山﨑 匡
| 情報・ネットワーク工学専攻 | 准教授 |
| Ⅰ類（情報系） | 准教授 |
| 脳・医工学研究センター | 准教授 |
| 人工知能先端研究センター | 准教授 |
研究者情報
委員歴
- 2020年05月 - 現在
「革新脳」プロジェクト データベース有識者委員会 - 2020年01月 - 現在
一般社団法人 日本神経回路学会 理事, 学協会 - 2016年01月 - 現在
機関紙編集委員, 日本神経回路学会, 学協会 - 2016年08月 - 2020年06月
総務省「次世代人工知能技術の研究開発」評価検討委員会 - 2019年11月 - 2020年02月
External expert evaluator for Calls for Expression of Interest, the Human Brain Project - 2020年
External experts for EBRAINS Infrastructure Allocation Committee, the Human Brain Project - 2020年
日本神経科学学会 情報基盤整備委員会 - 2015年12月25日 - 2016年07月
総務省情報通信審議会 AI・脳研究WG, 政府 - 2015年
電子情報通信学会 ニューロコンピューティング研究専門委員会, 電子情報通信学会 ニューロコンピューティング研究専門委員会, 学協会
研究活動情報
受賞
論文
- Discrimination and learning of temporal input sequences in a cerebellar Purkinje cell model
Kaaya Tamura; Yuki Yamamoto; Taira Kobayashi; Rin Kuriyama; Tadashi Yamazaki
責任著者, Frontiers in Cellular Neuroscience, Frontiers Media SA, 17, 2023年02月02日, 査読付, 10.3389/fncel.2023.1075005, Introduction
Temporal information processing is essential for sequential contraction of various muscles with the appropriate timing and amplitude for fast and smooth motor control. These functions depend on dynamics of neural circuits, which consist of simple neurons that accumulate incoming spikes and emit other spikes. However, recent studies indicate that individual neurons can perform complex information processing through the nonlinear dynamics of dendrites with complex shapes and ion channels. Although we have extensive evidence that cerebellar circuits play a vital role in motor control, studies investigating the computational ability of single Purkinje cells are few.
Methods
We found, through computer simulations, that a Purkinje cell can discriminate a series of pulses in two directions (from dendrite tip to soma, and from soma to dendrite), as cortical pyramidal cells do. Such direction sensitivity was observed in whatever compartment types of dendrites (spiny, smooth, and main), although they have dierent sets of ion channels.
Results
We found that the shortest and longest discriminable sequences lasted for 60 ms (6 pulses with 10 ms interval) and 4,000 ms (20 pulses with 200 ms interval), respectively. and that the ratio of discriminable sequences within the region of the interesting parameter space was, on average, 3.3% (spiny), 3.2% (smooth), and 1.0% (main). For the direction sensitivity, a T-type Ca2+ channel was necessary, in contrast with cortical pyramidal cells that have N-methyl-D-aspartate receptors (NMDARs). Furthermore, we tested whether the stimulus direction can be reversed by learning, specifically by simulated long-term depression, and obtained positive results.
Discussion
Our results show that individual Purkinje cells can perform more complex information processing than is conventionally assumed for a single neuron, and suggest that Purkinje cells act as sequence discriminators, a useful role in motor control and learning.
研究論文（学術雑誌） - Testing an explicit method for multi-compartment neuron model simulation on a GPU
Taira Kobayashi; Rin Kuriyama; Tadashi Yamazaki
Cognitive Computation, Springer, 15, 1118, 1131, 2023年, 査読付, 10.1007/s12559-021-09942-6, Large-scale simulation of multi-compartment models is important for understanding the role of morphological structures of individual neurons for information processing in the brain. In a simulation, partial differential equations (PDEs) that describe the dynamics of neurons have to be solved numerically for each time step. To solve PDEs, numerical methods called implicit methods are used for stability. Implicit methods need to solve simultaneous equations, which can make numerical simulation slow on graphics processing units (GPUs) hardware accelerators for parallel computing. To overcome this problem, we investigated the use of explicit methods for multi-compartment model simulation. We applied the Runge-Kutta-Chebyshev (RKC) method to several cerebellar neuron models including Purkinje cells, granule cells, Golgi cells, and inferior olive cells. Next, we implemented a cerebellar cortical model composed of granule cells, Golgi cells, and Purkinje cells, while using different numerical methods for different cell types. Although explicit methods can be unstable against PDEs, using the RKC method showed sufficient stability for most cases, better computational performance than implicit methods on a GPU, and good reproducibility. In the network simulation, choosing the suitable numerical methods for each cell type achieved faster simulation than that used an implicit method solely. Our results suggest that explicit methods are applicable to multi-compartment models and can accelerate computational speed of simulations. Furthermore, to conduct large-scale simulation of multi-compartment models, choosing efficient numerical methods will be more important.
