“Self” versus “Non-Self” Connectivity Dictates Properties of Synaptic Transmission and Plasticity

Autapses are connections between a neuron and itself. These connections are morphologically similar to “normal” synapses between two different neurons, and thus were long thought to have similar properties of synaptic transmission. However, this has not been directly tested. Here, using a micro-island culture assay in which we can define the number of interconnected cells, we directly compared synaptic transmission in excitatory autapses and in two-neuron micronetworks consisting of two excitatory neurons, in which a neuron is connected to one other neuron and to itself. We discovered that autaptic synapses are optimized for maximal transmission, and exhibited enhanced EPSC amplitude, charge, and RRP size compared to interneuronal synapses. However, autapses are deficient in several aspects of synaptic plasticity. Short-term potentiation only became apparent when a neuron was connected to another neuron. This acquisition of plasticity only required reciprocal innervation with one other neuron; micronetworks consisting of just two interconnected neurons exhibited enhanced short-term plasticity in terms of paired pulse ratio (PPR) and release probability (Pr), compared to autapses. Interestingly, when a neuron was connected to another neuron, not only interneuronal synapses, but also the autaptic synapses on itself exhibited a trend toward enhanced short-term plasticity in terms of PPR and Pr. Thus neurons can distinguish whether they are connected via “self” or “non-self” synapses and have the ability to adjust their plasticity parameters when connected to other neurons.     

Computing the Local Field Potential (LFP) from Integrate-and-Fire Network Models

Leaky integrate-and-fire (LIF) network models are commonly used to study how the spiking dynamics of neural networks changes with stimuli, tasks or dynamic network states. However, neurophysiological studies in vivo often rather measure the mass activity of neuronal microcircuits with the local field potential (LFP). Given that LFPs are generated by spatially separated currents across the neuronal membrane, they cannot be computed directly from quantities defined in models of point-like LIF neurons. Here, we explore the best approximation for predicting the LFP based on standard output from point-neuron LIF networks. To search for this best “LFP proxy”, we compared LFP predictions from candidate proxies based on LIF network output (e.g, firing rates, membrane potentials, synaptic currents) with “ground-truth” LFP obtained when the LIF network synaptic input currents were injected into an analogous three-dimensional (3D) network model of multi-compartmental neurons with realistic morphology, spatial distributions of somata and synapses. We found that a specific fixed linear combination of the LIF synaptic currents provided an accurate LFP proxy, accounting for most of the variance of the LFP time course observed in the 3D network for all recording locations. This proxy performed well over a broad set of conditions, including substantial variations of the neuronal morphologies. Our results provide a simple formula for estimating the time course of the LFP from LIF network simulations in cases where a single pyramidal population dominates the LFP generation, and thereby facilitate quantitative comparison between computational models and experimental LFP recordings in vivo.     

Structural Synaptic Plasticity Has High Memory Capacity and Can Explain Graded Amnesia,Catastrophic Forgetting,and the Spacing Effect

Although already William James and, more explicitly, Donald Hebb''s theory of cell assemblies have suggested that activity-dependent rewiring of neuronal networks is the substrate of learning and memory, over the last six decades most theoretical work on memory has focused on plasticity of existing synapses in prewired networks. Research in the last decade has emphasized that structural modification of synaptic connectivity is common in the adult brain and tightly correlated with learning and memory. Here we present a parsimonious computational model for learning by structural plasticity. The basic modeling units are “potential synapses” defined as locations in the network where synapses can potentially grow to connect two neurons. This model generalizes well-known previous models for associative learning based on weight plasticity. Therefore, existing theory can be applied to analyze how many memories and how much information structural plasticity can store in a synapse. Surprisingly, we find that structural plasticity largely outperforms weight plasticity and can achieve a much higher storage capacity per synapse. The effect of structural plasticity on the structure of sparsely connected networks is quite intuitive: Structural plasticity increases the “effectual network connectivity”, that is, the network wiring that specifically supports storage and recall of the memories. Further, this model of structural plasticity produces gradients of effectual connectivity in the course of learning, thereby explaining various cognitive phenomena including graded amnesia, catastrophic forgetting, and the spacing effect.     

