Dynamic Synapses Research Articles

Enhanced Carrier Dynamics and Excitation of Optically Stimulated Artificial Synapse Using van der Waals Passivation Layers.

Advances in the semiconductor industry have been limited owing to the constraints imposed by silicon-based CMOS technology; hence, innovative device design approaches are necessary. This study focuses on "more than Moore" approaches, specifically in neuromorphic computing. Although MoS2 devices have attracted attention as neuromorphic computing candidates, their performances have been limited due to environment-induced perturbations to carrier dynamics and the formation of defect states. This study explores the integration of hydrocarbon (HC) layers onto active MoS2 channels to enhance neuromorphic computing characteristics. HC layers were employed in the proposed MoS2 field-effect transistor to facilitate stable optoelectrical control over the MoS2 channel under high-power stimulation. The improved electrical performance, stability, and synaptic behaviors of the HC-capped MoS2 devices compared to uncapped counterparts were experimentally demonstrated. The combination of optical and electrical tuning allowed for in-sensor computing applications that mimic human sensory behaviors. The impact of HC passivation on device performance was evaluated, and its potential for applications in neuromorphic computing with high stability was demonstrated across wide-ranging environmental conditions. The unique capabilities of HC-capped MoS2 devices were demonstrated by examining the spike duration-dependent plasticity and spiking timing-dependent plasticity. Thus, the proposed approach offers a promising avenue for advancing neuromorphic computing technologies.

A computational model for angular velocity integration in a locust heading circuit.

Accurate navigation often requires the maintenance of a robust internal estimate of heading relative to external surroundings. We present a model for angular velocity integration in a desert locust heading circuit, applying concepts from early theoretical work on heading circuits in mammals to a novel biological context in insects. In contrast to similar models proposed for the fruit fly, this circuit model uses a single 360° heading direction representation and is updated by neuromodulatory angular velocity inputs. Our computational model was implemented using steady-state firing rate neurons with dynamical synapses. The circuit connectivity was constrained by biological data, and remaining degrees of freedom were optimised with a machine learning approach to yield physiologically plausible neuron activities. We demonstrate that the integration of heading and angular velocity in this circuit is robust to noise. The heading signal can be effectively used as input to an existing insect goal-directed steering circuit, adapted for outbound locomotion in a steady direction that resembles locust migration. Our study supports the possibility that similar computations for orientation may be implemented differently in the neural hardware of the fruit fly and the locust.

Corrigendum to “Emergent population activity in metric-free and metric networks of neurons with stochastic spontaneous spikes and dynamic synapses” [Neurocomputing 461 (2021) 727–742

Corrigendum to “Emergent population activity in metric-free and metric networks of neurons with stochastic spontaneous spikes and dynamic synapses” [Neurocomputing 461 (2021) 727–742

Optimizing Clinical Translation of Bispecific T-cell Engagers through Context Unification with a Quantitative Systems Pharmacology Model.

Bispecific T-cell engagers (bsTCEs) have emerged as a promising class of cancer immunotherapy. BsTCEs enable physical connections between T cells and tumor cells to enhance T-cell activity against cancer. Despite several marketing approvals, the development of bsTCEs remains challenging, especially at early clinical translational stages. The intricate design of bsTCEs makes their pharmacologic effects and safety profiles highly dependent on patient's immunological and tumor conditions. Such context-dependent pharmacology introduces considerable uncertainty into translational efforts. In this study, we developed a Quantitative Systems Pharmacology (QSP) model, through context unification, that can facilitate the translation of bsTCEs preclinical data into clinical activity. Through characterizing the formation dynamics of immunological synapse (IS) induced by bsTCEs, this model unifies a broad range of contexts related to target affinity, tumor characteristics, and immunological conditions. After rigorous calibration using both experimental and clinical data, the model enables consistent translation of drug potency observed under diverse experimental conditions into predictable exposure-response relationships in patients. Moreover, the model can help identify optimal target-binding affinities and minimum efficacious concentrations across different clinical contexts. This QSP approach holds significant promise for the future development of bsTCEs.

