scholarly journals Spiking Recurrent Networks as a Model to Probe Neuronal Timescales Specific to Working Memory

2019 ◽  
Author(s):  
Robert Kim ◽  
Terrence J. Sejnowski
Keyword(s):  
Working Memory ◽  
Recurrent Neural Networks ◽  
Cortical Neurons ◽  
Receptive Fields ◽  
Hierarchical Organization ◽  
Recurrent Networks ◽  
Multiple Timescales ◽  
Underlying Mechanisms ◽  
Experimental Findings ◽  
Long Timescales

AbstractCortical neurons process and integrate information on multiple timescales. In addition, these timescales or temporal receptive fields display functional and hierarchical organization. For instance, areas important for working memory (WM), such as prefrontal cortex, utilize neurons with stable temporal receptive fields and long timescales to support reliable representations of stimuli. Despite of the recent advances in experimental techniques, the underlying mechanisms for the emergence of neuronal timescales long enough to support WM are unclear and challenging to investigate experimentally. Here, we demonstrate that spiking recurrent neural networks (RNNs) designed to perform a WM task reproduce previously observed experimental findings and that these models could be utilized in the future to study how neuronal timescales specific to WM emerge.

scholarly journals Strong inhibitory signaling underlies stable temporal dynamics and working memory in spiking neural networks

2020 ◽  
Author(s):  
Robert Kim ◽  
Terrence J. Sejnowski
Keyword(s):  
Working Memory ◽  
Task Performance ◽  
Cortical Neurons ◽  
Temporal Dynamics ◽  
Single Neurons ◽  
Multiple Timescales ◽  
Neural Data ◽  
Temporal Properties ◽  
Underlying Mechanisms ◽  
Inhibitory Signaling

AbstractCortical neurons process information on multiple timescales, and areas important for working memory (WM) contain neurons capable of integrating information over a long timescale. However, the underlying mechanisms for the emergence of neuronal timescales stable enough to support WM are unclear. By analyzing a spiking recurrent neural network (RNN) model trained on a WM task and activity of single neurons in the primate prefrontal cortex, we show that the temporal properties of our model and the neural data are remarkably similar. Dissecting our RNN model revealed strong inhibitory-to-inhibitory connections underlying a disinhibitory microcircuit as a critical component for long neuronal timescales and WM maintenance. We also found that enhancing inhibitory-to-inhibitory connections led to more stable temporal dynamics and improved task performance. Finally, we show that a network with such microcircuitry can perform other tasks without disrupting its pre-existing timescale architecture, suggesting that strong inhibitory signaling underlies a flexible WM network.

Adult plasticity in the visual system

1995 ◽  
Vol 73 (9) ◽  
pp. 1323-1338 ◽  
Author(s):  
Yuzo M. Chino
Keyword(s):  
Visual Cortex ◽  
Cortical Neurons ◽  
Time Course ◽  
Receptive Fields ◽  
Afferent Input ◽  
Contrast Threshold ◽  
Topographic Map ◽  
Retinal Lesions ◽  
Underlying Mechanisms ◽  
Adult Plasticity

When visual cortical neurons in adult mammals are deprived of their normal afferent input from retinae, they are capable of acquiring new receptive fields by modifying the effectiveness of existing intrinsic connections, a basis for topographic map reorganization. To gain insights into the underlying mechanisms and functional significance of this adult plasticity, we measured the spatial limits and time course of retinotopic map reorganization. We also determined whether reactivated neurons exhibit normal receptive field properties. We found that virtually all units in the denervated zone of cortex acquired new receptive fields (i.e., there were no silent areas in the cortex) and map reorganization can take place within hours of deafferentation provided that retinal lesions are relatively small (<5°). Furthermore, after long periods of recovery, reactivated units exhibited strikingly normal selectivity to stimulus orientation, direction of movement, and spatial frequency if relatively high contrast stimuli were used. However, responsiveness of these neurons, measured in terms of the maximum response amplitude and the contrast threshold, was clearly reduced. Thus, contrary to traditional belief, the adult visual cortex is capable of exhibiting considerable plasticity, and reactivated neurons are capable of contributing to an analysis of a visual scene.Key words: adult plasticity, visual cortex, retinal lesions, map reorganization, cat.