miR‐124‐dependent tagging of synapses by synaptopodin enables input‐specific homeostatic plasticity

Homeostatic synaptic plasticity is a process by which neurons adjust their synaptic strength to compensate for perturbations in neuronal activity. Whether the highly diverse synapses on a neuron respond uniformly to the same perturbation remains unclear. Moreover, the molecular determinants that underlie synapse‐specific homeostatic synaptic plasticity are unknown. Here, we report a synaptic tagging mechanism in which the ability of individual synapses to increase their strength in response to activity deprivation depends on the local expression of the spine‐apparatus protein synaptopodin under the regulation of miR‐124. Using genetic manipulations to alter synaptopodin expression or regulation by miR‐124, we show that synaptopodin behaves as a “postsynaptic tag” whose translation is derepressed in a subpopulation of synapses and allows for nonuniform homeostatic strengthening and synaptic AMPA receptor stabilization. By genetically silencing individual connections in pairs of neurons, we demonstrate that this process operates in an input‐specific manner. Overall, our study shifts the current view that homeostatic synaptic plasticity affects all synapses uniformly to a more complex paradigm where the ability of individual synapses to undergo homeostatic changes depends on their own functional and biochemical state.     

Longitudinal Associations between Self-Rated Health and Performance-Based Physical Function in a Population-Based Cohort of Older Adults

Background

Although self-rated health (SRH) and performance-based physical function (PPF) are both strong predictors of mortality, little research has investigated the relationships between them. The objective of this study was to evaluate longitudinal, bi-directional associations between SRH and PPF.

Methods

We evaluated longitudinal associations between SRH and PPF in 3,610 adults aged 65–89 followed for an average of 4.8 (standard deviation [SD]: 4.4) years between 1994 and July 2011 in the Adult Changes in Thought study, a population-based cohort in the Seattle area. SRH was assessed with a single-item question in the ACT study. Participants were asked at each evaluation to rate their health as “excellent”, “very good”, “good”, “fair”, or “poor” in response to the question “In general, how would you rate your health at this time”. PPF scores (ranging from 0–16, with higher indicating better performance) included walking speed, chair rises, grip strength, and balance.

Results

At the baseline visit, participants averaged 74.5 (SD: 5.8) years of age and 2,115 (58.6%) were female. In multivariable linear mixed models, PPF declined with age, with more rapid decreases associated with very good, good, and fair (vs. excellent) baseline SRH. Adjusted annual change in PPF was −0.17 points (95% confidence interval [CI]: −0.19, −0.15) for individuals with excellent baseline SRH and −0.21 points (95% CI: −0.22, −0.19) for participants with fair SRH. In multivariable generalized linear mixed models, lower baseline PPF quartiles were associated with lower odds of excellent/very good/good SRH at age 75, however, differences between baseline PPF quartiles diminished with age.

Conclusions

These results suggest that less than excellent SRH predicts decline in physical functioning, however, poor physical functioning may not predict change in SRH in a reciprocal fashion. SRH provides a simple assessment tool for identifying individuals at increased risk for decline in physical function.     

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.     

Evaluation of the Performance of Information Theory-Based Methods and Cross-Correlation to Estimate the Functional Connectivity in Cortical Networks

Functional connectivity of in vitro neuronal networks was estimated by applying different statistical algorithms on data collected by Micro-Electrode Arrays (MEAs). First we tested these “connectivity methods” on neuronal network models at an increasing level of complexity and evaluated the performance in terms of ROC (Receiver Operating Characteristic) and PPC (Positive Precision Curve), a new defined complementary method specifically developed for functional links identification. Then, the algorithms better estimated the actual connectivity of the network models, were used to extract functional connectivity from cultured cortical networks coupled to MEAs. Among the proposed approaches, Transfer Entropy and Joint-Entropy showed the best results suggesting those methods as good candidates to extract functional links in actual neuronal networks from multi-site recordings.     

