DOCSLIB.ORG

Corrigendum: Myelination Increases the Spatial Extent of Analog-Digital Modulation of Synaptic Transmission: a Modeling Study Mickaël Zbili, Dominique Debanne

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Corrigendum: Myelination Increases the Spatial Extent of Analog-Digital Modulation of Synaptic Transmission: a Modeling Study Mickaël Zbili, Dominique Debanne Corrigendum: Myelination Increases the Spatial Extent of Analog-Digital Modulation of Synaptic Transmission: A Modeling Study Mickaël Zbili, Dominique Debanne To cite this version: Mickaël Zbili, Dominique Debanne. Corrigendum: Myelination Increases the Spatial Extent of Analog- Digital Modulation of Synaptic Transmission: A Modeling Study. Frontiers in Cellular Neuroscience, Frontiers, 2020, 14, 10.3389/fncel.2020.00099. hal-03139013 HAL Id: hal-03139013 https://hal-amu.archives-ouvertes.fr/hal-03139013 Submitted on 11 Feb 2021 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Distributed under a Creative Commons Attribution| 4.0 International License CORRECTION published: 04 May 2020 doi: 10.3389/fncel.2020.00099 Corrigendum: Myelination Increases the Spatial Extent of Analog-Digital Modulation of Synaptic Transmission: A Modeling Study Mickaël Zbili 1,2* and Dominique Debanne 2* 1 Lyon Neuroscience Research Center, INSERM U1028-CNRS UMR5292-Université Claude Bernard Lyon1, Lyon, France, 2 UNIS UMR 1072 INSERM, AMU, Marseille, France Keywords: myelin, axon, axonal space constant, analog digital facilitation, spike shape, ion channels, axonal length constant Edited and reviewed by: Josef Bischofberger, A Corrigendum on University of Basel, Switzerland *Correspondence: Myelination Increases the Spatial Extent of Analog-Digital Modulation of Synaptic Mickaël Zbili Transmission: A Modeling Study [email protected] Dominique Debanne by Zbili, M., and Debanne, D. (2020). Front. Cell. Neurosci. 14:40. doi: 10.3389/fncel.2020.00040 [email protected] 2+ In the original article, there was an error in the equation W = A ∗ QCa describing how Specialty section: we computed the synaptic strength from the calcium charge in the presynaptic terminals. 2+ 2.5 This article was submitted to Actually, we used the following equation in the model: W = A ∗ (QCa ) . In consequence, Cellular Neurophysiology, a correction has been made to the Materials and Methods section, subsection Postsynaptic a section of the journal Responses, first paragraph: Frontiers in Cellular Neuroscience “To obtain the postsynaptic responses, we used Alpha Synapse Point Processes from Neuron 7.6 Received: 23 March 2020 inserted into postsynaptic cells. The weights of the synapses were calculated using the charge of the Accepted: 31 March 2020 spike-evoked Ca2+ entry in the presynaptic sites with the following power law: Published: 04 May 2020 2+ 2.5 Citation: W = A∗ (QCa ) Zbili M and Debanne D (2020) Corrigendum: Myelination Increases 2+ where W is the synaptic weight, A is a scaling factor and QCa is the charge of the spike-evoked Ca the Spatial Extent of Analog-Digital 2+ Modulation of Synaptic Transmission: current (Scott et al., 2008). Therefore, an increase in the Ca entry produced by an increase in A Modeling Study. presynaptic spike amplitude or duration led to an increase in the postsynaptic response amplitude.” Front. Cell. Neurosci. 14:99. The authors apologize for this error and state that this does not change the scientific conclusions doi: 10.3389/fncel.2020.00099 of the article in any way. The original article has been updated. Frontiers in Cellular Neuroscience | www.frontiersin.org 1 May 2020 | Volume 14 | Article 99 Zbili and Debanne Myelination Extends ADF Space Constant REFERENCES Copyright © 2020 Zbili and Debanne. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, Scott, R., Ruiz, A., Henneberger, C., Kullmann, D. M., and Rusakov, D. A. distribution or reproduction in other forums is permitted, provided the original (2008). Analog modulation of mossy fiber transmission is uncoupled author(s) and the copyright owner(s) are credited and that the original publication from changes in presynaptic Ca2+. J. Neurosci. 28, 7765–7773. in this journal is cited, in accordance with accepted academic practice. No use, doi: 10.1523/JNEUROSCI.1296-08.2008 distribution or reproduction is permitted which does not comply with these terms. Frontiers in Cellular Neuroscience | www.frontiersin.org 2 May 2020 | Volume 14 | Article 99.
