检索结果-内蒙古大学图书馆

Parallel and hierarchical neural mechanisms for adaptive and predictive behavioral control

NEURAL NETWORKS 2021年 144卷 507-521页

作者： Macpherson, Tom Matsumoto, Masayuki Gomi, Hiroaki Morimoto, Jun Uchibe, Eiji Hikida, Takatoshi Osaka Univ Inst Prot Res Lab Adv Brain Funct 3-2 Yamadaoka Suita Osaka 5650871 Japan Univ Tsukuba Fac Med Div Biomed Sci Tsukuba Ibaraki Japan Nippon Telegraph & Tel Co NTT Commun Sci Labs Atsugi Kanagawa Japan ATR Computat Neurosci Labs Dept Brain Robot Interface Kyoto Japan Kyoto Univ Grad Sch Informat Kyoto Japan

Our brain can be recognized as a network of largely hierarchically organized neural circuits that operate to control specific functions, but when acting in parallel, enable the performance of complex and simultaneous behaviors. Indeed, many of our daily actions require concurrent information processing in sensorimotor, associative, and limbic circuits that are dynamically and hierarchically modulated by sensory information and previous learning. This organization of information processing in biological organisms has served as a major inspiration for artificial intelligence and has helped to create in silico systems capable of matching or even outperforming humans in several specific tasks, including visual recognition and strategy-based games. However, the development of human-like robots that are able to move as quickly as humans and respond flexibly in various situations remains a major challenge and indicates an area where further use of parallel and hierarchical architectures may hold promise. In this article we review several important neural and behavioral mechanisms organizing hierarchical and predictive processing for the acquisition and realization of flexible behavioral control. Then, inspired by the organizational features of brain circuits, we introduce a multi-timescale parallel and hierarchical learning framework for the realization of versatile and agile movement in humanoid robots. (c) 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://***/licenses/by-nc-nd/4.0/).

关键词： Parallel processing hierarchical processing Behavioral flexibility Movement control Artificial intelligence Humanoid robotics

来源：评论

学校读者我要写书评

暂无评论

Multiscale Gossip for Efficient Decentralized Averaging in Wireless Packet Networks

引用

IEEE TRANSACTIONS ON SIGNAL processing 2013年第9期61卷 2137-2149页

作者： Tsianos, Konstantinos I. Rabbat, Michael G. McGill Univ Dept Elect & Comp Engn Montreal PQ H3A 0E9 Canada

This paper describes and analyzes a hierarchical algorithm called Multiscale Gossip for solving the distributed average consensus problem in wireless sensor networks. The algorithm proceeds by recursively partitioning a given network. Initially, nodes at the finest scale gossip to compute local averages. Then, using multi-hop communication and geographic routing to communicate between nodes that are not directly connected, these local averages are progressively fused up the hierarchy until the global average is computed. We show that the proposed hierarchical scheme with k = Theta(log log n) levels of hierarchy is competitive with state-of-the-art randomized gossip algorithms in terms of message complexity, achieving epsilon-accuracy with high probability after O(n log log n log 1/epsilon) single-hop messages. Key to our analysis is the way in which the network is recursively partitioned. We find that the above scaling law is achieved when subnetworks at scale j contain O(n((2/3)j)) nodes;then the message complexity at any individual scale is O(n log 1/epsilon) Another important consequence of the hierarchical construction is that the longest distance over which messages are exchanged is O(n(1/3)) hops (at the highest scale), and most messages (at lower scales) travel shorter distances. In networks that use link-level acknowledgements, this results in less congestion and resource usage by reducing message retransmissions. Simulations illustrate that the proposed scheme is more efficient than state-of-the-art randomized gossip algorithms based on averaging along paths.

关键词： Consensus algorithms distributed signal processing hierarchical processing scaling laws wireless sensor networks

来源：评论

学校读者我要写书评

暂无评论

Re-Thinking the Organization of Cortico-Basal Ganglia-Thalamo-Cortical Loops

引用

COGNITIVE COMPUTATION 2024年第5期16卷 2405-2410页

作者： Baladron, Javier Hamker, Fred H. Tech Univ Chemnitz Dept Comp Sci Str Nationen 62 D-09107 Chemnitz Germany Univ Santiago Chile Dept Ingn Informat Santiago Chile