研究論文（学術雑誌）, 英語 - Deploying and Optimizing Embodied Simulations of Large Scale Spiking Neural Networks on HPC Infrastructure
Benedikt Feldotto; Jochen Martin Eppler; Cristian Jimenez-Romero; Christopher Bignamini; Carlos Enrique Gutierrez; Ugo Albanese; Eloy Retamino; Viktor Vorobev; Vahid Zolfaghari; Alex Upton; Zhe Sun; Hiroshi Yamaura; Morteza Heidarinejad; Wouter Klijn; Abigail Morrison; Felipe Cruz; Colin McMurtrie; Alois C. Knoll; Jun Igarashi; Tadashi Yamazaki; Kenji Doya; Fabrice O. Morin
Frontiers in Neuroinformatics, 16, 884180, 2022年05月, 査読付
研究論文（学術雑誌）, 英語 - Human-scale Brain Simulation via Supercomputer: A Case Study on the Cerebellum
Tadashi Yamazaki; Jun Igarashi; Hiroshi Yamaura
Neuroscience, Elsevier, 462, 235, 246, 2021年01月20日, 査読付
研究論文（学術雑誌）, 英語 - Real-Time Simulation of a Cerebellar Scaffold Model on Graphics Processing Units
Rin Kuriyama; Claudia Casellato; Egidio D'Angelo; Tadashi Yamazaki
Frontiers in Cellular Neuroscience, 15, 623552, 2021年, 査読付
研究論文（学術雑誌）, 英語 - Simulation of a human-scale cerebellar network model on the K computer
Hiroshi Yamaura; Jun Igarashi; Tadashi Yamazaki
Frontiers in Neuroinformatics, 14, 16, 2020年, 査読付
研究論文（学術雑誌）, 英語 - A Pathological Condition Affects Motor Modules in a Bipedal Locomotion Model
Daisuke Ichimura; Tadashi Yamazaki
Frontiers in Neurorobotics, 13, 79, 79, 2019年, 査読付, 国際誌, 10.3389/fnbot.2019.00079, Bipedal locomotion is a basic motor activity that requires simultaneous control of multiple muscles. Physiological experiments suggest that the nervous system controls bipedal locomotion efficiently by using motor modules of synergistic muscle activations. If these modules were merged, abnormal locomotion patterns would be realized as observed in patients with neural impairments such as chronic stroke. However, sub-acute patients have been reported not to show such merged motor modules. Therefore, in this study, we examined what conditions in the nervous system merges motor modules. we built a two-dimensional bipedal locomotion model that included a musculoskeletal model with 7 segments and 18 muscles, a neural system with a hierarchical central pattern generator (CPG), and various feedback inputs from reflex organs. The CPG generated synergistic muscle activations comprising 5 motor modules to produce locomotion phases. Our model succeeded to acquire stable locomotion by using the motor modules and reflexes. Next, we examined how a pathological condition altered motor modules. Specifically, we weakened neural inputs to muscles on one leg to simulate a stroke condition. Immediately after the simulated stroke, the model did not walk. Then, internal parameters were modified to recover stable locomotion. We refitted either (a) reflex parameters or (b) CPG parameters to compensate the locomotion by adapting (a) reflexes or (b) the controller. Stable locomotion was recovered in both conditions. However the motor modules were merged only in (b). These results suggest that light or sub-acute stroke patients, who can compensate stable locomotion by just adapting reflexes, would not show merge of motor modules, whereas severe or chronic patients, who needed to adapt the controller for compensation, would show the merge, as consistent with experimental findings.
研究論文（学術雑誌）, 英語 - Large-scale simulation of a layered cortical sheet of spiking network model using a tile partitioning method
Jun Igarashi; Hiroshi Yamaura; Tadashi Yamazaki
Frontiers in Neuroinformatics, 13, 71, 2019年, 査読付
研究論文（学術雑誌）, 英語 - Realtime simulation of a cat-scale artificial cerebellum on PEZY-SC processors
Tadashi Yamazaki; Jun Igarashi; Junichiro Makino; Toshikazu Ebisuzaki
International Journal of High Performance Computing Applications, SAGE Publications, 33, 1, 155, 168, 2019年, 査読付
研究論文（学術雑誌）, 英語 - 小脳神経回路をコンピュータ上に作る
山浦 洋; 山﨑 匡
Clinical Neuroscience, 中外医学社, In press, 2019年
研究論文（学術雑誌）, 日本語 - Revisiting a theory of cerebellar cortex
Tadashi Yamazaki; William Lennon
Neuroscience Research, Elsevier, 2019, 148, 1, 8, 2019年, 査読付, 招待
研究論文（学術雑誌）, 日本語 - 脳による足底接地情報の遅れ補償を組み込んだ下肢筋骨格系モデルの歩行シミュレーション
市村 大輔; 矢野 諭; 山﨑 匡
電子情報通信学会論文誌D, J100-D, 8, 808, 816, 2017年08月, 査読付
研究論文（学術雑誌）, 日本語 - 高性能神経計算による神経回路モデルのリアルタイムシミュレーション
山﨑匡; 五十嵐潤
日本神経回路学会誌, 24, 4, 172, 181, 2017年, 招待
研究論文（学術雑誌）, 日本語 - Real-Time Simulation of Passage-of-Time Encoding in Cerebellum Using a Scalable FPGA-Based System
Junwen Luo; Graeme Coapes; Terrence Mak; Tadashi Yamazaki; Chung Tin; Patrick Degenaar
IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS, IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 10, 3, 742, 753, 2016年06月, 査読付, 10.1109/TBCAS.2015.2460232, The cerebellum plays a critical role for sensorimotor control and learning. However, dysmetria or delays in movements' onsets consequent to damages in cerebellum cannot be cured completely at the moment. Neuroprosthesis is an emerging technology that can potentially substitute such motor control module in the brain. A pre-requisite for this to become practical is the capability to simulate the cerebellum model in real-time, with low timing distortion for proper interfacing with the biological system. In this paper, we present a frame-based network-on-chip (NoC) hardware architecture for implementing a bio-realistic cerebellum model with similar to 100 000 neurons, which has been used for studying timing control or passage-of-time (POT) encoding mediated by the cerebellum. The simulation results verify that our implementation reproduces the POT representation by the cerebellum properly. Furthermore, our field-programmable gate array (FPGA)-based system demonstrates excellent computational speed that it can complete 1sec real world activities within 25.6 ms. It is also highly scalable such that it can maintain approximately the same computational speed even if the neuron number increases by one order of magnitude. Our design is shown to outperform three alternative approaches previously used for implementing spiking neural network model. Finally, we show a hardware electronic setup and illustrate how the silicon cerebellum can be adapted as a potential neuroprosthetic platform for future biological or clinical application.
研究論文（学術雑誌）, 英語 - Real-World-Time Simulation of Memory Consolidation in a Large-Scale Cerebellar Model
Masato Gosui; Tadashi Yamazaki
FRONTIERS IN NEUROANATOMY, FRONTIERS MEDIA SA, 10, 21, 1, 10, 2016年03月, 査読付, 10.3389/fnana.2016.00021, We report development of a large-scale spiking network model of the cerebellum composed of more than 1 million neurons. The model is implemented on graphics processing units (GPUs), which are dedicated hardware for parallel computing. Using 4 GPUs simultaneously, we achieve realtime simulation, in which computer simulation of cerebellar activity for 1 s completes within 1 s in the real-world time, with temporal resolution of 1 ms. This allows us to carry out a very long-term computer simulation of cerebellar activity in a practical time with millisecond temporal resolution. Using the model, we carry out computer simulation of long-term gain adaptation of optokinetic response (OKR) eye movements for 5 days aimed to study the neural mechanisms of posttraining memory consolidation. The simulation results are consistent with animal experiments and our theory of posttraining memory consolidation. These results suggest that realtime computing provides a useful means to study a very slow neural process such as memory consolidation in the brain.