Brain-derived neurotrophic factor scales presynaptic calcium transients to modulate excitatory neurotransmission

Brain-derived neurotrophic factor (BDNF) plays a critical role in synaptic physiology, as well as mechanisms underlying various neuropsychiatric diseases and their treatment. Despite its clear physiological role and disease relevance, BDNF's function at the presynaptic terminal, a fundamental unit of neurotransmission, remains poorly understood. In this study, we evaluated single synapse dynamics using optical imaging techniques in hippocampal cell cultures. We find that exogenous BDNF selectively increases evoked excitatory neurotransmission without affecting spontaneous neurotransmission. However, acutely blocking endogenous BDNF has no effect on evoked or spontaneous release, demonstrating that different approaches to studying BDNF may yield different results. When we suppressed BDNF-Tropomyosin receptor kinase B (TrkB) activity chronically over a period of days to weeks using a mouse line enabling conditional knockout of TrkB, we found that evoked glutamate release was significantly decreased while spontaneous release remained unchanged. Moreover, chronic blockade of BDNF-TrkB activity selectively downscales evoked calcium transients without affecting spontaneous calcium events. Via pharmacological blockade by voltage-gated calcium channel (VGCC) selective blockers, we found that the changes in evoked calcium transients are mediated by the P/Q subtype of VGCCs. These results suggest that BDNF-TrkB activity increases presynaptic VGCC activity to selectively increase evoked glutamate release.

Temporal dynamics and stoichiometry in human Notch signaling from Notch synaptic complex formation to nuclear entry of the Notch intracellular domain

Mammalian Notch signaling occurs when the binding of Delta or Jagged to Notch stimulates the proteolytic release of the Notch intracellular domain (NICD), which enters the nucleus to control target gene expression. To determine the temporal dynamics of events associated with Notch signaling under native conditions, we fluorescently tagged Notch and Delta at their endogenous genomic loci and visualized them upon pairing of receiver (Notch) and sender (Delta) cells as a function of time after cell contact. At contact sites, Notch and Delta immediately accumulated at 1:1 stoichiometry in synapses, which resolved by 15-20min after contact. Synapse formation preceded the entrance of the Notch extracellular domain into the sender cell and accumulation of NICD in the nucleus of the receiver cell, which approached a maximum after ∼45min and was prevented by chemical and genetic inhibitors of signaling. These findings directly link Notch-Delta synapse dynamics to NICD production with spatiotemporal precision.

Artificial neural network with dynamic synapse model

The purpose of this study is to develop and investigate a new short-term memory model based on an artificial neural network without short-term memory effect and a dynamic short-term memory model with astrocytic modulation. Methods. The artificial neural network is represented by a classical convolutional neural network that does not have short-term memory. Short-term memory is modeled in our hybrid model using the Tsodyks-Markram model, which is a system of third-order ordinary differential equations. Astrocyte dynamics is modeled by a mean field model of gliotransmitter concentration. Results. A new hybrid short-term memory model was developed and investigated using a convolutional neural network and a dynamic synapse model for an image recognition problem. Graphs of dependence of accuracy and error on the number of epochs for the presented model are given. The sensitivity metric of image recognition d-prime has been introduced. The developed model was compared with the recurrent neural network and the configuration of the new model without taking into account astrocytic modulation. A comparative table has been constructed showing the best recognition accuracy for the introduced model. Conclusion. As a result of the study, the possibility of combining an artificial neural network and a dynamic model that expands its functionality is shown. Comparison of the proposed model with short-term memory using a convolutional neural network and a dynamic synapse model with astrocytic modulation with a recurrent network showed the effectiveness of the proposed approach in simulating short-term memory.

