| Project/Area Number |
21K19781
|
|---|
| Research Category |
Grant-in-Aid for Challenging Research (Exploratory)
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Medium-sized Section 61:Human informatics and related fields
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2021-07-09 – 2024-03-31
|
| Project Status |
Completed (Fiscal Year 2023)
|
| Budget Amount *help |
¥6,500,000 (Direct Cost: ¥5,000,000、Indirect Cost: ¥1,500,000)
Fiscal Year 2023: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Fiscal Year 2022: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Fiscal Year 2021: ¥3,380,000 (Direct Cost: ¥2,600,000、Indirect Cost: ¥780,000)
|
|---|
| Keywords | 身体運動 / 深層学習 / マウス / 機械学習 / 強化学習 / 行動実験 / 神経活動計測 |
|---|
| Outline of Research at the Start |
ロボットが行動を学習する理論として強化学習がある．中脳ドーパミン細胞による強化学習の実装が発見されて以来，脳の意思決定研究は，神経細胞と強化学習の関係を調べてきた．しかし，強化学習は，動物の行動選択を100%説明できるわけではない．本研究は，脳の意思決定モデルの再構築に向けて，深層学習で，マウスの行動選択を予測する．情報学的手法と，マウスの神経活動計測を組み合わせて，脳の意思決定の新因子同定を目指す．
|
| Outline of Final Research Achievements |
This study used a machine learning and large-scale measurement of physical body movements of mice, and investigated whether the body movement predicted the future action selection of mice. During a tone frequency discrimination task with head-fixed mice, we measured the physical body movements with four or five video cameras. We used DeepLabCut to extract the XY trajectories of 30 or 40 major body parts of mice from the movie data.
We used a machine-learning tool and predicted the left- or right-spout choice during the task from the trajectories of physical movements of mice. We found that the body movement predicted the action selection of mice at least 2.5 seconds before the choice. Our next step predicts the action selection of mice from the neural activity recorded with electrophysiology or two-photon microscopy. We are going to identify the neural activity that predict choices faster than the physical movements to identify the neural basis of decision making.
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究は，脳の意思決定の神経基盤解明に向けて，マウスで，身体運動の大規模計測と解析を実施した．その結果，マウスの身体運動から，2.5秒後の行動選択を予測できた．今後の神経活動解析は必要だが，本研究は，2.5秒前の身体や脳の状態が，行動選択に影響することを示唆する．意思決定とは，脳の高次機能であり，その不具合は，高次機能障害につながる．また，意思決定の不調は，精神疾患とも関連が深いと考える．さらに，脳の意思決定機構の解明は，感覚情報処理や行動選択を伴う将来の脳型人工知能の開発につながる．本研究の成果は，医学・工学・情報学の発展に寄与する．
|