From decision to action: Detailed modelling of frog tadpoles reveals neuronal mechanisms of decision-making and reproduces unpredictable swimming movements in response to sensory signals

How does the brain process sensory stimuli, and decide whether to initiate locomotor behaviour? To investigate this question we develop two whole body computer models of a tadpole. The “Central Nervous System” (CNS) model uses evidence from whole-cell recording to define 2300 neurons in 12 classes to study how sensory signals from the skin initiate and stop swimming. In response to skin stimulation, it generates realistic sensory pathway spiking and shows how hindbrain sensory memory populations on each side can compete to initiate reticulospinal neuron firing and start swimming. The 3-D “Virtual Tadpole” (VT) biomechanical model with realistic muscle innervation, body flexion, body-water interaction, and movement is then used to evaluate if motor nerve outputs from the CNS model can produce swimming-like movements in a volume of “water”. We find that the whole tadpole VT model generates reliable and realistic swimming. Combining these two models opens new perspectives for experiments.     

Time of Progression to Osteopenia/Osteoporosis in Chronically HIV-Infected Patients: Screening DXA Scan

Background

Algorithms for bone mineral density (BMD) management in HIV-infected patients are lacking. Our objective was to assess how often a dual-energy x-ray absorptiometry (DXA) scan should be performed by assessing time of progression to osteopenia/osteoporosis.

Methods

All DXA scans performed between 2000 and 2009 from HIV-infected patients with at least two DXA were included. Time to an event (osteopenia and osteoporosis) was assessed using the Kaplan–Meier method. Strata (tertiles) were defined using baseline minimum T scores. Differences between strata in time to an event were compared with the log-rank test.

Results

Of 391 patients (1,639 DXAs), 49.6% had osteopenia and 21.7% osteoporosis at their first DXA scan. Of the 112 (28.6%) with normal BMD, 35.7% progressed to osteopenia; median progression time was 6.7 years. These patients were stratified: “low-risk" (baseline minimum T score >−0.2 SD), “middle-risk" (between −0.2 and −0.6 SD), and “high-risk" (from −0.6 to −1 SD); median progression time to osteopenia was 8.7, >7.2, and 1.7 years, respectively (p<0.0001). Of patients with osteopenia, 23.7% progressed to osteoporosis; median progression time was >8.5 years. Progression time was >8.2 years in “low-risk" tertile (T score between −1.1 and −1.6 SD), >8.5 years in “middle-risk" (between −1.6 and −2), and 3.2 years in “high-risk" (from −2 to −2.4) (p<0.0001).

Conclusions

Our results may help to define the BMD testing interval. The lowest T score tertiles would suggest recommending a subsequent DXA in 1–2 years; in the highest tertiles, ≥6 years. Early intervention in patients with bone demineralization could reduce fracture–related morbidity/mortality.     

A Computational Model of Neuro-Glio-Vascular Loop Interactions

We present a computational, biophysical model of neuron-astrocyte-vessel interaction. Unlike other cells, neurons convey “hunger” signals to the vascular network via an intervening layer of glial cells (astrocytes); vessels dilate and release glucose which fuels neuronal firing. Existing computational models focus on only parts of this loop (neuron→astrocyte→vessel→neuron), whereas the proposed model describes the entire loop. Neuronal firing causes release of a neurotransmitter like glutamate which triggers release of vasodilator by astrocytes via a cascade of biochemical events. Vasodilators released from astrocytic endfeet cause blood vessels to dilate and release glucose into the interstitium, part of which is taken up by the astrocyticendfeet. Glucose is converted into lactate in the astrocyte and transported into the neuron. Glucose from the interstitium and lactate (produced from glucose) influx from astrocyte are converted into ATP in the neuron. Neuronal ATP is used to drive the Na⁺/K⁺ATPase pumps, which maintain ionic gradients necessary for neuronal firing. When placed in the metabolic loop, the neuron exhibits sustained firing only when the stimulation current is more than a minimum threshold. For various combinations of initial neuronal [ATP] and external current, the neuron exhibits a variety of firing patterns including sustained firing, firing after an initial pause, burst firing etc. Neurovascular interactions under conditions of constricted vessels are also studied. Most models of cerebral circulation describe neurovascular interactions exclusively in the “forward” neuron→vessel direction. The proposed model indicates possibility of “reverse” influence also, with vasomotion rhythms influencing neural firing patterns. Another idea that emerges out of the proposed work is that brain''s computations may be more comprehensively understood in terms of neuro-glial-vascular dynamics and not in terms of neural firing alone.     