研究論文（学術雑誌）, 英語 - A Model of In vitro Plasticity at the Parallel Fiber-Molecular Layer Interneuron Synapses
William Lennon; Tadashi Yamazaki; Robert Hecht-Nielsen
FRONTIERS IN COMPUTATIONAL NEUROSCIENCE, FRONTIERS MEDIA SA, 9, 150, 1, 12, 2015年12月, 査読付, 10.3389/fncom.2015.00150, Theoretical and computational models of the cerebellum typically focus on the role of parallel fiber (PF)-Purkinje cell (PKJ) synapses for learned behavior, but few emphasize the role of the molecular layer interneurons (MLIs) the stellate and basket cells. A number of recent experimental results suggest the role of Mils is more important than previous models put forth. We investigate learning at PF MLI synapses and propose a mathematical model to describe plasticity at this synapse. We perform computer simulations with this form of learning using a spiking neuron model of the MLI and show that it reproduces six in vitro experimental results in addition to simulating four novel protocols. Further, we show how this plasticity model can predict the results of other experimental protocols that are not simulated. Finally, we hypothesize what the biological mechanisms are for changes in synaptic efficacy that embody the phenomenological model proposed here.
研究論文（学術雑誌）, 英語 - Modeling memory consolidation during posttraining periods in cerebellovestibular learning
Tadashi Yamazaki; Soichi Nagao; William Lennon; Shigeru Tanaka
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, NATL ACAD SCIENCES, 112, 11, 3541, 3546, 2015年03月, 査読付, 10.1073/pnas.1413798112, Long-term depression (LTD) at parallel fiber-Purkinje cell (PF-PC) synapses is thought to underlie memory formation in cerebellar motor learning. Recent experimental results, however, suggest that multiple plasticity mechanisms in the cerebellar cortex and cerebellar/vestibular nuclei participate in memory formation. To examine this possibility, we formulated a simple model of the cerebellum with a minimal number of components based on its known anatomy and physiology, implementing both LTD and long-term potentiation (LTP) at PF-PC synapses and mossy fiber-vestibular nuclear neuron (MF-VN) synapses. With this model, we conducted a simulation study of the gain adaptation of optokinetic response (OKR) eye movement. Our model reproduced several important aspects of previously reported experimental results in wild-type and cerebellum-related gene-manipulated mice. First, each 1-h training led to the formation of short-term memory of learned OKR gain at PF-PC synapses, which diminished throughout the day. Second, daily repetition of the training gradually formed long-term memory that was maintained for days at MF-VN synapses. We reproduced such memory formation under various learning conditions. Third, long-term memory formation occurred after training but not during training, indicating that the memory consolidation occurred during posttraining periods. Fourth, spaced training outperformed massed training in long-term memory formation. Finally, we reproduced OKR gain changes consistent with the changes in the vestibuloocular reflex (VOR) previously reported in some gene-manipulated mice.
研究論文（学術雑誌）, 英語 - 小脳の計算機シミュレーション
山﨑 匡
人工知能学会誌, 日本人工知能学会, 30, 5, 639, 646, 2015年
研究論文（学術雑誌）, 日本語 - A spiking network model of cerebellar Purkinje cells and molecular layer interneurons exhibiting irregular firing
William Lennon; Robert Hecht-Nielsen; Tadashi Yamazaki
FRONTIERS IN COMPUTATIONAL NEUROSCIENCE, FRONTIERS RESEARCH FOUNDATION, 8, 157, 1, 10, 2014年12月, 査読付, 10.3389/fncom.2014.00157, While the anatomy of the cerebellar microcircuit is well-studied, how it implements cerebellar function is not understood. A number of models have been proposed to describe this mechanism but few emphasize the role of the vast network Purkinje cells (PKJs) form with the molecular layer interneurons (MLIs)-the stellate and basket cells. We propose a model of the MLI-PKJ network composed of simple spiking neurons incorporating the major anatomical and physiological features. In computer simulations, the model reproduces the irregular firing patterns observed in PKJs and MLIs in vitro and a shift toward faster, more regular firing patterns when inhibitory synaptic currents are blocked. In the model, the time between PKJ spikes is shown to be proportional to the amount of feedforward inhibition from an MLI on average. The two key elements of the model are: (1) spontaneously active PKJs and MLIs due to an endogenous depolarizing current, and (2) adherence to known anatomical connectivity along a parasagittal strip of cerebellar cortex. We propose this model to extend previous spiking network models of the cerebellum and for further computational investigation into the role of irregular firing and MLIs in cerebellar learning and function.
研究論文（学術雑誌）, 英語 - Consensus Paper: The Cerebellum's Role in Movement and Cognition
Leonard F. Koziol; Deborah Budding; Nancy Andreasen; Stefano D'Arrigo; Sara Bulgheroni; Hiroshi Imamizu; Masao Ito; Mario Manto; Cherie Marvel; Krystal Parker; Giovanni Pezzulo; Narender Ramnani; Daria Riva; Jeremy Schmahmann; Larry Vandervert; Tadashi Yamazaki
CEREBELLUM, SPRINGER, 13, 1, 151, 177, 2014年02月, 査読付, 招待, 10.1007/s12311-013-0511-x, While the cerebellum's role in motor function is well recognized, the nature of its concurrent role in cognitive function remains considerably less clear. The current consensus paper gathers diverse views on a variety of important roles played by the cerebellum across a range of cognitive and emotional functions. This paper considers the cerebellum in relation to neurocognitive development, language function, working memory, executive function, and the development of cerebellar internal control models and reflects upon some of the ways in which better understanding the cerebellum's status as a "supervised learning machine" can enrich our ability to understand human function and adaptation. As all contributors agree that the cerebellum plays a role in cognition, there is also an agreement that this conclusion remains highly inferential. Many conclusions about the role of the cerebellum in cognition originate from applying known information about cerebellar contributions to the coordination and quality of movement. These inferences are based on the uniformity of the cerebellum's compositional infrastructure and its apparent modular organization. There is considerable support for this view, based upon observations of patients with pathology within the cerebellum.