Development and Application of Technology for Neural Circuit Visualization - Secondary Publication

The dynamics of neurite extension and synaptic connections are central issues in neural circuit research. The development of technologies for labeling purified cytoskeletal proteins with fluorescent dyes and introducing them into living neurons using microinjection greatly facilitated our understanding of cytoskeletal dynamics in neuronal axons. Imaging data showed that the cytoskeleton repeatedly polymerized and depolymerized within the axon, and elongation was driven by the new cytoskeleton formed at the axon tip. This finding significantly revised previously proposed models that explained slow axonal transport. After the discovery of green fluorescent protein (GFP), its potential application to the live imaging of neurons was recognized in the 1990s, and a new method for visualizing synapses using GFP-tagged postsynaptic scaffolding molecules was established. This method revealed the continuous turnover of synapses during development, which overturned the established theory that synapses are highly stable once they are formed. Live imaging of synapses also demonstrated that the molecular composition of synapses changes rapidly, driven by the rapid replacement of synaptic molecules. Fluorescence measurement of single GFP molecules enabled estimation of the absolute number of postsynaptic molecules in a single synapse. Furthermore, in multiple mouse models of autism spectrum disorders (ASDs), enhanced synapse turnover was detected as a common circuit-level phenotype. This study provides solid experimental evidence that an increase in synapse dynamics underlies the pathophysiology in mouse models of ASDs. The introduction of fluorescence imaging in neurobiology revealed that the neuronal cytoskeleton and synaptic structure are not static but dynamic cellular components. Imaging technology is expected to further advance our understanding of the dynamic properties of neurons and neural circuits.

ARTIFICIAL NEURAL NETWORKS WITH DYNAMIC SYNAPSES: A REVIEW

Artificial neural networks (ANNs) are widely applied to solve real-world problems. Most of the actions we take and the processes around us are time-varying. ANNs with dynamic properties allow processing time-dependent data and solving tasks such as speech and text processing, prediction models, face and emotion recognition, game strategy development. Dynamics in neural networks can appear in the input data, the architecture of the neural network, and the individual elements of the neural network – synapses and neurons. Unlike static synapses, dynamic synapses can change their connection strength based on incoming information. This is a fundamental principle allows neural networks to perform complex tasks like word processing or face recognition more efficiently. Dynamic synapses play a key role in the ability of artificial neural networks to learn from experience and change over time, which is one of the key aspects of artificial intelligence. The scientific works examined in this article show that there are no literature sources that review and compare dynamic DNTs according to their synapses. To fill this gap, the article reviews and groups DNTs with dynamic synapses. Dynamic neural networks are defined by providing a general mathematical expression. A dynamic synapse is described by specifying its main properties and presenting a general mathematical expression. Also an explanation, how these synapses can be modelled and integrated into 11 different dynamic ANNs is shown. Moreover, structures of dynamic ANNs are compared according to the properties of dynamic synapses.

Investigation into Molecular Brain Aging in Senescence-Accelerated Mouse (SAM) Model Employing Whole Transcriptomic Analysis in Search of Potential Molecular Targets for Therapeutic Interventions.

With the progression of an aging society, cognitive aging has emerged as a pressing concern necessitating attention. The senescence-accelerated mouse-prone 8 (SAMP8) model has proven instrumental in investigating the early stages of cognitive aging. Through an extensive examination of molecular changes in the brain cortex, utilizing integrated whole-genome transcriptomics, our principal aim was to uncover potential molecular targets with therapeutic applications and relevance to drug screening. Our investigation encompassed four distinct conditions, comparing the same strain at different time points (1 year vs. 16 weeks) and the same time point across different strains (SAMP8 vs. SAMR1), namely: physiological aging, accelerated aging, early events in accelerated aging, and late events in accelerated aging. Focusing on key functional alterations associated with aging in the brain, including neurogenesis, synapse dynamics, neurometabolism, and neuroinflammation, we identified candidate genes linked to these processes. Furthermore, employing protein-protein interaction (PPI) analysis, we identified pivotal hub genes involved in interactions within these functional domains. Additionally, gene-set perturbation analysis allowed us to uncover potential upstream genes or transcription factors that exhibited activation or inhibition across the four conditions. In summary, our comprehensive analysis of the SAMP8 mouse brain through whole-genome transcriptomics not only deepens our understanding of age-related changes but also lays the groundwork for a predictive model to facilitate drug screening for cognitive aging.