Taste Neurons Consist of Both a Large TrkB-Receptor-Dependent and a Small TrkB-Receptor-Independent Subpopulation

Brain-derived neurotrophic factor (BDNF) and neurotrophin-4 (NT-4) are two neurotrophins that play distinct roles in geniculate (taste) neuron survival, target innervation, and taste bud formation. These two neurotrophins both activate the tropomyosin-related kinase B (TrkB) receptor and the pan-neurotrophin receptor p75. Although the roles of these neurotrophins have been well studied, the degree to which BDNF and NT-4 act via TrkB to regulate taste development in vivo remains unclear. In this study, we compared taste development in TrkB^−/− and Bdnf^−/−/Ntf4^−/− mice to determine if these deficits were similar. If so, this would indicate that the functions of both BDNF and NT-4 can be accounted for by TrkB-signaling. We found that TrkB^−/− and Bdnf^−/−/Ntf4^−/− mice lose a similar number of geniculate neurons by E13.5, which indicates that both BDNF and NT-4 act primarily via TrkB to regulate geniculate neuron survival. Surprisingly, the few geniculate neurons that remain in TrkB^−/− mice are more successful at innervating the tongue and taste buds compared with those neurons that remain in Bdnf^−/−/Ntf4^−/− mice. The remaining neurons in TrkB^−/− mice support a significant number of taste buds. In addition, these remaining neurons do not express the TrkB receptor, which indicates that either BDNF or NT-4 must act via additional receptors to influence tongue innervation and/or targeting.     

A model of naturalistic decision making in preference tests

Decisions as to whether to continue with an ongoing activity or to switch to an alternative are a constant in an animal’s natural world, and in particular underlie foraging behavior and performance in food preference tests. Stimuli experienced by the animal both impact the choice and are themselves impacted by the choice, in a dynamic back and forth. Here, we present model neural circuits, based on spiking neurons, in which the choice to switch away from ongoing behavior instantiates this back and forth, arising as a state transition in neural activity. We analyze two classes of circuit, which differ in whether state transitions result from a loss of hedonic input from the stimulus (an “entice to stay” model) or from aversive stimulus-input (a “repel to leave” model). In both classes of model, we find that the mean time spent sampling a stimulus decreases with increasing value of the alternative stimulus, a fact that we linked to the inclusion of depressing synapses in our model. The competitive interaction is much greater in “entice to stay” model networks, which has qualitative features of the marginal value theorem, and thereby provides a framework for optimal foraging behavior. We offer suggestions as to how our models could be discriminatively tested through the analysis of electrophysiological and behavioral data.     

Analyzing Self-Similar and Fractal Properties of the C. elegans Neural Network

The brain is one of the most studied and highly complex systems in the biological world. While much research has concentrated on studying the brain directly, our focus is the structure of the brain itself: at its core an interconnected network of nodes (neurons). A better understanding of the structural connectivity of the brain should elucidate some of its functional properties. In this paper we analyze the connectome of the nematode Caenorhabditis elegans. Consisting of only 302 neurons, it is one of the better-understood neural networks. Using a Laplacian Matrix of the 279-neuron “giant component” of the network, we use an eigenvalue counting function to look for fractal-like self similarity. This matrix representation is also used to plot visualizations of the neural network in eigenfunction coordinates. Small-world properties of the system are examined, including average path length and clustering coefficient. We test for localization of eigenfunctions, using graph energy and spacial variance on these functions. To better understand results, all calculations are also performed on random networks, branching trees, and known fractals, as well as fractals which have been “rewired” to have small-world properties. We propose algorithms for generating Laplacian matrices of each of these graphs.     