研究論文（学術雑誌）, 英語 - Realtime cerebellum: A large-scale spiking network model of the cerebellum that runs in realtime using a graphics processing unit
Tadashi Yamazaki; Jun Igarashi
NEURAL NETWORKS, PERGAMON-ELSEVIER SCIENCE LTD, 47, 103, 111, 2013年11月, 査読付, 10.1016/j.neunet.2013.01.019, The cerebellum plays an essential role in adaptive motor control. Once we are able to build a cerebellar model that runs in realtime, which means that a computer simulation of 1 s in the simulated world completes within 1 s in the real world, the cerebellar model could be used as a realtime adaptive neural controller for physical hardware such as humanoid robots. In this paper, we introduce "Realtime Cerebellum (RC)", a new implementation of our large-scale spiking network model of the cerebellum, which was originally built to study cerebellar mechanisms for simultaneous gain and timing control and acted as a general-purpose supervised learning machine of spatiotemporal information known as reservoir computing, on a graphics processing unit (GPU). Owing to the massive parallel computing capability of a GPU, RC runs in realtime, while reproducing qualitatively the same simulation results of the Pavlovian delay eyeblink conditioning with the previous version. RC is adopted as a realtime adaptive controller of a humanoid robot, which is instructed to learn a proper timing to swing a bat to hit a flying ball online. These results suggest that RC provides a means to apply the computational power of the cerebellum as a versatile supervised learning machine towards engineering applications. (C) 2013 Elsevier Ltd. All rights reserved.
研究論文（学術雑誌）, 英語 - Transfer of memory trace of cerebellum-dependent motor learning in human prism adaptation: A model study
Soichi Nagao; Takeru Honda; Tadashi Yamazaki
Neural Networks, 47, 72, 80, 2013年11月, 査読付, 10.1016/j.neunet.2013.01.017, Accumulating experimental evidence suggests that the memory trace of ocular reflex adaptation is initially encoded in the cerebellar cortex, and later transferred to the cerebellar nuclei for consolidation through repetitions of training. However, the memory transfer is not well characterized in the learning of voluntary movement. Here, we implement our model of memory transfer to interpret the data of prism adaptation (Martin, Keating, Goodkin, Bastian, &
Thach, 1996a, 1996b), assuming that the cerebellar nuclear memory formed by memory transfer is used for normal throwing. When the subject was trained to throw darts wearing prisms in 30-40 trials, the short-term memory for recalibrating the throwing direction by gaze would be formed in the cerebellar cortex, which was extinguished by throwing with normal vision in a similar number of trials. After weeks of repetitions of short-term prism adaptation, the long-term memory would be formed in the cerebellar nuclei through memory transfer, which enabled one to throw darts to the center wearing prisms without any training. These two long-term memories, one for throwing with normal vision and the other for throwing wearing prisms, are assumed to be utilized automatically under volitional control. Moreover, when the prisms were changed to new prisms, a new memory for adapting to the new prisms would be formed in the cerebellar cortex, just to counterbalance the nuclear memory of long-term adaptation to the original prisms in a similar number of trials. These results suggest that memory transfer may occur in the learning of voluntary movements. © 2013 Elsevier Ltd.
研究論文（学術雑誌）, 英語 - A Computational Mechanism for Unified Gain and Timing Control in the Cerebellum
Tadashi Yamazaki; Soichi Nagao
PLOS ONE, PUBLIC LIBRARY SCIENCE, 7, 3, e33319, 2012年03月, 査読付, 10.1371/journal.pone.0033319, Precise gain and timing control is the goal of cerebellar motor learning. Because the basic neural circuitry of the cerebellum is homogeneous throughout the cerebellar cortex, a single computational mechanism may be used for simultaneous gain and timing control. Although many computational models of the cerebellum have been proposed for either gain or timing control, few models have aimed to unify them. In this paper, we hypothesize that gain and timing control can be unified by learning of the complete waveform of the desired movement profile instructed by climbing fiber signals. To justify our hypothesis, we adopted a large-scale spiking network model of the cerebellum, which was originally developed for cerebellar timing mechanisms to explain the experimental data of Pavlovian delay eyeblink conditioning, to the gain adaptation of optokinetic response (OKR) eye movements. By conducting large-scale computer simulations, we could reproduce some features of OKR adaptation, such as the learning-related change of simple spike firing of model Purkinje cells and vestibular nuclear neurons, simulated gain increase, and frequency-dependent gain increase. These results suggest that the cerebellum may use a single computational mechanism to control gain and timing simultaneously.
研究論文（学術雑誌）, 英語 - Simulation Platform: A cloud-based online simulation environment
Tadashi Yamazaki; Hidetoshi Ikeno; Yoshihiro Okumura; Shunji Satoh; Yoshimi Kamiyama; Yutaka Hirata; Keiichiro Inagaki; Akito Ishihara; Takayuki Kannon; Shiro Usui
NEURAL NETWORKS, PERGAMON-ELSEVIER SCIENCE LTD, 24, 7, 693, 698, 2011年09月, 10.1016/j.neunet.2011.06.010, For multi-scale and multi-modal neural modeling, it is needed to handle multiple neural models described at different levels seamlessly. Database technology will become more important for these studies, specifically for downloading and handling the neural models seamlessly and effortlessly. To date, conventional neuroinformatics databases have solely been designed to archive model files, but the databases should provide a chance for users to validate the models before downloading them. In this paper, we report our on-going project to develop a cloud-based web service for online simulation called "Simulation Platform". Simulation Platform is a cloud of virtual machines running GNU/Linux. On a virtual machine, various software including developer tools such as compilers and libraries, popular neural simulators such as GENESIS, NEURON and NEST, and scientific software such as Gnuplot, R and Octave, are pre-installed. When a user posts a request, a virtual machine is assigned to the user, and the simulation starts on that machine. The user remotely accesses to the machine through a web browser and carries out the simulation, without the need to install any software but a web browser on the user's own computer. Therefore, Simulation Platform is expected to eliminate impediments to handle multiple neural models that require multiple software. (C) 2011 Elsevier Ltd. All rights reserved.