A Noelin-organized extracellular network of proteins required for constitutive and context-dependent anchoring of AMPA-receptors

Information processing and storage in the brain rely on AMPA-receptors (AMPARs) and their context-dependent dynamics in synapses and extra-synaptic sites. We found that distribution and dynamics of AMPARs in the plasma membrane are controlled by Noelins, a three-member family of conserved secreted proteins expressed throughout the brain in a cell-type-specific manner. Noelin tetramers tightly assemble with the extracellular domains of AMPARs and interconnect them in a network-like configuration with a variety of secreted and membrane-anchored proteins including Neurexin1, Neuritin1, and Seizure 6-like. Knock out of Noelins1-3 profoundly reduced AMPARs in synapses onto excitatory and inhibitory (inter)neurons, decreased their density and clustering in dendrites, and abolished activity-dependent synaptic plasticity. Our results uncover an endogenous mechanism for extracellular anchoring of AMPARs and establish Noelin-organized networks as versatile determinants of constitutive and context-dependent neurotransmission.

Mapping the dynamic high-density lipoprotein synapse

Background and aimsHeterogeneous high-density lipoprotein (HDL) particles, which can contain hundreds of proteins, affect human health and disease through dynamic molecular interactions with cell surface proteins. How HDL mediates its long-range signaling functions and interactions with various cell types is largely unknown. Due to the complexity of HDL, we hypothesize that multiple receptors engage with HDL particles resulting in condition-dependent receptor-HDL interaction clusters at the cell surface. MethodsHere we used the mass spectrometry-based and light-controlled proximity labeling strategy LUX-MS in a discovery-driven manner to decode HDL-receptor interactions. ResultsSurfaceome nanoscale organization analysis of hepatocytes and endothelial cells using LUX-MS revealed that the previously known HDL-binding protein scavenger receptor B1 (SCRB1) is embedded in a cell surface protein community, which we term HDL synapse. Modulating the endothelial HDL synapse, composed of 60 proteins, by silencing individual members, showed that the HDL synapse can be assembled in the absence of SCRB1 and that the members are interlinked. The aminopeptidase N (AMPN) (also known as CD13) was identified as an HDL synapse member that directly influences HDL uptake into the primary human aortic endothelial cells (HAECs). ConclusionsOur data indicate that preformed cell surface residing protein complexes modulate HDL function and suggest new theragnostic opportunities.

Hippocampal engram networks for fear memory recruit new synapses and modify pre-existing synapses in vivo

As basic units of neural networks, ensembles of synapses underlie cognitive functions such as learning and memory. These synaptic engrams show elevated synaptic density among engram cells following contextual fear memory formation. Subsequent analysis of the CA3-CA1 engram synapse revealed larger spine sizes, as the synaptic connectivity correlated with the memory strength. Here, we elucidate the synapse dynamics between CA3 and CA1 by tracking identical synapses at multiple time points by adapting two-photon microscopy and dual-eGRASP technique invivo. After memory formation, synaptic connections between engram populations are enhanced in conjunction with synaptogenesis within the hippocampal network. However, extinction learning specifically correlated with the disappearance of CA3 engram to CA1 engram (E-E) synapses. We observed "newly formed" synapses near pre-existing synapses, which clustered CA3-CA1 engram synapses after fear memory formation. Overall, we conclude that dynamics at CA3 to CA1 E-E synapses are key sites for modification during fear memory states.

Perineuronal nets affect memory and learning after synapse withdrawal

Perineuronal nets (PNNs) enwrap mature neurons, playing a role in the control of plasticity and synapse dynamics. PNNs have been shown to have effects on memory formation, retention and extinction in a variety of animal models. It has been proposed that the cavities in PNNs, which contain synapses, can act as a memory store and that they remain stable after events that cause synaptic withdrawal such as anoxia or hibernation. We examine this idea by monitoring place memory before and after synaptic withdrawal caused by acute hibernation-like state (HLS). Animals lacking hippocampal PNNs due to enzymatic digestion by chondroitinase ABC or knockout of the PNN component aggrecan were compared with wild type controls. HLS-induced synapse withdrawal caused a memory deficit, but not to the level of untreated naïve animals and not worsened by PNN attenuation. After HLS, only animals lacking PNNs showed memory restoration or relearning. Absence of PNNs affected the restoration of excitatory synapses on PNN-bearing neurons. The results support a role for hippocampal PNNs in learning, but not in long-term memory storage for correction of deficits.