Load more

Recommended publications
  • The Synaptic Weight Matrix
    The Synaptic Weight Matrix: Dynamics, Symmetry Breaking, and Disorder Francesco Fumarola Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Graduate School of Arts and Sciences COLUMBIA UNIVERSITY 2018 © 2018 Francesco Fumarola All rights reserved ABSTRACT The Synaptic Weight Matrix: Dynamics, Symmetry Breaking, and Disorder Francesco Fumarola A key role in simplified models of neural circuitry (Wilson and Cowan, 1972) is played by the matrix of synaptic weights, also called connectivity matrix, whose elements describe the amount of influence the firing of one neuron has on another. Biologically, this matrix evolves over time whether or not sensory inputs are present, and symmetries possessed by the internal dynamics of the network may break up sponta- neously, as found in the development of the visual cortex (Hubel and Wiesel, 1977). In this thesis, a full analytical treatment is provided for the simplest case of such a biologi- cal phenomenon, a single dendritic arbor driven by correlation-based dynamics (Linsker, 1988). Borrowing methods from the theory of Schrödinger operators, a complete study of the model is performed, analytically describing the break-up of rotational symmetry that leads to the functional specialization of cells. The structure of the eigenfunctions is calcu- lated, lower bounds are derived on the critical value of the correlation length, and explicit expressions are obtained for the long-term profile of receptive fields, i.e. the dependence of neural activity on external inputs. The emergence of a functional architecture of orientation preferences in the cortex is another crucial feature of visual information processing.
    [Show full text]
  • Synaptic Plasticity in Correlated Balanced Networks
    bioRxiv preprint doi: https://doi.org/10.1101/2020.04.26.061515; this version posted April 28, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. Synaptic Plasticity in Correlated Balanced Networks Alan Eric Akil Department of Mathematics, University of Houston, Houston, Texas, United States of America Robert Rosenbaum Department of Applied and Computational Mathematics and Statistics, University of Notre Dame, Notre Dame, Indiana, United States of America∗ Kreˇsimir Josi´cy Department of Mathematics, University of Houston, Houston, Texas, United States of America z (Dated: April 2020) 1 bioRxiv preprint doi: https://doi.org/10.1101/2020.04.26.061515; this version posted April 28, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. Abstract The dynamics of local cortical networks are irregular, but correlated. Dynamic excitatory{ inhibitory balance is a plausible mechanism that generates such irregular activity, but it remains unclear how balance is achieved and maintained in plastic neural networks. In particular, it is not fully understood how plasticity induced changes in the network affect balance, and in turn, how correlated, balanced activity impacts learning. How does the dynamics of balanced networks change under different plasticity rules? How does correlated spiking activity in recurrent networks change the evolution of weights, their eventual magnitude, and structure across the network? To address these questions, we develop a general theory of plasticity in balanced networks.