研究論文（学術雑誌）, 英語 - Stimulus-Dependent State Transition between Synchronized Oscillation and Randomly Repetitive Burst in a Model Cerebellar Granular Layer
Takeru Honda; Tadashi Yamazaki; Shigeru Tanaka; Soichi Nagao; Tetsuro Nishino
PLOS COMPUTATIONAL BIOLOGY, PUBLIC LIBRARY SCIENCE, 7, 7, e1002087, 2011年07月, 査読付, 10.1371/journal.pcbi.1002087, Information processing of the cerebellar granular layer composed of granule and Golgi cells is regarded as an important first step toward the cerebellar computation. Our previous theoretical studies have shown that granule cells can exhibit random alternation between burst and silent modes, which provides a basis of population representation of the passage-of-time (POT) from the onset of external input stimuli. On the other hand, another computational study has reported that granule cells can exhibit synchronized oscillation of activity, as consistent with observed oscillation in local field potential recorded from the granular layer while animals keep still. Here we have a question of whether an identical network model can explain these distinct dynamics. In the present study, we carried out computer simulations based on a spiking network model of the granular layer varying two parameters: the strength of a current injected to granule cells and the concentration of Mg2+ which controls the conductance of NMDA channels assumed on the Golgi cell dendrites. The simulations showed that cells in the granular layer can switch activity states between synchronized oscillation and random burst-silent alternation depending on the two parameters. For higher Mg2+ concentration and a weaker injected current, granule and Golgi cells elicited spikes synchronously (synchronized oscillation state). In contrast, for lower Mg2+ concentration and a stronger injected current, those cells showed the random burst-silent alternation (POT-representing state). It is suggested that NMDA channels on the Golgi cell dendrites play an important role for determining how the granular layer works in response to external input.
研究論文（学術雑誌）, 英語 - Reprint of: Simulation Platform: A cloud-based online simulation environment.
Tadashi Yamazaki; Hidetoshi Ikeno; Yoshihiro Okumura; Shunji Satoh; Yoshimi Kamiyama; Yutaka Hirata; Keiichiro Inagaki; Akito Ishihara; Takayuki Kannon; Shiro Usui
Neural Networks, 24, 9, 927, 932, 2011年, 査読付, 10.1016/j.neunet.2011.08.007 - 脳スパイキングネットワークモデルにおける条件刺激強度依存性タイミング制御
本多武尊; 山崎匡; 田中繁; 西野哲朗
情報処理学会論文誌数理モデル化と応用, 3, 2, 1--10, 2010年, 査読付
研究論文（学術雑誌）, 日本語 - Robust reservoir generation by correlation-based learning
Tadashi Yamazaki; Shigeru Tanaka
Advances in Articial Neural Networks, 2009, 1, Article ID: 467128, 2009年, 査読付
研究論文（学術雑誌）, 英語 - 小脳顆粒層をモデル化したスパイキングネットワークの研究：NMDA 受容体を介した同期発火状態と時間表現状態の遷移
本多武尊; 山崎匡; 田中繁; 西野哲朗
電子情報通信学会 論文誌D, J91-D, 11, 2709--2718, 2008年, 査読付
研究論文（学術雑誌）, 日本語 - A spiking network model for passage-of-time representation in the cerebellum
Tadashi Yamazaki; Shigeru Tanaka
EUROPEAN JOURNAL OF NEUROSCIENCE, BLACKWELL PUBLISHING, 26, 8, 2279, 2292, 2007年10月, 査読付, 10.1111/j.1460-9568.2007.05837.x, In Pavlovian delay eyeblink conditioning, the cerebellum represents the passage-of-time (POT) between onsets of conditioned and unconditioned stimuli (CS and US, respectively). To study possible computational mechanisms of the POT representation we built a large-scale spiking network model of the cerebellum. Consistent with our previous rate-coding model, we found two conditions necessary for the present model to represent the POT with a dynamic population of active granule cells: (i) long temporal integration of input signals; and (ii) random recurrent connections between granule and Golgi cells. When these conditions were satisfied, a nonrecurrent sequence of active granule cell populations was generated in response to a CS and, conversely, the POT from the CS onset was able to be read out from the sequence. Specifically, simulated N-methyl-D-aspartate (NMDA) channels with a long decay time constant at granule and Golgi cells were responsible for the long temporal integration. Thus, blocking the NMDA channels or ablating Golgi cells impaired the POT representation. Simulated glomerulus structure made POT representation robust against noise in mossy fibre inputs. Long-term potentiation induced at mossy fibre synapses on granule cells also served to enhance the robustness. We reproduced some experimental results of Pavlovian delay eyeblink conditioning using the present model. These results suggest that the recurrent network in the granular layer and NMDA channels in granule and Golgi cells play an essential role in the timing mechanisms in the cerebellum, whereas the glomerulus serves to realize a robust representation of time.
研究論文（学術雑誌）, 英語 - The cerebellum as a liquid state machine
Tadashi Yamazaki; Shigeru Tanaka
NEURAL NETWORKS, PERGAMON-ELSEVIER SCIENCE LTD, 20, 3, 290, 297, 2007年04月, 査読付, 10.1016/j.neunet.2007.04.004, We examined closely the cerebellar circuit model that we have proposed previously. The model granular layer generates a finite but very long sequence of active neuron populations without recurrence, which is able to represent the passage of time. For all the possible binary patterns fed into mossy fibres, the circuit generates the same number of different sequences of active neuron populations. Model Purkinje cells that receive parallel fiber inputs from neurons in the granular layer learn to stop eliciting spikes at the timing instructed by the arrival of signals from the inferior olive. These functional roles of the granular layer and Purkinje cells are regarded as a liquid state generator and readout neurons, respectively. Thus, the cerebellum that has been considered to date as a biological counterpart of a perceptron is reinterpreted to be a liquid state machine that possesses powerful information processing capability more than a perceptron. (c) 2007 Published by Elsevier Ltd.