First concepts of cortex-basal ganglia interactions suggested that cognitive functions are implemented through different parallel, segregated cortico-basal ganglia-cortical loops. Recent evidence however shows that there are at least 4 ways by which different loops could interact: overlapping cortico-pallidal projections, overlapping cortico-striatal projections, striato-nigro-striatal spirals, and cortico-thalamo-striatal projections. We propose that current evidence, if incorporated into neuro-computational models, provides new avenues for explaining cognitive functions. Using a recently introduced hierarchical neuro-computational model of multiple cortico-basal ganglia-cortical loops, we exemplify how interaction between loops can explain behavioral data linked to the ideomotor theory. During training of a stimulus-response task, a task-irrelevant tone is played after each action. Then, during a test period, the same tones are played but together with the stimuli. The model learns a distributed representation of action outcomes which is then used to select actions. Similar to observations in human subjects, the model's response time is larger if an inconsistent tone (previously listened together with an alternative action) is played with the stimulus. Understanding the function of different communication strategies between loops could be the key to fully unravel the neural basis of the numerous functions supported by the basal ganglia. The multiple-loop structure allows for a more complex representation of behavior in which action outcomes are considered.

关键词： Cognition Basal ganglia hierarchical processing Ideomotor theory

来源：评论

学校读者我要写书评

暂无评论

hierarchical-Task Reservoir for Online Semantic Analysis From Continuous Speech

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2022年第6期33卷 2654-2663页

作者： Pedrelli, Luca Hinaut, Xavier INRIA Bordeaux Sud Ouest F-33405 Bordeaux France CNRS Bordeaux INP LaBRI UMR 5800 Paris France Univ Bordeaux CNRS Inst Malad Neurodegenerat UMR 5293 Bordeaux France

In this article, we propose a novel architecture called hierarchical-task reservoir (HTR) suitable for real-time applications for which different levels of abstraction are available. We apply it to semantic role labeling (SRL) based on continuous speech recognition. Taking inspiration from the brain, this demonstrates the hierarchies of representations from perceptive to integrative areas, and we consider a hierarchy of four subtasks with increasing levels of abstraction (phone, word, part-of-speech (POS), and semantic role tags). These tasks are progressively learned by the layers of the HTR architecture. Interestingly, quantitative and qualitative results show that the hierarchical-task approach provides an advantage to improve the prediction. In particular, the qualitative results show that a shallow or a hierarchical reservoir, considered as baselines, does not produce estimations as good as the HTR model would. Moreover, we show that it is possible to further improve the accuracy of the model by designing skip connections and by considering word embedding (WE) in the internal representations. Overall, the HTR outperformed the other state-of-the-art reservoir-based approaches and it resulted in extremely efficient with respect to typical recurrent neural networks (RNNs) in deep learning (DL) [e.g., long short term memory (LSTMs)]. The HTR architecture is proposed as a step toward the modeling of online and hierarchical processes at work in the brain during language comprehension.

关键词： Computer architecture Brain modeling Task analysis Reservoirs Speech recognition Semantics Computational modeling Anytime process hierarchical processing hierarchical reservoir computing (HRC) natural language processing (NLP) part-of-speech (POS) POS tagging recurrent neural networks (RNNs) semantic role labeling (SRL) speech recognition

来源：评论

学校读者我要写书评

暂无评论

Early visual cortical responses produced by checkerboard pattern stimulation

引用

NEUROIMAGE 2016年 134卷 532-539页

作者： Shigihara, Yoshihito Hoshi, Hideyuki Zeki, Semir UCL Wellcome Lab Neurobiol Gower St London WC1E 6BT England

Visual evoked potentials have been traditionally triggered with flash or reversing checkerboard stimuli and recorded with electroencephalographic techniques, largely but not exclusively in clinical or clinically related settings. They have been crucial in determining the healthy functioning or otherwise of the visual pathways up to and including the cerebral cortex. They have typically given early response latencies of 100 ms, the source of which has been attributed to V1, with the prestriate cortex being secondarily activated somewhat later. On the other hand, magnetoencephalographic studies using stimuli better tailored to the physiology of individual, specialized, visual areas have given early latencies of <50 ms with the sources localized in both striate (V1) and prestriate cortex. In this study, we used the reversing checkerboard pattern as a stimulus and recorded cortical visual evoked magnetic fields with magnetoencephalography, to establish whether very early responses can be traced to (estimated) in both striate and prestriate cortex, since such a demonstration would enhance considerably the power of this classical approach in clinical investigations. Our results show that cortical responses evoked by checkerboard patterns can be detected before 50 ms post-stimulus onset and that their sources can be estimated in both striate and prestriate cortex, suggesting a strong parallel input from the sub-cortex to both striate and prestriate divisions of the visual cortex. (C) 2016 Published by Elsevier Inc.