    [Show full text]
  • Energy Efficient Synaptic Plasticity Ho Ling Li1, Mark CW Van Rossum1,2*
    RESEARCH ARTICLE Energy efficient synaptic plasticity Ho Ling Li1, Mark CW van Rossum1,2* 1School of Psychology, University of Nottingham, Nottingham, United Kingdom; 2School of Mathematical Sciences, University of Nottingham, Nottingham, United Kingdom Abstract Many aspects of the brain’s design can be understood as the result of evolutionary drive toward metabolic efficiency. In addition to the energetic costs of neural computation and transmission, experimental evidence indicates that synaptic plasticity is metabolically demanding as well. As synaptic plasticity is crucial for learning, we examine how these metabolic costs enter in learning. We find that when synaptic plasticity rules are naively implemented, training neural networks requires extremely large amounts of energy when storing many patterns. We propose that this is avoided by precisely balancing labile forms of synaptic plasticity with more stable forms. This algorithm, termed synaptic caching, boosts energy efficiency manifold and can be used with any plasticity rule, including back-propagation. Our results yield a novel interpretation of the multiple forms of neural synaptic plasticity observed experimentally, including synaptic tagging and capture phenomena. Furthermore, our results are relevant for energy efficient neuromorphic designs. Introduction The human brain only weighs 2% of the total body mass but is responsible for 20% of resting metab- olism (Attwell and Laughlin, 2001; Harris et al., 2012). The brain’s energy need is believed to have shaped many aspects of its design, such as its sparse coding strategy (Levy and Baxter, 1996; Len- nie, 2003), the biophysics of the mammalian action potential (Alle et al., 2009; Fohlmeister, 2009), *For correspondence: and synaptic failure (Levy and Baxter, 2002; Harris et al., 2012).
    [Show full text]
  • Arxiv:2103.12649V1 [Q-Bio.NC] 23 Mar 2021
    A synapse-centric account of the free energy principle David Kappel and Christian Tetzlaff March 24, 2021 Abstract The free energy principle (FEP) is a mathematical framework that describes how biological systems self- organize and survive in their environment. This principle provides insights on multiple scales, from high-level behavioral and cognitive functions such as attention or foraging, down to the dynamics of specialized cortical microcircuits, suggesting that the FEP manifests on several levels of brain function. Here, we apply the FEP to one of the smallest functional units of the brain: single excitatory synaptic connections. By focusing on an experimentally well understood biological system we are able to derive learning rules from first principles while keeping assumptions minimal. This synapse-centric account of the FEP predicts that synapses interact with the soma of the post-synaptic neuron through stochastic synaptic releases to probe their behavior and use back-propagating action potentials as feedback to update the synaptic weights. The emergent learning rules are regulated triplet STDP rules that depend only on the timing of the pre- and post-synaptic spikes and the internal states of the synapse. The parameters of the learning rules are fully determined by the parameters of the post-synaptic neuron model, suggesting a close interplay between the synaptic and somatic compartment and making precise predictions about the synaptic dynamics. The synapse-level uncertainties automatically lead to representations of uncertainty on the network level that manifest in ambiguous situations. We show that the FEP learning rules can be applied to spiking neural networks for supervised and unsupervised learning and for a closed loop learning task where a behaving agent interacts with an environment.
    [Show full text]
  • Improved Expressivity Through Dendritic Neural Networks
    Improved Expressivity Through Dendritic Neural Networks Xundong Wu Xiangwen Liu Wei Li Qing Wu School of Computer Science and Technology Hangzhou Dianzi University, Hangzhou, China [email protected], [email protected] A typical biological neuron, such as a pyramidal neuron of the neocortex, receives thousands of afferent synaptic inputs on its dendrite tree and sends the efferent axonal output downstream. In typical artificial neural networks, dendrite trees are modeled as linear structures that funnel weighted synaptic inputs to the cell bodies. However, numerous experimental and theoretical studies have shown that dendritic arbors are far more than simple linear accumulators. That is, synaptic inputs can actively modulate their neighboring synaptic activities; therefore, the dendritic structures are highly nonlinear. In this study, we model such local nonlinearity of dendritic trees with our dendritic neural network (DENN) structure and apply this structure to typical machine learning tasks. Equipped with localized nonlinearities, DENNs can attain greater model expressivity than regular neural networks while maintaining efficient network inference. Such strength is evidenced by the increased fitting power when we train DENNs with supervised machine learning tasks. We also empirically show that the locality structure of DENNs can improve the generalization performance, as exemplified by DENNs outranking naive deep neural network architectures when tested on classification tasks from the UCI machine learning repository. 1 Introduction Deep learning algorithms have made remarkable achievements in a vast array of fields over the past few years. Notably, deep convolutional neural networks have revolutionized computer vision research and real-world applications. Inspired by biological neuronal networks in our brains, originally in the form of a perceptron [37], an artificial neural network unit is typically constructed as a simple weighted sum of synaptic inputs followed by feeding the summation result through an activation function.