RBL-2H3 Mast Cell Receptor Dynamics in the Immunological Synapse.

The RBL-2H3 mast cell immunological synapse dynamics is often simulated with reaction-diffusion and Fokker-Planck equations. The equations focus on how the cell synapse captures receptors following an immune response, where the receptor capture at the immunological site appears to be a delayed process. This article investigates the physical nature and mathematics behind such time-dependent delays. Using signal processing methods, convolution and cross-correlation-type delay capture simulations give a -squared range of 22 to 60, in good agreement with experimental results. The cell polarization event is offered as a possible explanation for these capture delays, where polarizing rates measure how fast the cell polarization event occurs. In the case of RBL-2H3 mast cells, polarization appears to be associated with cytoskeletal rearrangement; thus, both cytoskeletal and diffusional components are considered. From these simulations, a maximum polarizing rate ranging from 0.0057 s-2 to 0.031 s-2 is obtained. These results indicate that RBL-2H3 mast cells possess both temporal and spatial memory, and cell polarization is possibly linked to a Turing-type pattern formation.

Characteristic columnar connectivity caters to cortical computation: Replication, simulation, and evaluation of a microcircuit model.

The neocortex, and with it the mammalian brain, achieves a level of computational efficiency like no other existing computational engine. A deeper understanding of its building blocks (cortical microcircuits), and their underlying computational principles is thus of paramount interest. To this end, we need reproducible computational models that can be analyzed, modified, extended and quantitatively compared. In this study, we further that aim by providing a replication of a seminal cortical column model. This model consists of noisy Hodgkin-Huxley neurons connected by dynamic synapses, whose connectivity scheme is based on empirical findings from intracellular recordings. Our analysis confirms the key original finding that the specific, data-based connectivity structure enhances the computational performance compared to a variety of alternatively structured control circuits. For this comparison, we use tasks based on spike patterns and rates that require the systems not only to have simple classification capabilities, but also to retain information over time and to be able to compute nonlinear functions. Going beyond the scope of the original study, we demonstrate that this finding is independent of the complexity of the neuron model, which further strengthens the argument that it is the connectivity which is crucial. Finally, a detailed analysis of the memory capabilities of the circuits reveals a stereotypical memory profile common across all circuit variants. Notably, the circuit with laminar structure does not retain stimulus any longer than any other circuit type. We therefore conclude that the model's computational advantage lies in a sharper representation of the stimuli.

Kinetic Activity of Chromosomes and Expression of Recombination Genes in Achiasmatic Meiosis of Tityus (Archaeotityus) Scorpions.

Several species of Tityus (Scorpiones, Buthidae) present multi-chromosomal meiotic associations and failures in the synaptic process, originated from reciprocal translocations. Holocentric chromosomes and achiasmatic meiosis in males are present in all members of this genus. In the present study, we investigated synapse dynamics, transcriptional silencing by γH2AX, and meiotic microtubule association in bivalents and a quadrivalent of the scorpion Tityus maranhensis. Additionally, we performed RT-PCR to verify the expression of mismatch repair enzymes involved in crossing-over formation in Tityus silvestris gonads. The quadrivalent association in T. maranhensis showed delay in the synaptic process and long asynaptic regions during pachytene. In this species, γH2AX was recorded only at the chromosome ends during early stages of prophase I; in metaphase I, bivalents and quadrivalents of T. maranhensis exhibited binding to microtubules along their entire length, while in metaphase II/anaphase II transition, spindle fibers interacted only with telomeric regions. Regarding T. silvestris, genes involved in the recombination process were transcribed in ovaries, testes and embryos, without significant difference between these tissues. The expression of these genes during T. silvestris achiasmatic meiosis is discussed in the present study. The absence of meiotic inactivation by γH2AX and holo/telokinetic behavior of the chromosomes are important factors for the maintenance of the quadrivalent in T. maranhensis and the normal continuation of the meiotic cycle in this species.