scholarly journals Information diversity in individual auditory cortical neurons is associated with functionally distinct coordinated neuronal ensembles

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Jermyn Z. See ◽  
Natsumi Y. Homma ◽  
Craig A. Atencio ◽  
Vikaas S. Sohal ◽  
Christoph E. Schreiner
Keyword(s):  
Cortical Neurons ◽  
Single Neuron ◽  
Receptive Fields ◽  
Acoustic Feature ◽  
Neuronal Ensembles ◽  
Neuron Spike ◽  
Information Diversity ◽  
Feature Selectivity ◽  
Stimulus Features ◽  
Time Specific

AbstractNeuronal activity in auditory cortex is often highly synchronous between neighboring neurons. Such coordinated activity is thought to be crucial for information processing. We determined the functional properties of coordinated neuronal ensembles (cNEs) within primary auditory cortical (AI) columns relative to the contributing neurons. Nearly half of AI cNEs showed robust spectro-temporal receptive fields whereas the remaining cNEs showed little or no acoustic feature selectivity. cNEs can therefore capture either specific, time-locked information of spectro-temporal stimulus features or reflect stimulus-unspecific, less-time specific processing aspects. By contrast, we show that individual neurons can represent both of those aspects through membership in multiple cNEs with either high or absent feature selectivity. These associations produce functionally heterogeneous spikes identifiable by instantaneous association with different cNEs. This demonstrates that single neuron spike trains can sequentially convey multiple aspects that contribute to cortical processing, including stimulus-specific and unspecific information.

scholarly journals Attention-based hierarchical recurrent neural networks for MOOC forum posts analysis

2020 ◽  
Author(s):  
Nicola Capuano ◽  
Santi Caballé ◽  
Jordi Conesa ◽  
Antonio Greco
Keyword(s):  
Neural Networks ◽  
Recurrent Neural Networks ◽  
Text Categorization ◽  
Online Courses ◽  
Recurrent Network ◽  
Subject Area ◽  
Conversational Agents ◽  
Collaborative Tools ◽  
Experimental Findings ◽  
Productive Talk

AbstractMassive open online courses (MOOCs) allow students and instructors to discuss through messages posted on a forum. However, the instructors should limit their interaction to the most critical tasks during MOOC delivery so, teacher-led scaffolding activities, such as forum-based support, can be very limited, even impossible in such environments. In addition, students who try to clarify the concepts through such collaborative tools could not receive useful answers, and the lack of interactivity may cause a permanent abandonment of the course. The purpose of this paper is to report the experimental findings obtained evaluating the performance of a text categorization tool capable of detecting the intent, the subject area, the domain topics, the sentiment polarity, and the level of confusion and urgency of a forum post, so that the result may be exploited by instructors to carefully plan their interventions. The proposed approach is based on the application of attention-based hierarchical recurrent neural networks, in which both a recurrent network for word encoding and an attention mechanism for word aggregation at sentence and document levels are used before classification. The integration of the developed classifier inside an existing tool for conversational agents, based on the academically productive talk framework, is also presented as well as the accuracy of the proposed method in the classification of forum posts.

scholarly journals On the structural connectivity of large-scale models of brain networks at cellular level

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Giuseppe Giacopelli ◽  
Domenico Tegolo ◽  
Emiliano Spera ◽  
Michele Migliore
Keyword(s):  
Large Scale ◽  
Brain Networks ◽  
Structural Connectivity ◽  
Network Models ◽  
Brain Regions ◽  
Cellular Level ◽  
Scale Model ◽  
Mathematical Framework ◽  
Underlying Mechanisms ◽  
Experimental Findings