Multiplex Networks of Cortical and Hippocampal Neurons Revealed at Different Timescales

Recent studies have emphasized the importance of multiplex networks – interdependent networks with shared nodes and different types of connections – in systems primarily outside of neuroscience. Though the multiplex properties of networks are frequently not considered, most networks are actually multiplex networks and the multiplex specific features of networks can greatly affect network behavior (e.g. fault tolerance). Thus, the study of networks of neurons could potentially be greatly enhanced using a multiplex perspective. Given the wide range of temporally dependent rhythms and phenomena present in neural systems, we chose to examine multiplex networks of individual neurons with time scale dependent connections. To study these networks, we used transfer entropy – an information theoretic quantity that can be used to measure linear and nonlinear interactions – to systematically measure the connectivity between individual neurons at different time scales in cortical and hippocampal slice cultures. We recorded the spiking activity of almost 12,000 neurons across 60 tissue samples using a 512-electrode array with 60 micrometer inter-electrode spacing and 50 microsecond temporal resolution. To the best of our knowledge, this preparation and recording method represents a superior combination of number of recorded neurons and temporal and spatial recording resolutions to any currently available in vivo system. We found that highly connected neurons (“hubs”) were localized to certain time scales, which, we hypothesize, increases the fault tolerance of the network. Conversely, a large proportion of non-hub neurons were not localized to certain time scales. In addition, we found that long and short time scale connectivity was uncorrelated. Finally, we found that long time scale networks were significantly less modular and more disassortative than short time scale networks in both tissue types. As far as we are aware, this analysis represents the first systematic study of temporally dependent multiplex networks among individual neurons.     

Dissecting neurotransmission with artificial synapses

JGP study demonstrates how recordings from neuron–HEK cell cocultures provide a clearer picture of the factors involved in synaptic transmission.

Resolving the rapid series of steps involved in synaptic transmission and assessing the contributions of different molecules to each of them is an enormous challenge. In this issue of JGP, Chiang et al. show that the process can be studied with greater resolution at the artificial synapses formed between neurons and cocultured human embryonic kidney (HEK) cells (1).Chung-Wei Chiang (left), Meyer Jackson (right), and colleagues studied synaptic transmission between hippocampal neurons and cocultured HEK cells. Mutations in the SNARE protein synaptobrevin 2 alter the shape of mEPSCs generated in HEK cells, an effect made clearer by the absence of dendritic filtering in this artificial system.Meyer Jackson and colleagues at the University of Wisconsin School of Medicine and Public Health are particularly interested in exocytosis. Though this process can be measured directly in endocrine cells, its role in controlling the dynamics of synaptic transmission can be difficult to separate from all the downstream steps required to elicit a response in the postsynaptic neuron. “We wanted to study a surrogate synapse with a simplified response to neurotransmitter that would allow us to focus on vesicle release with greater resolution,” Jackson explains.For years, researchers have studied synaptogenesis by transfecting HEK cells with a handful of postsynaptic factors that enable them to assemble functional synapses when cocultured with neurons (2, 3). Jackson realized that these artificial synapses lack two key sources of variability that can obscure the contribution of vesicle release to synaptic transmission. First, the postsynaptic apparatus of neuron–HEK synapses is consistent and can be precisely controlled (in contrast to regular synapses, where the molecular composition may vary from synapse to synapse). Second, the compact shape of HEK cells greatly reduces the influence of dendritic filtering, the phenomenon by which synaptic inputs take varying lengths of time to reach the cell body, depending on the location of the synapse within the dendritic arbor.Jackson and colleagues, including first author Chung-Wei Chiang, transfected HEK cells with four postsynaptic proteins—the AMPA receptor subunit GluA2, the adhesion molecule neuroligin 1, the scaffold protein PSD-95, and the accessory factor stargazin—and cocultured them with hippocampal neurons (1). The researchers then measured the miniature excitatory postsynaptic currents (mEPSCs) evoked in the HEK cells by the spontaneous release of single synaptic vesicles from neighboring neurons.The mEPSCs generated at these artificial synapses were larger and faster than mEPSCs produced by regular, neuron–neuron synapses (though they involved comparable amounts of charge, suggesting that the vesicle populations are similar at both types of synapse).Notably, the rise rate of mEPSCs in HEK cells was faster and less variable, in keeping with the absence of dendritic filtering and the consistent, shorter distance between the artificial synapses and the HEK cell body. This allowed Chiang et al. (1) to resolve the contribution of vesicle release to mEPSC dynamics using mutant versions of the SNARE protein synaptobrevin 2 that impede the flux of neurotransmitters through synaptic fusion pores (4). These mutant proteins decreased the rise rate and decay rate of mEPSCs at artificial synapses. “However, the effect was much clearer in HEK cells than we’d previously seen in neurons,” Jackson says.Chiang et al. (1) were also able to resolve the contribution of postsynaptic receptors to mEPSC shape, but Jackson is most interested in using the neuron–HEK coculture system to investigate synaptic vesicle release in more detail. “It opens up a new approach that will allow us to study synaptic exocytosis more precisely and look for much subtler effects,” Jackson says. For example, Jackson hopes to explain why mutations in some exocytotic proteins have major effects on endocrine cells but little to no effect at synapses.     

