Optimization of retinal prosthetic stimulation by artificial intelligence

2022 Fiscal Year Final Research Report

• Project/Area Number: 19K09949
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 56060:Ophthalmology-related
• Research Institution: Osaka University
• Principal Investigator: Miyoshi Tomomitsu 大阪大学, 大学院医学系研究科, 助教 (70314309)
• Co-Investigator(Kenkyū-buntansha): 内藤 智之 大阪大学, 大学院医学系研究科, 准教授 (90403188) ; 森本 壮 大阪大学, 大学院医学系研究科, 寄附講座准教授 (00530198)
• Project Period (FY): 2019-04-01 – 2023-03-31
• Keywords: 人工網膜 / 電気刺激 / 誘発反応 / 機械学習

Outline of Final Research Achievements

We attempted to evaluate the evoked response in the brain caused by stimulation from the artificial retina of the STS type by using machine learning. By training a recurrent neural network, it was possible to estimate which stimulating electrode in the retina was stimulated, and it was thought that the characteristics of the stimulation could be evaluated using the estimation rate as an index. In the single neuron recording, there was variation in the strength of the response even under the same stimulation conditions, but this was not due to factors such as whether the stimulated site was cell body or axon, as is the case in other stimulation methods. On the other hand, Off-centered cells were found to be excited at a lower threshold than On-centered cells.

Free Research Field

感覚生理学

Academic Significance and Societal Importance of the Research Achievements

視細胞の変性疾患に対して残存細胞を電気刺激する人工網膜による視機能は、健常人の視覚よりもはるかに分解能が低く、より優れた刺激を開発する必要がある。本研究では誘発反応を使って網膜刺激を定量的に評価する方法を確立した。動物実験の結果ではあるが、将来、人工網膜を埋植した人においても、脳から誘発反応などを非侵襲的に記録し同様の方法を用いることで、刺激パターンを最適化することが原理的に可能であり、QOLの向上に資することが期待される。