2022 Fiscal Year Final Research Report
Time dynamics on information propagation in human brain performing hierarchical structure learning
| Project/Area Number |
19K16894
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 51020:Cognitive and brain science-related
|
|---|
| Research Institution | Kyoto University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2019-04-01 – 2023-03-31
|
| Keywords | ベイズ学習 / 強化学習 / EEG |
|---|
| Outline of Final Research Achievements |
The objective of this research was to examine the temporal dynamics of brain information processing during the updating of multiple parameters using Bayesian learning. While Bayesian learning regulates the computation of updates with multiple parameters, it does not specify the order in which the parameters should be updated. In this study, we defined the temporal dynamics as the update order and update speed (the time required for updating), and investigated them using a behavioral model and EEG analysis. Our findings indicate the following: 1) The update order was adaptively determined based on the certainty of the parameter estimate, which depends on the learning progress, and 2) The update computation was completed within 400 ms after the arrival of new information.
|
| Free Research Field |
計算認知神経科学
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究は，情報をヒト脳に入力したときに，脳がどのようにその情報を処理し，脳内の外部環境に対する推定を更新するかを，時間ダイナミクスの観点から調査した．これまでの認知神経科学の研究では，学習を担う脳領域に着目することが多かった．これは脳の情報処理回路の回路構造を同定することを目的としている．一方で，時間ダイナミクスは，回路の構成部品がどのような役割を持っているかを知る手掛かりになる．本研究の成果は，脳回路に関する新たな知見と研究アプローチを提供しており，ヒトがどのように情報処理を行うかの解明に貢献する．
|