AbstractThe brain’s structural connectivity plays a fundamental role in determining how neuron networks generate, process, and transfer information within and between brain regions. The underlying mechanisms are extremely difficult to study experimentally and, in many cases, large-scale model networks are of great help. However, the implementation of these models relies on experimental findings that are often sparse and limited. Their predicting power ultimately depends on how closely a model’s connectivity represents the real system. Here we argue that the data-driven probabilistic rules, widely used to build neuronal network models, may not be appropriate to represent the dynamics of the corresponding biological system. To solve this problem, we propose to use a new mathematical framework able to use sparse and limited experimental data to quantitatively reproduce the structural connectivity of biological brain networks at cellular level.

scholarly journals Tactile orientation perception: an ideal observer analysis of human psychophysical performance in relation to macaque area 3b receptive fields

2015 ◽  
Vol 114 (6) ◽  
pp. 3076-3096 ◽  
Author(s):  
Ryan M. Peters ◽  
Phillip Staibano ◽  
Daniel Goldreich
Keyword(s):  
Cortical Neurons ◽  
Human Performance ◽  
Spatial Perception ◽  
Receptive Fields ◽  
Human Perception ◽  
Ideal Observer ◽  
Orientation Discrimination ◽  
Orientation Perception ◽  
Area 3B ◽  
The Ideal

The ability to resolve the orientation of edges is crucial to daily tactile and sensorimotor function, yet the means by which edge perception occurs is not well understood. Primate cortical area 3b neurons have diverse receptive field (RF) spatial structures that may participate in edge orientation perception. We evaluated five candidate RF models for macaque area 3b neurons, previously recorded while an oriented bar contacted the monkey's fingertip. We used a Bayesian classifier to assign each neuron a best-fit RF structure. We generated predictions for human performance by implementing an ideal observer that optimally decoded stimulus-evoked spike counts in the model neurons. The ideal observer predicted a saturating reduction in bar orientation discrimination threshold with increasing bar length. We tested 24 humans on an automated, precision-controlled bar orientation discrimination task and observed performance consistent with that predicted. We next queried the ideal observer to discover the RF structure and number of cortical neurons that best matched each participant's performance. Human perception was matched with a median of 24 model neurons firing throughout a 1-s period. The 10 lowest-performing participants were fit with RFs lacking inhibitory sidebands, whereas 12 of the 14 higher-performing participants were fit with RFs containing inhibitory sidebands. Participants whose discrimination improved as bar length increased to 10 mm were fit with longer RFs; those who performed well on the 2-mm bar, with narrower RFs. These results suggest plausible RF features and computational strategies underlying tactile spatial perception and may have implications for perceptual learning.

scholarly journals Recurrent Neural Networks for Narrowband Signal Detection in the Time-Frequency Domain

2004 ◽  
Vol 213 ◽  
pp. 483-486
Author(s):  
David Brodrick ◽  
Douglas Taylor ◽  
Joachim Diederich
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Signal Detection ◽  
Frequency Domain ◽  
Recurrent Neural Networks ◽  
Radio Frequency Interference ◽  
Recurrent Networks ◽  
Time Frequency ◽  
Narrowband Signal ◽  
Radio Signals

A recurrent neural network was trained to detect the time-frequency domain signature of narrowband radio signals against a background of astronomical noise. The objective was to investigate the use of recurrent networks for signal detection in the Search for Extra-Terrestrial Intelligence, though the problem is closely analogous to the detection of some classes of Radio Frequency Interference in radio astronomy.

Recurrent Neural Networks with Small Weights Implement Definite Memory Machines

2003 ◽  
Vol 15 (8) ◽  
pp. 1897-1929 ◽  
Author(s):  
Barbara Hammer ◽  
Peter Tiňo
Keyword(s):  
Neural Networks ◽  
Recurrent Neural Networks ◽  
Experimental Studies ◽  
Activation Function ◽  
Transition Function ◽  
Recurrent Network ◽  
Point Of View ◽  
Recurrent Networks ◽  
Contraction Parameter ◽  
Arbitrary Precision