Mirroring intentional forgetting in a shared-goal learning situation

Background

Intentional forgetting refers to the surprising phenomenon that we can forget previously successfully encoded memories if we are instructed to do so. Here, we show that participants cannot only intentionally forget episodic memories but they can also mirror the “forgetting performance” of an observed model.

Methodology/Principal Findings

In four experiments a participant observed a model who took part in a memory experiment. In Experiment 1 and 2 observers saw a movie about the experiment, whereas in Experiment 3 and 4 the observers and the models took part together in a real laboratory experiment. The observed memory experiment was a directed forgetting experiment where the models learned two lists of items and were instructed either to forget or to remember the first list. In Experiment 1 and 3 observers were instructed to simply observe the experiment (“simple observation” instruction). In Experiment 2 and 4, observers received instructions aimed to induce the same learning goal for the observers and the models (“observation with goal-sharing” instruction). A directed forgetting effect (the reliably lower recall of to-be-forgotten items) emerged only when models received the “observation with goal-sharing” instruction (P<.001 in Experiment 2, and P<.05 in Experiment 4), and it was absent when observers received the “simple observation” instruction (P>.1 in Experiment 1 and 3).

Conclusion

If people observe another person with the same intention to learn, and see that this person is instructed to forget previously studied information, then they will produce the same intentional forgetting effect as the person they observed. This seems to be a an important aspect of human learning: if we can understand the goal of an observed person and this is in line with our behavioural goals then our learning performance will mirror the learning performance of the model.     

Hyperpolarization-Activated Current (Ih) Is Reduced in Hippocampal Neurons from Gabra5?/? Mice