    [Show full text]
  • Synaptic Plasticity As Bayesian Inference
    ARTICLES https://doi.org/10.1038/s41593-021-00809-5 Synaptic plasticity as Bayesian inference Laurence Aitchison 1,2 ✉ , Jannes Jegminat 3,4, Jorge Aurelio Menendez 1,5, Jean-Pascal Pfister3,4, Alexandre Pouget 1,6,7 and Peter E. Latham 1,7 Learning, especially rapid learning, is critical for survival. However, learning is hard; a large number of synaptic weights must be set based on noisy, often ambiguous, sensory information. In such a high-noise regime, keeping track of probability distribu- tions over weights is the optimal strategy. Here we hypothesize that synapses take that strategy; in essence, when they esti- mate weights, they include error bars. They then use that uncertainty to adjust their learning rates, with more uncertain weights having higher learning rates. We also make a second, independent, hypothesis: synapses communicate their uncertainty by linking it to variability in postsynaptic potential size, with more uncertainty leading to more variability. These two hypotheses cast synaptic plasticity as a problem of Bayesian inference, and thus provide a normative view of learning. They generalize known learning rules, offer an explanation for the large variability in the size of postsynaptic potentials and make falsifiable experimental predictions. o survive, animals must accurately estimate the state of the state-of-the-art adaptive optimization algorithms for artificial neu- world. This estimation problem is plagued by uncertainty: ral networks4. Tnot only is information often extremely limited (for example, Bayesian plasticity is a hypothesis about what synapses com- because it is dark) or ambiguous (for example, a rustle in the bushes pute. It does not, however, tell synapses how to set their weights; could be the wind, or it could be a predator), but sensory recep- for that, a second hypothesis is needed.
    [Show full text]
  • Mathematical Models of Dynamic Behavior of Individual Neural Networks of Central Nervous System Dimitra-Despoina Pagania, Adam Adamopoulos, Spiridon Likothanassis
    Mathematical Models of Dynamic Behavior of Individual Neural Networks of Central Nervous System Dimitra-Despoina Pagania, Adam Adamopoulos, Spiridon Likothanassis To cite this version: Dimitra-Despoina Pagania, Adam Adamopoulos, Spiridon Likothanassis. Mathematical Models of Dynamic Behavior of Individual Neural Networks of Central Nervous System. 12th Engineering Appli- cations of Neural Networks (EANN 2011) and 7th Artificial Intelligence Applications and Innovations (AIAI), Sep 2011, Corfu, Greece. pp.433-442, 10.1007/978-3-642-23957-1_48. hal-01571328 HAL Id: hal-01571328 https://hal.inria.fr/hal-01571328 Submitted on 2 Aug 2017 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Distributed under a Creative Commons Attribution| 4.0 International License Mathematical Models of Dynamic Behavior of Individual Neural Networks of Central Nervous System Dimitra – Despoina Pagania1, Adam Adamopoulos1,2 and Spiridon D. Likothanassis1 1 Pattern Recognition Laboratory, Dept. of Computer Engineering & Informatics, University of Patras, 26500, Patras, Greece 2Dept. of Medicine, Democritus University of Thrace, Alexandroupolis, Greece {pagania, likothan}@ceid.upatras.gr, [email protected] Abstract. We present mathematical models that describe individual neural networks of the Central Nervous System. Three cases are examined, varying in each case the values of the refractory period and the synaptic delay of a neuron.