研究論文（学術雑誌）, 英語 - Chronically mountable goggles for persistent exposure to single orientation
Shigeru Tanaka; Toshiki Tani; Jerome Ribot; Tadashi Yamazaki
JOURNAL OF NEUROSCIENCE METHODS, ELSEVIER SCIENCE BV, 160, 2, 206, 214, 2007年03月, 査読付, 10.1016/j.jneumeth.2006.09.004, To examine the effect of experience on the developmental plasticity of functional maps in the visual cortex, we need to establish a method for a stable visual experience manipulation under the freely moving condition. For this purpose, we fabricated goggles that are chronically mounted stably on the animal's head, but easy to replace according to the animal's growth. Here we report the design of the goggles and the method of mounting them on the head of animals. By this method, combined with the intrinsic signal optical imaging technique, we were able to observe a rapid and robust reorganization of orientation maps. (c) 2006 Elsevier B.V. All rights reserved.
研究論文（学術雑誌）, 英語 - インターナルクロックモデルに基づくロボット制御法の実現
真鍋秀聡; 西野哲朗; 山崎匡; 田中繁
情報処理学会論文誌数理モデル化と応用, 48, SIG19, 139--154, 2007年, 査読付
研究論文（学術雑誌）, 日本語 - Neural Modeling of an internal clock
T Yamazaki; S Tanaka
NEURAL COMPUTATION, M I T PRESS, 17, 5, 1032, 1058, 2005年05月, 査読付, We studied a simple random recurrent inhibitory network. Despite its simplicity, the dynamics was so rich that activity patterns of neurons evolved with time without recurrence due to random recurrent connections among neurons. The sequence of activity patterns was generated by the trigger of an external signal, and the generation was stable against noise. Moreover, the same sequence was reproducible using a strong transient signal, that is, the sequence generation could be reset. Therefore, a time passage from the trigger of an external signal could be represented by the sequence of activity patterns, suggesting that this model could work as an internal clock. The model could generate different sequences of activity patterns by providing different external signals; thus, spatiotemporal information could be represented by this model. Moreover, it was possible to speed up and slow down the sequence generation.
研究論文（学術雑誌）, 英語 - A neural network model for trace conditioning
T Yamazaki; S Tanaka
INTERNATIONAL JOURNAL OF NEURAL SYSTEMS, WORLD SCIENTIFIC PUBL CO PTE LTD, 15, 1-2, 23, 30, 2005年02月, 査読付, We studied the dynamics of a neural network that has both recurrent excitatory and random inhibitory connections. Neurons started to become active when a relatively weak transient excitatory signal was presented and the activity was sustained due to the recurrent excitatory connections. The sustained activity stopped when a strong transient signal was presented or when neurons were disinhibited. The random inhibitory connections modulated the activity patterns of neurons so that the patterns evolved without recurrence with time. Hence, a time passage between the onsets of the two transient signals was represented by the sequence of activity patterns. We then applied this model to represent the trace eyeblink conditioning, which is mediated by the hippocampus. We assumed this model as CA3 of the hippocampus and considered an output neuron corresponding to a neuron in CA1. The activity pattern of the output neuron was similar to that of CA1 neurons during trace eyeblink conditioning, which was experimentally observed.
研究論文（学術雑誌）, 英語 - Sound frequency representation in cat auditory cortex
Tsytsarev, V; T Yamazaki; J Ribot; S Tanaka
NEUROIMAGE, ACADEMIC PRESS INC ELSEVIER SCIENCE, 23, 4, 1246, 1255, 2004年12月, 査読付, 10.1016/j.neuroimage.2004.08.021, Using the intrinsic signal optical recording technique, we reconstructed the two-dimensional pattern of stimulus-evoked neuronal activities in the auditory cortex of anesthetized and paralyzed cats. The average magnitude of intrinsic signal in response to a pure tone stimulus increased steadily as the sound pressure level increased. A detailed analysis demonstrated that the evoked signals at early frames were scaled by the sound pressure level, which in turn indicated the presence of a minimum level of sound pressure beyond which stimulus-related intrinsic signal can be generated. Intrinsic signals evoked significantly by pure tone stimuli of different frequencies were localized and arranged in an orderly manner in the middle ectosylvian gyros, which indicates that the primary auditory field (AI) is tonotopically organized. The arrangement of optimal frequencies obtained from optical recordings of the same auditory cortex, which were conducted on different days, was highly reproducible. Furthermore, other auditory fields surrounding AI in the recorded area were allocated based on the observed tonotopicity. We also conducted unit recordings on the cats used for optical recording with the same set of acoustic stimuli. The gross feature of the arrangement of optimal frequencies determined by unit recordings agreed with the tonotopic arrangement determined by the optical recording, although the precise agreement was not obtained. (C) 2004 Elsevier Inc. All rights reserved.
研究論文（学術雑誌）, 英語 - Mathematical analysis of a correlation-based model for orientation map formation
T Yamazaki
NEURAL NETWORKS, PERGAMON-ELSEVIER SCIENCE LTD, 16, 1, 47, 54, 2003年01月, 査読付, We consider a correlation-based model for the orientation map formation proposed by Miller [Journal of Neuroscience 14 (1994) 409] and study the formation mathematically. We perform the Fourier transform and compute the principal component of the model. With our analysis, the roles of functions are clarified and the result indicates that the developed orientation maps have the following properties. (1) Maps have oriented receptive fields. (2) Preferred orientations smoothly change on the cortical surface. (3) Periodicity of preferred orientations does not appear. (4) Periodicity of phases appears. (5) Singular points appear irregularly on the cortical surface. Our analytical results are justified by computer simulations. (C) 2002 Elsevier Science Ltd. All rights reserved.
研究論文（学術雑誌）, 英語 - A mathematical analysis of the development of oriented receptive fields in Linsker's model
T Yamazaki
NEURAL NETWORKS, PERGAMON-ELSEVIER SCIENCE LTD, 15, 2, 201, 207, 2002年03月, 査読付, We consider Linsker's neural network model and study how oriented receptive fields (RFs) are developed at layer F --> G mathematically. We concentrate on the Mexican-hat correlation function and show that this function determines the spatial frequency of RFs. We also focus on the role of the arbor function, Two types of arbor functions, a Gaussian type and a step type, are considered, and it is shown that the Gaussian arbor function develops center-surround RFs while the step arbor function enables the development of multi-lobed RFs in addition to center-surround RFs. The roles of Linsker's parameters k(1), k(2) are also discussed. (C) 2002 Elsevier Science Ltd. All rights reserved.
