Integrative understanding of various symptoms of psychiatric disorders by constructive approach using robots
| Project/Area Number |
17K12754
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Soft computing
|
|---|
| Research Institution | National Institute of Informatics (2018-2019)
Waseda University (2017) |
|---|
Principal Investigator |
MURATA Shingo 国立情報学研究所, 情報学プリンシプル研究系, 助教 (80778168)
|
|---|
| Project Period (FY) |
2017-04-01 – 2020-03-31
|
| Project Status |
Completed (Fiscal Year 2019)
|
| Budget Amount *help |
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2018: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2017: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000)
|
|---|
| Keywords | 計算論的精神医学 / 認知ロボティクス / 神経回路モデル / ニューラルネットワーク / 予測符号化 |
|---|
| Outline of Final Research Achievements |
This research project aims at integrative and system-level understanding of the brain and neural mechanisms that produce various symptoms of psychiatric disorders by a constructive approach using robots. We hypothesize that the prediction error minimization considering uncertainty estimation is the fundamental computational principle in the neural networks of the brain. We developed a hierarchical recurrent neural network model based on the principle and evaluated the hypothesis by implementing the model into a robot. Experimental results on robot behavior learning demonstrate that aberrant uncertainty estimation decreases or increases prediction error signals, resulting in abnormalities in perception and action. Furthermore, the results also suggest that the aberrant uncertainty estimation is derived from disconnection between hierarchies and homogenization of neural activities.
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究成果は，不確実性を考慮した予測誤差最小化が人間の認知や行動における柔軟性の実現に欠かせない基本的な計算原理であり，その失調によって精神障害における多様な症状が理解できるということを示した．これまでの精神障害を対象とした研究において，神経レベル，情報処理レベル，認知・行動レベル等，異なるレベルにおいて様々な知見・理論が提案されてきた．本成果は，神経科学・計算論・身体性認知科学といった観点を統合したロボット構成論的アプローチによって，これらの知見・理論の橋渡しをするものであり，精神障害の統合的理解に貢献することが期待される．
|
Report
(4 results)
Research Products
(15 results)