关键词： Magnetoencephalography Parallel processing Prestriate cortex V1 hierarchical processing

来源：评论

学校读者我要写书评

暂无评论

Optimizing the imaging of the monkey auditory cortex: sparse vs. continuous fMRI

引用

MAGNETIC RESONANCE IMAGING 2009年第8期27卷 1065-1073页

作者： Petkov, Christopher I. Kayser, Christoph Augath, Mark Logothetis, Nikos K. Max Planck Inst Biol Cybernet D-72076 Tubingen Germany Newcastle Univ Inst Neurosci Newcastle Upon Tyne NE2 4HH Tyne & Wear England Univ Manchester Manchester N13 9PT Lancs England

The noninvasive imaging of the monkey auditory system with functional magnetic resonance imaging (fMRI) can bridge the gap between electrophysiological Studies in monkeys and imaging studies in humans. Some of the recent imaging of monkey auditory cortical and subcortical structures relies on a technique of "sparse imaging," which was developed in human Studies to sidestep the negative influence of scanner noise by adding periods of silence in between volume acquisition. Among the various aspects that have gone into the ongoing optimization of fMRI of the monkey auditory cortex, replacing the more common continuous-imaging paradigm with sparse irnaging seemed to LIS to make the most obvious difference in the amount of activity that we could reliably obtain from awake or anesthetized animals. Here. we directly compare the sparse- and continuous-imaging paradigms in anesthetized animals. We document a strikingly greater auditory response with sparse irnaging, both quantitatively and qualitatively, which includes a more expansive and robust tonotopic organization. There were instances where continuous imaging Could better reveal organizational properties that sparse irnaging missed, Such as aspects of the hierarchical organization of auditory cortex. We consider the choice of imaging paradigm its a key component in optimizing the fMRI of the monkey auditory cortex. (C) 2009 Elsevier Inc. All rights reserved.

关键词： Monkey Macaque Auditory Cortex Tonotopy hierarchical processing Human Communication Hearing Sparse Continuous Imaging Sequence Paradigm

来源：评论

学校读者我要写书评

暂无评论

Neural mechanism of dynamic responses of neurons in inferior temporal cortex in face perception

引用

COGNITIVE NEURODYNAMICS 2013年第1期7卷 23-38页

作者： Yamada, Yuichiro Kashimori, Yoshiki Univ Electrocommun Grad Sch Informat Syst Chofu Tokyo 1828585 Japan Univ Electrocommun Dept Engn Sci Chofu Tokyo 1828585 Japan

Understanding the neural mechanisms of object and face recognition is one of the fundamental challenges of visual neuroscience. The neurons in inferior temporal (IT) cortex have been reported to exhibit dynamic responses to face stimuli. However, little is known about how the dynamic properties of IT neurons emerge in the face information processing. To address this issue, we made a model of IT cortex, which performs face perception via an interaction between different IT networks. The model was based on the face information processed by three resolution maps in early visual areas. The network model of IT cortex consists of four kinds of networks, in which the information about a whole face is combined with the information about its face parts and their arrangements. We show here that the learning of face stimuli makes the functional connections between these IT networks, causing a high spike correlation of IT neuron pairs. A dynamic property of subthreshold membrane potential of IT neuron, produced by Hodgkin-Huxley model, enables the coordination of temporal information without changing the firing rate, providing the basis of the mechanism underlying face perception. We show also that the hierarchical processing of face information allows IT cortex to perform a "coarse-to-fine" processing of face information. The results presented here seem to be compatible with experimental data about dynamic properties of IT neurons.

关键词： Inferior temporal cortex Dynamic response hierarchical processing Face perception Neural model

来源：评论

学校读者我要写书评

暂无评论

Multiple stages of auditory speech perception reflected in event-related fMRI

引用

CEREBRAL CORTEX 2007年第10期17卷 2251-2257页

作者： Obleser, Jonas Zimmermann, Jonas Van Meter, John Rauschecker, Josef P. Georgetown Univ Sch Med Washington DC 20057 USA

Speech processing in auditory cortex and beyond is a remarkable yet poorly understood faculty of the listening brain. Here we show that stop consonants, as the most transient constituents of speech, are sufficient to involve speech perception circuits in the human superior temporal cortex. Left anterolateral superior temporal cortex showed a stronger response in blood oxygenation level-dependent functional magnetic resonance imaging (fMRI) to intelligible consonantal bursts compared with incomprehensible control sounds matched for spectrotemporal complexity. Simultaneously, the left posterior superior temporal plane (including planum temporale [PT]) exhibited a noncategorical responsivity to complex stimuluscoustics across all trials, showing no preference for intelligible peech sounds. Multistage hierarchical processing of speech sounds is thus revealed with fMRI, providing evidence for a role of the PT in the fundamental stages of the acoustic analysis of complex sounds, including speech.