    [Show full text]
  • Synaptic Mechanisms of Memory Consolidation During Sleep Slow Oscillations
    The Journal of Neuroscience, April 13, 2016 • 36(15):4231–4247 • 4231 Cellular/Molecular Synaptic Mechanisms of Memory Consolidation during Sleep Slow Oscillations X Yina Wei, Giri P. Krishnan, and Maxim Bazhenov Department of Cell Biology and Neuroscience, University of California at Riverside, Riverside, California 92521 Sleep is critical for regulation of synaptic efficacy, memories, and learning. However, the underlying mechanisms of how sleep rhythms contribute to consolidating memories acquired during wakefulness remain unclear. Here we studied the role of slow oscillations, 0.2–1 Hz rhythmic transitions between Up and Down states during stage 3/4 sleep, on dynamics of synaptic connectivity in the thalamocortical network model implementing spike-timing-dependent synaptic plasticity. We found that the spatiotemporal pattern of Up-state propa- gation determines the changes of synaptic strengths between neurons. Furthermore, an external input, mimicking hippocampal ripples, delivered to the cortical network results in input-specific changes of synaptic weights, which persisted after stimulation was removed. These synaptic changes promoted replay of specific firing sequences of the cortical neurons. Our study proposes a neuronal mechanism on how an interaction between hippocampal input, such as mediated by sharp wave-ripple events, cortical slow oscillations, and synaptic plasticity, may lead to consolidation of memories through preferential replay of cortical cell spike sequences during slow-wave sleep. Key words: cortex; hippocampus; learning and memory; memory consolidation; sleep slow oscillations; thalamus Significance Statement Sleep is critical for memory and learning. Replay during sleep of temporally ordered spike sequences related to a recent experience was proposed to be a neuronal substrate of memory consolidation.
    [Show full text]
  • Synaptic Weight Modulation and Adaptation Part I: Introduction and Presynaptic Mechanisms
    Synaptic Weight I Synaptic Weight Modulation and Adaptation Part I: Introduction and Presynaptic Mechanisms A Technical Strategy Brief for the Neurofuzzy Soft Computing Program R. Wells May 15, 2003 The Biological Basis of Synaptic Weight A “weight” in an artificial neuron model is meant to be a representation of the amount of effect that a synaptic input has on the overall response of the neuron. The neuron that generates the synaptic input (the “action potential” or AP) is called the “presynaptic neuron”; the neuron receiving the effect of the action potential is called the “postsynaptic neuron”. The synapse is the biological structure that connects these two neurons. As such, “the” synapse belongs to neither the presynaptic nor the postsynaptic neuron. The part of the synaptic structure contained within the cell of the presynaptic neuron is called the “presynaptic terminal”. The part of the synaptic structure contained within the cell of the postsynaptic neuron is called the “postsynaptic compartment”. Between these two is a small, liquid-filled gap called the “synaptic cleft.” Transmission from the presynaptic terminal to the postsynaptic compartment takes place by chemical signaling. The presynaptic terminal secretes particular molecules, called neurotransmitters (NTXs), into the cleft. These NTXs bind to receptor proteins on the surface of the postsynaptic compartment, and this chemical binding causes electro-chemical reactions to take place in the postsynaptic cell. Figure 1 illustrates the generic form of a synapse. The “weight” of a synapse can therefore be regarded as a measure of the amount of postsynaptic response that occurs when an action potential invades the presynaptic terminal.
    [Show full text]
  • A New Learning Method for Inference Accuracy, Core Occu- Pation, and Performance Co-Optimization on Truenorth Chip
    A New Learning Method for Inference Accuracy, Core Occu- pation, and Performance Co-optimization on TrueNorth Chip 1 Wei Wen, 1 Chunpeng Wu, 1 Yandan Wang, 1 Kent Nixon, 2 Qing Wu, 2 Mark Barnell, 1 Hai Li, 1 Yiran Chen 1 University of Pittsburgh, Pittsburgh, PA, USA; 2 Air Force Research Laboratory, Rome, NY, USA 1{wew57, chw127, yaw46, kwn2, hal66, yic52}@pitt.edu, 2{qing.wu.2, Mark.Barnell.1}@us.af.mil ABSTRACT ues. Besides the stochastic scheme, TrueNorth also supports many IBM TrueNorth chip uses digital spikes to perform neuromor- deterministic neural coding schemes (i.e., rate code, population phic computing and achieves ultrahigh execution parallelism and code, time-to-spike code and rank code [9]) in temporal domain to power efficiency. However, in TrueNorth chip, low quantization represent non-binary message using multiple binary spikes. Rate resolution of the synaptic weights and spikes significantly limits code, for example, uses the number of the spikes occurring within the inference (e.g., classification) accuracy of the deployed neural a predefined number of time steps to indicate the amplitude of a network model. Existing workaround, i.e., averaging the results signal [7]. However, this “official” workaround rapidly exhausts over multiple copies instantiated in spatial and temporal domains, hardware resources when the network scale increases and a large rapidly exhausts the hardware resources and slows down the number of spatial samples (copies) are instantiated. It also slows computation. In this work, we propose a novel learning method on down the computation when many spike samples are generated. TrueNorth platform that constrains the random variance of each In this work, we first formulate the standard TrueNorth learning computation copy and reduces the number of needed copies.
