Neural mechanism for computing information value as abstract reward

2018 Fiscal Year Final Research Report

• Project/Area Number: 15K01843
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Basic / Social brain science
• Research Institution: Tokyo Institute of Technology
• Principal Investigator: Nakamura Kiyohiko 東京工業大学, 情報理工学院, 教授 (10172397)
• Project Period (FY): 2015-04-01 – 2019-03-31
• Keywords: 抽象的報酬 / 情報量 / 不確実性削減 / 好奇心 / サル / 細胞集団活動

Outline of Final Research Achievements

The desire to obtain information is a powerful motivator in daily life, as well as desires for primary rewards such as food and water. Psychological studies have presented several criteria that human subjects use to assess information . We examined which criterion the nervous system uses to assess information among the following criteria, the value of information (VOI), the probability gain , and Shannon information. We recorded the activity of about 1000 neurons from the lateral prefrontal cortex (lPFC) of two monkeys while they conducted information seeking tasks. Results: 1. A principal component analysis showed that responses of the entire populations of the recorded neurons effectively discriminated between the tasks performed. 2. Dynamics of the population responses showed that the population responses could encode the VOI longer than the probability gain, and not Shannon information. These indicate that the lPFC encodes the VOI and the probability gain to assess information.

Free Research Field

神経行動学

Academic Significance and Societal Importance of the Research Achievements

ヒトは自らの外界からより多くの情報を効率良く収集する能力を持つ。この行動には複数の行動選択肢のそれぞれで得られる情報の量を評価する機能が必要である。この機能はこれまで心理学および経済学で研究されてきた。この機能の脳内機構を明らかにできれば、ヒトと同様の外界から情報を効率的に収集する情報機器を設計することができる。本研究は情報量評価の神経機構を明らかにして新しい情報処理機構の設計原理を提示するとともにヒトの情報処理に関連する疾患の機序解明の基礎的知見を与えるものである。