研究論文（学術雑誌）, 英語
MISC
- まだ解かれていない七つの小脳の基本的研究課題
永雄総一; 山崎匡
金原一郎記念医学医療振興財団, 2012年, 生体の科学, 63, 1, 3, 10, 日本語, 記事・総説・解説・論説等（その他）, 0370-9531, 40019196266, AN10360633 - 小脳内部時計の神経機構と機能的役割
山崎匡
Timing control and gain control are twins in motor control. Deficits in precise timing control could cause various motor disturbances. The cerebellum is responsible for timing control within the range of tens to hundreds of milliseconds in both motor and cognitive domains. This implies that the cerebellum has a certain mechanism to represent the passage of time internally. This article reviews the neural mechanisms of timing control in the cerebellum., 2012年, 臨床神経学, 52, 11, 990, 993, 日本語, 記事・総説・解説・論説等（その他）, 10.5692/clinicalneurol.52.990, 0009-918X, 23196494, 84880544555 - 現代の小脳パーセプトロン仮説
山崎匡
2011年, 日本神経回路学会論文誌, 18, 1, 22, 30, 日本語, 記事・総説・解説・論説等（その他） - Computational Models of Timing Mechanisms in the Cerebellar Granular Layer
Tadashi Yamazaki; Shigeru Tanaka
A long-standing question in neuroscience is how the brain controls movement that requires precisely timed muscle activations. Studies using Pavlovian delay eyeblink conditioning provide good insight into this question. In delay eyeblink conditioning, which is believed to involve the cerebellum, a subject learns an interstimulus interval (ISI) between the onsets of a conditioned stimulus (CS) such as a tone and an unconditioned stimulus such as an airpuff to the eye. After a conditioning phase, the subject's eyes automatically close or blink when the ISI time has passed after CS onset. This timing information is thought to be represented in some way in the cerebellum. Several computational models of the cerebellum have been proposed to explain the mechanisms of time representation, and they commonly point to the granular layer network. This article will review these computational models and discuss the possible computational power of the cerebellum., SPRINGER, 2009年12月, CEREBELLUM, 8, 4, 423, 432, 英語, 査読付, 記事・総説・解説・論説等（その他）, 10.1007/s12311-009-0115-7, 1473-4222, WOS:000272384000002 - 小脳の計算機構の完全理解とその応用を目指して
山崎匡
2009年, 日本神経回路学会誌, 16, 4, 190, 195, 日本語, 査読付, 記事・総説・解説・論説等（その他） - 万能シミュレータとしての小脳は非線形性をどのように活用しているか
山崎匡
2008年, 岩波『科学』, 78, 11, 1224, 1227, 日本語, 記事・総説・解説・論説等（その他）
書籍等出版物
- Cerebellum as a CNS Hub
Tadashi Yamazaki
学術書, 英語, 共著, Evolution of the Marr-Albus-Ito Model, Springer, 2021年11月11日, 無償でダウンロードできない - 人工知能学大辞典
山﨑 匡
事典・辞書, 日本語, 分担執筆, リザーバーコンピューティング, 共立出版社, 2017年 - Annual Review 神経 2016
山﨑 匡
学術書, 日本語, 共著, 小脳と時計, 中外医学社, 2016年03月09日 - Cerebellar learning
Masao Ito; Kazuhiko Yamaguchi; Soichi Nagao; Tadashi Yamazaki
学術書, 英語, 共著, Chapter 1. Long-Term Depression as a Model of Cerebellar Plasticity, Elsevier, 2014年06月20日, 9780444633569
講演・口頭発表等
- シミュレーション神経科学は新しい種類の神経科学か？
山﨑匡
口頭発表（招待・特別）, 日本語, NEURO2022, 招待, 国際会議
2022年07月02日 - スパコン「富岳」で脳をつくる
口頭発表（招待・特別）, 日本語, 「富岳」成果創出加速プログラムシンポジウム「富岳」百景, 招待, 国内会議
2022年03月29日 - Large-scale simulation of the little brain
Tadashi Yamazaki
口頭発表（招待・特別）, 英語, RIKEN Workshop on Neuromorphic Computing, 招待, RIKEN R-CCS, 国際会議
2019年03月11日 - Computer simulation of a monkey-scale cerebellum with 8 billion spiking neurons in realtime and its applications
Tadashi Yamazaki
口頭発表（招待・特別）, 英語, 75th Fujihara Seminar "Cerebellum as a CNS hub", 招待, Fujihara Foundation, 国際会議
2018年12月03日 - ヒト規模の脳神経回路シミュレーションを目指して: 小脳の場合
口頭発表（招待・特別）, 日本語, 科学技術計算分科会2017年度会合「拡がるHPC ～新たな鼓動～」, 国内会議
2017年10月26日 - Cat-scale artificial cerebellum on an energy-efficient supercomputer Shoubu
口頭発表（招待・特別）, 英語, Workshop on Brain-inspired Hardware, 招待, AI Research Center, AIST, 国際会議
2017年03月30日 - Shoubuで実現するネコ一匹分の人工小脳
口頭発表（招待・特別）, 日本語, 理研シンポジウム「スーパーコンピュータHOKUSAIとShoubu、研究開発の最前線」, 国内会議
2016年06月08日 - Building a 1 mm^3 cerebellar module on a computer
口頭発表（招待・特別）, 英語, Neuroinformatics 2015, 招待, International Neuroinformatics Coordinating Facility, Cairns, Australia, 国際会議
2015年08月20日 - 人工小脳の構築を目指して