关键词： auditory cortex consonants fMRI hierarchical processing planum temporale speech

来源：评论

学校读者我要写书评

暂无评论

Altered visual character and object recognition in Japanese-speaking adolescents with developmental dyslexia

引用

NEUROSCIENCE LETTERS 2020年 723卷 134841页

作者： Higuchi, Hiroki Iwaki, Sunao Uno, Akira Natl Inst Adv Ind Sci & Technol Automot Human Factors Res Ctr 1-1-1 Higashi Tsukuba Ibaraki 3058566 Japan Univ Tsukuba Fac Human Sci Lab Adv Res D 1-1-1 Tennodai Tsukuba Ibaraki 3058577 Japan NTT Commun Sci Labs 2-4 HikaridaiSeika Cho Kyoto 6190237 Japan

Many studies have confirmed a brain dysfunction in people with developmental dyslexia (DD) in certain brain regions, including the left superior temporal gyrus and the left fusiform gyrus. However, the neurobiological substrates in Japanese-speaking people with dyslexia are not fully understood, mostly due to the uniqueness of the orthographic systems. Since a substantial part of the written Japanese includes the logographic Kanji as well as the phonographic Kana systems, the reading disability might be caused not only by a dysfunction in the phonological system, but also by a dysfunction in the visual recognition system. Previous studies reported altered hierarchical visual word form processing in the left occipitotemporal cortex;however, it remains unclear whether the altered hierarchical visual processing is language stimuli-specific. Therefore, we aimed to investigate whether (a) Japanese-speaking individuals with DD exhibit atypical hierarchical visual processing, and if so, (b) whether the altered hierarchical visual processing is language stimuli-specific or not. The present study investigated the brain activation pattern for the hierarchical component of the Kanji characters and object stimuli in typically developing (TD) adolescents and adolescents with DD using functional magnetic resonance imaging. For the Kanji characters, adolescents with DD showed a greater activation in the left occipital gyrus and right occipital fusiform gyrus, and this hyperactivity was also found for pseudo and artificial Kanji characters. These results imply reliance on an early visual system in Kanji reading in Japanese-speaking adolescents with DD. Additionally, we also investigated the brain activity for object stimuli, and adolescents with DD showed a greater activation in the bilateral occipital gyri compared with the TD adolescents. These results imply an altered hierarchical visual processing characterized by overactivation in the early visual areas, which is a not restric

关键词： Developmental dyslexia Visual recognition Japanese Kanji character Object hierarchical processing

来源：评论

学校读者我要写书评

暂无评论

A revival of Homo loquens as a builder of labeled structures: Neurocognitive considerations

引用

NEUROSCIENCE AND BIOBEHAVIORAL REVIEWS 2017年第PT B期81卷 213-224页

作者： Goucha, T. Zaccarella, E. Friederici, A. D. Max Planck Inst Human Cognit & Brain Sci Dept Neuropsychol Leipzig Germany Humboldt Univ Berlin Sch Mind & Brain Berlin Germany

The core capacity of human language is described as the faculty to combine words into hierarchical structures. This review aims to isolate the fundamental computation behind the language faculty together with its neural implementation. First, we present our central hypothesis by confronting recent linguistic theory with evolutionary arguments: linguistic humaniqueness is reflected in the labeling of word combinations forming asymmetric hierarchical structures. Second, we review the neurolinguistic literature, especially focusing on dual-stream connectivity models. We put forward that the dorsal pathway, especially the arcuate fascicle, is responsible for the rule-based combinatorial system, implementing labeling and giving rise to hierarchical structures. Conversely, the ventral stream is rather responsible for semantic associative operations. We further present evolutionary neuroanatomical evidence grounding our hypothesis. We conclude by suggesting further avenues of research as well as open questions to be addressed. With the aim to expand our knowledge on the neurobiology of language, we hope to provide a testable hypothesis for the origin of language syntax bringing together evidence from different fields. (C) 2017 Elsevier Ltd. All rights reserved.

关键词： Labeling Language evolution Syntax hierarchical processing Arcuate fascicle

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：