    [Show full text]
  • Drawing Inspiration from Biological Dendrites to Empower Artificial
    DRAWING INSPIRATION FROM BIOLOGICAL DENDRITES TO EMPOWER ARTIFICIAL NEURAL NETWORKS OPINION ARTICLE Spyridon Chavlis Institute of Molecular Biology and Biotechnology (IMBB) Foundation for Research and Technology Hellas (FORTH) Plastira Avenue 100, Heraklion, 70013, Greece [email protected] Panayiota Poirazi Institute of Molecular Biology and Biotechnology (IMBB) Foundation for Research and Technology Hellas (FORTH) Plastira Avenue 100, Heraklion, 70013, Greece [email protected] ABSTRACT This article highlights specific features of biological neurons and their dendritic trees, whose adop- tion may help advance artificial neural networks used in various machine learning applications. Advancements could take the form of increased computational capabilities and/or reduced power consumption. Proposed features include dendritic anatomy, dendritic nonlinearities, and compart- mentalized plasticity rules, all of which shape learning and information processing in biological networks. We discuss the computational benefits provided by these features in biological neurons and suggest ways to adopt them in artificial neurons in order to exploit the respective benefits in machine learning. Keywords Artificial Neural Networks · Biological dendrites · Plasticity 1 Introduction Artificial Neural Networks (ANN) implemented in Deep Learning (DL) architectures have been extremely successful in solving challenging machine learning (ML) problems such as image [1], speech [2], and face [3] recognition, arXiv:2106.07490v1 [q-bio.NC] 14 Jun 2021 playing online games [4], autonomous driving [5], etc. However, despite their huge successes, state-of-the-art DL architectures suffer from problems that seem rudimentary to real brains. For example, while becoming true experts in solving a specific task, they typically fail to transfer these skills to new problems without extensive retraining – a property known as “transfer learning”.
    [Show full text]
  • Effects of Synaptic Weight Diffusion on Learning in Decision Making Networks
    Effects of Synaptic Weight Diffusion on Learning in Decision Making Networks Kentaro Katahira1;2;3, Kazuo Okanoya1;3 and Masato Okada1;2;3 1ERATO Okanoya Emotional Information Project, Japan Science Technology Agency 2Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Chiba 277-8561, Japan 3RIKEN Brain Science Institute, Wako, Saitama 351-0198, Japan [email protected] [email protected] [email protected] Abstract When animals repeatedly choose actions from multiple alternatives, they can al- locate their choices stochastically depending on past actions and outcomes. It is commonly assumed that this ability is achieved by modifications in synaptic weights related to decision making. Choice behavior has been empirically found to follow Herrnstein’s matching law. Loewenstein & Seung (2006) demonstrated that matching behavior is a steady state of learning in neural networks if the synap- tic weights change proportionally to the covariance between reward and neural ac- tivities. However, their proof did not take into account the change in entire synap- tic distributions. In this study, we show that matching behavior is not necessarily a steady state of the covariance-based learning rule when the synaptic strength is sufficiently strong so that the fluctuations in input from individual sensory neu- rons influence the net input to output neurons. This is caused by the increasing variance in the input potential due to the diffusion of synaptic weights. This effect causes an undermatching phenomenon, which has been observed in many behav- ioral experiments. We suggest that the synaptic diffusion effects provide a robust neural mechanism for stochastic choice behavior.
    [Show full text]