口頭発表（一般）, 日本語, 東海大学シンポジウム「若手研究者による小脳研究の最前線」, 国内会議
2015年01月10日 - GPUによる小脳スパイキングネットワークモデルの実時間シミュレーション
山﨑 匡
口頭発表（招待・特別）, 日本語, 第5回アクセラレーション技術発表討論会「脳を超えよう!」, 招待, 国内会議
2013年09月 - Realtime Cerebellum: GPU-accelerated numerical simulation of a cerebellar spiking network model in realtime
Tadashi Yamazaki
口頭発表（招待・特別）, 英語, Workshop on Computations in the cerebellar circuit: advances on the modeling front, 招待, Computational Neuroscience (CNS) 2013 Paris, Paris, France, 国際会議
2013年07月 - Building the artificial cerebellum on a computer
Tadashi Yamazaki
口頭発表（招待・特別）, 英語, Neuro2013 Sattelite Symposium "Neuroscience of the cerebellum: from molecular biology to cognitive science", 招待, 国際会議
2013年06月22日 - Realtime simulation of a cerebellar spiking network model using a GPU
Tadashi Yamazaki
口頭発表（招待・特別）, 英語, Open Source Brain Kickoff Meeting, 招待, Sardinia, Italy, 国際会議
2013年05月 - コンピュータ上に小脳を創る。
山崎匡
その他, 日本語, 第６回次世代医師・研究者交流会, 自治医科大学
2013年02月 - Numerical Brain: 計算機上に小脳を創る
山崎匡
口頭発表（招待・特別）, 日本語, 平成24年度第5回ブレインウェア研究会, 東北大学
2012年12月 - 小脳内部時計の神経機構と機能的役割
山崎匡
口頭発表（招待・特別）, 日本語, 第53回日本神経学会学術講演会, 日本神経学会
2012年05月 - 小脳の計算機構の完全理解とその応用を目指して
山崎匡
口頭発表（招待・特別）, 日本語, 第19 回日本神経回路学会全国大会, 日本神経回路学会
2009年09月 - Large-scale computational model of the cerebellum for simultaneous control of timing and gain of movement
山崎匡
口頭発表（招待・特別）, 日本語, 平成20 年度ミニシンポジウム「小脳・脳幹による運動の制御と学習：神経生理，数理モデル，実機制御」, 中部大学
2008年11月
担当経験のある科目_授業
- Information Engineering Laboratory
The University of Electro-Communications - 情報工学工房
電気通信大学 - Mathematical Information Science Program Laboratory II
The University of Electro-Communications - 情報数理工学実験第二
電気通信大学 - Advanced Topics on Simulation Science and Engineering
The University of Electro-Communications - シミュレーション理工学特論
電気通信大学 - Numerical Calculus
The University of Electro-Communications - 数値計算
電気通信大学 - Graduate Technical English
The University of Electro-Communications - 大学院技術英語
電気通信大学 - 情報数理工学実験第二
電気通信大学 - 情報数理工学実験第二
電気通信大学 - 情報数理工学実験第一
The University of Electro-Communications - 情報数理工学実験第一
電気通信大学 - 情報工学工房
電気通信大学 - 情報工学工房
電気通信大学 - 情報・通信演習1
電気通信大学 - 情報・通信演習1
電気通信大学 - 情報領域演習第二
電気通信大学 - 情報領域演習第二
電気通信大学 - 情報数理工学実験第2
The University of Electro-Communications - 情報数理工学実験第2
電気通信大学 - 情報工学演習第二
The University of Electro-Communications - 情報工学演習第二
電気通信大学 - 情報数理工学実験第一
The University of Electro-Communications - 情報数理工学実験第一
電気通信大学 - 情報・通信演習1
The University of Electro-Communications - 情報・通信演習1
電気通信大学 - 情報数理工学実験第2
The University of Electro-Communications - 情報数理工学実験第2
電気通信大学 - 情報工学演習第二
The University of Electro-Communications - 情報工学演習第二
電気通信大学 - 情報数理工学実験第一
電気通信大学 - 情報数理工学実験第一
電気通信大学 - 情報・通信演習1
The University of Electro-Communications - 情報・通信演習1
電気通信大学 - キャリア教育演習
電気通信大学 - キャリア教育演習
電気通信大学 - 情報数理工学実験第2
電気通信大学 - 情報数理工学実験第2
電気通信大学 - 情報数理工学実験第1
電気通信大学 - 情報数理工学実験第1
電気通信大学 - 情報工学演習第二
電気通信大学 - 情報工学演習第二
電気通信大学 - 情報・通信演習1
The University of Electro-Communications - 情報・通信演習1
電気通信大学 - Advanced Topics on Simulation Science and Engineering
The University of Electro-Communications - シミュレーション理工学特論
電気通信大学 - Mathematical Information Science Program Laboratory II
The University of Electro-Communications - 情報数理工学実験第二
電気通信大学 - Numerical Calculus
The University of Electro-Communications - 数値計算
電気通信大学 - Graduate Technical English
The University of Electro-Communications - 大学院技術英語
電気通信大学
共同研究・競争的資金等の研究課題
- 行動変容の表出を規定する運動回路動態の解明
2022年 - 2026年 - 小脳機能ゾーンの生理学的意義の解明
2022年 - 2025年 - 神経回路シミュレーションによる知覚の点火を促す生物物理学的メカニズムの解明
内藤記念科学振興財団, 研究代表者
2022年11月 - 2024年10月 - 時間符号化スパイキングネットワークの探求
栢森情報科学振興財団, 研究代表者
2021年10月 - 2023年09月 - 小脳神経細胞のミクロな形態が回路全体のマクロな情報処理に及ぼす影響のモデル研究
2020年 - 2023年 - 脳情報動態を規定する他領野連関と並列処理
2017年 - 2022年 - 日本神経回路学会創立30周年記念助成金
2020年01月 - 2020年12月 - 小脳学習における神経細胞の空間形状の機能的役割に関するシミュレーション研究
2017年 - 2020年 - 小脳内に分散された複数の可塑性による相補的運動学習メカニズムの解明
山﨑 匡
研究代表者
2014年 - 2017年 - 細胞集団活動の遷移による時間経過表現のモデル研究
山﨑 匡
研究代表者
2014年 - 2016年 - 人工小脳の学習と予測にもとづく小型ヒューマノイドロボットの運動制御に関する研究
山﨑 匡
人工知能研究振興財団, 研究助成, 研究代表者
2014年04月01日 - 2015年03月31日