シンポジウム
11 計算神経科学の新展開
11 Advances in Computational Neuroscience
座長：豊泉 太郎（理化学研究所 脳神経科学研究センター） |
|---|
| 2022年7月3日 9:00～9:30 沖縄コンベンションセンター 会議場B1 第3会場
4S03m-01
Sequences and modularity of dynamic attractors in inhibition-dominated neural networks
*Carina Curto(1)
1. Penn State University
Keyword: threshold-linear networks, graphs, sequences, central pattern generators (CPGs)
Threshold-linear networks (TLNs) display a wide variety of nonlinear dynamics including multistability, limit cycles, quasiperiodic attractors, and chaos. Over the past few years, we have developed a detailed mathematical theory relating stable and unstable fixed points of TLNs to graph-theoretic properties of the underlying network. These results enable us to design networks that count stimulus pulses, track position, and encode multiple locomotive gaits in a single central pattern generator circuit. | | 2022年7月3日 9:30～10:00 沖縄コンベンションセンター 会議場B1 第3会場
4S03m-02
An economic decision-making model of anticipated surprise with dynamic expectation
*Ho Ka Chan(1), Taro Toyoizumi(1,2)
1. RIKEN Center for Brain Science, Wako, Saitama, Japan, 2. Dept Mathematical Informatics, Grad Sch Information Science and Technology, Univ of Tokyo, Tokyo, Japan
Keyword: decision making, surprise, sequential anticipation, economic paradox
When making decision under risk, people often exhibit behaviors that cannot be explained by classical economic theories. The reason is that the axioms of the Expected Utility Theory are violated. Various models, e.g. the famous Prospect Theory by Kahneman and Tversky, are made in an attempt to explain these ‘irrational’ behavior. However, many of these models lack neuroscience bases and require introduction of subjective and problem-specific constructs. Here, we present a decision-making model inspired by the prediction error signals and introspective neuronal replay reported in the brain. In the model, decisions are chosen based on ‘anticipated surprise’, defined by a nonlinear average of the differences between individual outcomes and a reference point. The reference point is determined by the expected value of the possible outcomes, which can dynamically change during the mental simulation of potential events in decision-making problems that involves sequential stages. With this formalism, the contribution of each stage to the appeal of available options in a decision-making problem can be elucidated. The model does not depend on non-linear weighting of the probabilities of outcomes and other problem-specific construct as in many other models, thus making it more objective and unambiguous. Using the model, several economic paradoxes, e.g. Allais Paradox and Ellsberg Paradox, as well as gambling behaviors, e.g. Blackjack gambling in casino, can be explained. Our work could provide a unified framework for understanding various types of ‘irrationality’ in decision making. It could also provide a novel direction for exploring the neuroscience bases of decision-making in economics, which help bridge the gap between the two distinct fields. | | 2022年7月3日 10:00～10:30 沖縄コンベンションセンター 会議場B1 第3会場
4S03m-03
The roles of inhibitory networks in spatial orientation working memory in insects
*Lo Chung-Chuan(1,2)、Hsuan-Pei Huang(1)、Ning Chang(1)
*Chung-Chuan Lo(1,2), Hsuan-Pei Huang(1), Ning Chang(1)
1. Inst of Sys Neurosci, Natl Tsing Hua Uni, Taiwan, 2. Brain Res Ctr, Natl Tsing Hua Uni, Taiwan
Keyword: spatial orientation, working memory, attractor networks, Drosophila
Spatial orientation memory plays a crucial role in navigation of animals as they need to maintain a stable sense of the direction of heading regardless the availability of external cues or landmarks. Recent studies have revealed stunning details about the activity and the functions of head-direction cells of Drosophila melanogaster (fruit fly), and suggested that the head-direction system can be well explained by the classical ring attractor models. However, connectomic data reveal networks of projecting and local inhibitory neurons in the head-direction system with a patterns much more complex than the classical models required. To address this issue, we propose a detailed neural network model of the fruit fly head-direction system based on the latest connectomic data. Our model suggests that the inhibitory neurons in the fruit fly head-direction system provide more functional roles than the classical models proposed. Specifically, the inhibitory networks coordinate the activation of different sub-circuits in the system, allowing the system to stabilize itself, to control the width of the activity bump and to update the bump location in accordance with the body rotation. | | 2022年7月3日 10:30～11:00 沖縄コンベンションセンター 会議場B1 第3会場
4S03m-04
The role of population structure in computations through neural dynamics
*Srdjan Ostojic(1)
1. Ecole Normale Superieure Paris
Keyword: Neural networks, Latent dynamics, Cognitive computations
Neural computations are currently investigated using two separate approaches: sorting neurons into functional populations, or examining the low-dimensional dynamics of collective activity. Whether and how these two aspects interact to shape computations is currently unclear. Using a novel approach to extract computational mechanisms from networks trained on neuroscience tasks, here we show that the dimensionality of the dynamics and cell-class structure play fundamentally complementary roles. While various tasks can be implemented by increasing the dimensionality in networks with fully random population structure, flexible input-output mappings instead required a non-random population structure that can be described in terms of multiple sub-populations. Our analyses revealed that such a population structure enabled flexible computations through a mechanism based on gain-controlled modulations that flexibly shape the dynamical landscape of collective dynamics. Our results lead to task-specific predictions for the structure of neural selectivity, inactivation experiments, and for the implication of different neurons in multi-tasking. | |
|
|
