Elucidation of cognitive and learning mechanism of cerebral cortex by multiscale optogenetics

2023 Fiscal Year Final Research Report

• Project/Area Number: 19H05642
• Research Category: Grant-in-Aid for Scientific Research (S)
• Allocation Type: Single-year Grants
• Review Section: Broad Section G
• Research Institution: The University of Tokyo
• Principal Investigator: Ohki Kenichi 東京大学, 大学院医学系研究科(医学部), 教授 (50332622)
• Co-Investigator(Kenkyū-buntansha): 松井 鉄平 同志社大学, 脳科学研究科, 教授 (10725948)
• Project Period (FY): 2019-06-26 – 2024-03-31
• Keywords: 視覚野 / イメージング / 光遺伝学 / 神経回路 / 情報処理

Outline of Final Research Achievements

We developed a functional imaging method for spines using optogenetics to elucidate the mechanisms of orientation and phase selectivity formation in cells of the visual cortex. We developed a method to systematically investigate the complex receptive fields of cells in higher visual cortex using deep learning. We elucidated the mechanisms and their learning rules by which the connections between cortical areas are precisely formed during development. We reconstructed the visual information represented in cell populations in the visual cortex, paving the way for understanding perception in normal and diseased animals. We trained mice in a visual discrimination task for novel shapes and found that with learning, a sparse representation pattern emerges. We developed a method to activate and inactivate cell populations while measuring neural activity in the whole brain, and were able to artificially induce perception and plasticity using light stimulation.

Free Research Field

神経科学

Academic Significance and Societal Importance of the Research Achievements

視覚野の細胞集団に表現されている視覚情報を再構成する方法の開発は、マウスが何を見ているのかを理解する道を開き、今後の正常・疾患マウスを用いた知覚研究に対して強力な方法論となる。深層学習を用いた視覚野の神経細胞の非線形で複雑な受容野を系統的に調べる方法の開発については、実際の脳活動をニューラルネットワーク上に写しこんで解析しており、脳活動を模した脳型人工知能の開発にもつながる。大脳皮質の領野をつなぐ階層的か並列的なネットワークが発達時に精密に形成されるメカニズムを初めて解明したことは、網膜からの活動伝播の重要性を示唆し、早期失明の病態理解にもつながる。