events occur as temporally ordered sequences with rich probabilistic structure. Sequences can be characterized by the probability of transitioning from the current state to upcoming states (forward probability), as well as the probability of having transitioned to the current state from previous states (backward probability). Despite the prevalence of probabilistic sequencing of both sensory and motor events, the Hebbian mechanisms that mold synapses to reflect the statistics of experienced probabilistic sequences are not well understood. Here, we show through analytic calculations and numerical simulations that Hebbian plasticity (correlation, covariance, and STDP) with pre-synaptic competition can develop synaptic weights equal to the conditional forward transition probabilities present in the input sequence. In contrast, post-synaptic competition can develop synaptic weights proportional to the conditional backward probabilities of the same input sequence. We demonstrate that to stably reflect the conditional probability of a neuron's inputs and outputs, local Hebbian plasticity requires balance between competitive learning forces that promote synaptic differentiation and homogenizing learning forces that promote synaptic stabilization. The balance between these forces dictates a prior over the distribution of learned synaptic weights, strongly influencing both the rate at which structure emerges and the entropy of the final distribution of synaptic weights. Together, these results demonstrate a simple correspondence between the biophysical organization of neurons, the site of synaptic competition, and the temporal flow of information encoded in synaptic weights by Hebbian plasticity while highlighting the utility of balancing learning forces to accurately encode probability distributions, and prior expectations over such probability distributions.