Recent experimental studies indicate that recurrent neural networks initialized with “small” weights are inherently biased toward definite memory machines (Tiňno, Čerňanský, & Beňušková, 2002a, 2002b). This article establishes a theoretical counterpart: transition function of recurrent network with small weights and squashing activation function is a contraction. We prove that recurrent networks with contractive transition function can be approximated arbitrarily well on input sequences of unbounded length by a definite memory machine. Conversely, every definite memory machine can be simulated by a recurrent network with contractive transition function. Hence, initialization with small weights induces an architectural bias into learning with recurrent neural networks. This bias might have benefits from the point of view of statistical learning theory: it emphasizes one possible region of the weight space where generalization ability can be formally proved. It is well known that standard recurrent neural networks are not distribution independent learnable in the probably approximately correct (PAC) sense if arbitrary precision and inputs are considered. We prove that recurrent networks with contractive transition function with a fixed contraction parameter fulfill the so-called distribution independent uniform convergence of empirical distances property and hence, unlike general recurrent networks, are distribution independent PAC learnable.

Abnormal auditory N400 in a case of zolpidem dependence, during a working memory test

2006 ◽  
Vol 21 (2) ◽  
pp. 135-137 ◽  
Author(s):  
Ioannis A. Liappas ◽  
Charalabos C. Papageorgiou ◽  
Andreas D. Rabavilas
Keyword(s):  
Working Memory ◽  
Spatial Working Memory ◽  
Event Related Potentials ◽  
Short Term ◽  
Context Processing ◽  
Gaba A ◽  
Related Potentials ◽  
N400 Component ◽  
Underlying Mechanisms ◽  
Term Management

AbstractZolpidem is a GABA (A) agonist, which is indicated for the short-term management of insomnia. Recent research provide evidence suggesting that zolpidem produces spatial working memory (WM) deficits and dependence; however, the underlying mechanisms of these effects are unknown. Since the auditory N400 component of event-related potentials (ERPS) is considered as an index of memory use of context processing, the present study focused on N400 waveform of ERPs elicited during a WM task in a case suffering from zolpidem dependence. The patterns of N400 waveform of this case were compared to the patterns obtained from healthy controls. This comparison revealed that zolpidem dependence is accompanied by reduced amplitudes located at posterior brain areas and diffuse prolongation of N400. These findings may indicate that zolpidem dependence manifests alterations with regard to the memory use of context processing, involving or affecting a wide-ranging network of the brain's structures.

Two-dimensional receptive-field organization in striate cortical neurons of the cat

1994 ◽  
Vol 11 (4) ◽  
pp. 703-720 ◽  
Author(s):  
Ming Sun ◽  
A. B. Bonds
Keyword(s):  
Receptive Field ◽  
Cortical Neurons ◽  
Ocular Dominance ◽  
Receptive Fields ◽  
Field Size ◽  
Two Dimensional ◽  
Cortical Layers ◽  
Field Organization ◽  
Receptive Field Organization ◽  
Cortical Receptive Fields

AbstractThe two-dimensional organization of receptive fields (RFs) of 44 cells in the cat visual cortex and four cells from the cat LGN was measured by stimulation with either dots or bars of light. The light bars were presented in different positions and orientations centered on the RFs. The RFs found were arbitrarily divided into four general types: Punctate, resembling DOG filters (11%); those resembling Gabor filters (9%); elongate (36%); and multipeaked-type (44%). Elongate RFs, usually found in simple cells, could show more than one excitatory band or bifurcation of excitatory regions. Although regions inhibitory to a given stimulus transition (e.g. ON) often coincided with regions excitatory to the opposite transition (e.g. OFF), this was by no means the rule. Measurements were highly repeatable and stable over periods of at least 1 h. A comparison between measurements made with dots and with bars showed reasonable matches in about 40% of the cases. In general, bar-based measurements revealed larger RFs with more structure, especially with respect to inhibitory regions. Inactivation of lower cortical layers (V-VI) by local GABA injection was found to reduce sharpness of detail and to increase both receptive-field size and noise in upper layer cells, suggesting vertically organized RF mechanisms. Across the population, some cells bore close resemblance to theoretically proposed filters, while others had a complexity that was clearly not generalizable, to the extent that they seemed more suited to detection of specific structures. We would speculate that the broadly varying forms of cat cortical receptive fields result from developmental processes akin to those that form ocular-dominance columns, but on a smaller scale.

Sign in / Sign up

Export Citation Format

Share Document