Changes in the expression of γ-aminobutyric acid type A (GABA_A) receptors can either drive or mediate homeostatic alterations in neuronal excitability. A homeostatic relationship between α5 subunit-containing GABA_A (α5GABA_A) receptors that generate a tonic inhibitory conductance, and HCN channels that generate a hyperpolarization-activated cation current (I_h) was recently described for cortical neurons, where a reduction in I_h was accompanied by a reciprocal increase in the expression of α5GABA_A receptors resulting in the preservation of dendritosomatic synaptic function. Here, we report that in mice that lack the α5 subunit gene (Gabra5−/−), cultured embryonic hippocampal pyramidal neurons and ex vivo CA1 hippocampal neurons unexpectedly exhibited a decrease in I_h current density (by 40% and 28%, respectively), compared with neurons from wild-type (WT) mice. The resting membrane potential and membrane hyperpolarization induced by blockade of I_h with ZD-7288 were similar in cultured WT and Gabra5−/− neurons. In contrast, membrane hyperpolarization measured after a train of action potentials was lower in Gabra5−/− neurons than in WT neurons. Also, membrane impedance measured in response to low frequency stimulation was greater in cultured Gabra5−/− neurons. Finally, the expression of HCN1 protein that generates I_h was reduced by 41% in the hippocampus of Gabra5−/− mice. These data indicate that loss of a tonic GABAergic inhibitory conductance was followed by a compensatory reduction in I_h. The results further suggest that the maintenance of resting membrane potential is preferentially maintained in mature and immature hippocampal neurons through the homeostatic co-regulation of structurally and biophysically distinct cation and anion channels.     

Homeostatic control of synaptic rewiring in recurrent networks induces the formation of stable memory engrams

Brain networks store new memories using functional and structural synaptic plasticity. Memory formation is generally attributed to Hebbian plasticity, while homeostatic plasticity is thought to have an ancillary role in stabilizing network dynamics. Here we report that homeostatic plasticity alone can also lead to the formation of stable memories. We analyze this phenomenon using a new theory of network remodeling, combined with numerical simulations of recurrent spiking neural networks that exhibit structural plasticity based on firing rate homeostasis. These networks are able to store repeatedly presented patterns and recall them upon the presentation of incomplete cues. Storage is fast, governed by the homeostatic drift. In contrast, forgetting is slow, driven by a diffusion process. Joint stimulation of neurons induces the growth of associative connections between them, leading to the formation of memory engrams. These memories are stored in a distributed fashion throughout connectivity matrix, and individual synaptic connections have only a small influence. Although memory-specific connections are increased in number, the total number of inputs and outputs of neurons undergo only small changes during stimulation. We find that homeostatic structural plasticity induces a specific type of “silent memories”, different from conventional attractor states.     

Action Potential Modulation in CA1 Pyramidal Neuron Axons Facilitates OLM Interneuron Activation in Recurrent Inhibitory Microcircuits of Rat Hippocampus

Oriens-lacunosum moleculare (O-LM) interneurons in the CA1 region of the hippocampus play a key role in feedback inhibition and in the control of network activity. However, how these cells are efficiently activated in the network remains unclear. To address this question, I performed recordings from CA1 pyramidal neuron axons, the presynaptic fibers that provide feedback innervation of these interneurons. Two forms of axonal action potential (AP) modulation were identified. First, repetitive stimulation resulted in activity-dependent AP broadening. Broadening showed fast onset, with marked changes in AP shape following a single AP. Second, tonic depolarization in CA1 pyramidal neuron somata induced AP broadening in the axon, and depolarization-induced broadening summated with activity-dependent broadening. Outside-out patch recordings from CA1 pyramidal neuron axons revealed a high density of α-dendrotoxin (α-DTX)-sensitive, inactivating K⁺ channels, suggesting that K⁺ channel inactivation mechanistically contributes to AP broadening. To examine the functional consequences of axonal AP modulation for synaptic transmission, I performed paired recordings between synaptically connected CA1 pyramidal neurons and O-LM interneurons. CA1 pyramidal neuron–O-LM interneuron excitatory postsynaptic currents (EPSCs) showed facilitation during both repetitive stimulation and tonic depolarization of the presynaptic neuron. Both effects were mimicked and occluded by α-DTX, suggesting that they were mediated by K⁺ channel inactivation. Therefore, axonal AP modulation can greatly facilitate the activation of O-LM interneurons. In conclusion, modulation of AP shape in CA1 pyramidal neuron axons substantially enhances the efficacy of principal neuron–interneuron synapses, promoting the activation of O-LM interneurons in recurrent inhibitory microcircuits.