A model of interacting quantum neurons with a dynamic synapse

Motivated by recent advances in neuroscience, in this work, we explore the emergent behaviour of quantum systems with a dynamical biologically-inspired qubits interaction. We use a minimal model of two interacting qubits with an activity-dependent dynamic interplay as in classical dynamic synapses that induces the so-called synaptic depression, that is, synapses that present synaptic fatigue after heavy presynaptic stimulation. Our study shows that in absence of synaptic depression the two-qubits quantum system shows typical Rabi oscillations whose frequency decreases when synaptic depression is introduced, so one can trap excitations for a large period of time. This creates a population imbalance between the qubits even though the Hamiltonian is Hermitian. This imbalance can be sustained in time by introducing a small energy shift between the qubits. In addition, we report that long time entanglement between the two qubits raises naturally in the presence of synaptic depression. Moreover, we propose and analyse a plausible experimental setup of our two-qubits system which demonstrates that these results are robust and can be experimentally obtained in a laboratory.

Sequence learning, prediction, and replay in networks of spiking neurons

Sequence learning, prediction and replay have been proposed to constitute the universal computations performed by the neocortex. The Hierarchical Temporal Memory (HTM) algorithm realizes these forms of computation. It learns sequences in an unsupervised and continuous manner using local learning rules, permits a context specific prediction of future sequence elements, and generates mismatch signals in case the predictions are not met. While the HTM algorithm accounts for a number of biological features such as topographic receptive fields, nonlinear dendritic processing, and sparse connectivity, it is based on abstract discrete-time neuron and synapse dynamics, as well as on plasticity mechanisms that can only partly be related to known biological mechanisms. Here, we devise a continuous-time implementation of the temporal-memory (TM) component of the HTM algorithm, which is based on a recurrent network of spiking neurons with biophysically interpretable variables and parameters. The model learns high-order sequences by means of a structural Hebbian synaptic plasticity mechanism supplemented with a rate-based homeostatic control. In combination with nonlinear dendritic input integration and local inhibitory feedback, this type of plasticity leads to the dynamic self-organization of narrow sequence-specific subnetworks. These subnetworks provide the substrate for a faithful propagation of sparse, synchronous activity, and, thereby, for a robust, context specific prediction of future sequence elements as well as for the autonomous replay of previously learned sequences. By strengthening the link to biology, our implementation facilitates the evaluation of the TM hypothesis based on experimentally accessible quantities. The continuous-time implementation of the TM algorithm permits, in particular, an investigation of the role of sequence timing for sequence learning, prediction and replay. We demonstrate this aspect by studying the effect of the sequence speed on the sequence learning performance and on the speed of autonomous sequence replay.

A Novel Image Edge Detection Method Based on the Asymmetric STDP Mechanism of the Visual Path

The detection of image edges plays an important role for image processing. In view of the fact that these existing methods cannot effectively detect the edge of the image when facing the image with rich details. This paper proposes a novel method of asymmetric spike-timing-dependent plasticity (STDP) image edge detection based on the visual physiological mechanism. In the proposed method, the original image is preprocessed by the Gabor filter to simulate the visual physiological orientation characteristics to obtain the image information in different directions, and the orientation feature fusion is used to reconstruct the primary edge feature information of the image. Then, based on the mechanism of the visual nervous system, a neuron network composed of dynamic synapses based on the asymmetric STDP mechanism is constructed to further process it to obtain impulse response images. In order to eliminate disturbance of the neuron’s system noise on the impulse response image, the impulse response image is filtered by a Gaussian filter. Then, the lateral inhibition between neurons is applied to refine the filtered image edges. Finally, the result is normalized, and the final edge of the experimental image is obtained. Experimental results based on the colony image data set collected in the laboratory indicate that the proposed method achieved better performance than these state-of-the-art methods; meanwhile, the AUC value remains above 0.6.