Development of high-speed general-purpose cell sorting method by sparse modeling and image processing

2021 Fiscal Year Final Research Report

• Project/Area Number: 19K12104
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 61030:Intelligent informatics-related
• Research Institution: Kogakuin University
• Principal Investigator: Takekawa Takashi 工学院大学, 情報学部(情報工学部), 准教授 (50415220)
• Project Period (FY): 2019-04-01 – 2022-03-31
• Keywords: 蛍光イメージング / 信号検出 / 細胞検出 / スパース最適化 / 非負行列因子分解

Outline of Final Research Achievements

In recent years, fluorescence imaging technology that can observe the activity of a large number of neurons has been greatly improved, and it is expected to greatly enhance our understanding of the brain. In order to utilize the data, it is important to have a technology to automatically determine the position and activity state of cells from the recorded data, but existing systems have not been sufficient in terms of accuracy, speed, and stability. In this study, we developed a system that significantly outperforms existing systems in terms of accuracy and speed by making full use of image processing, stochastic models, optimization algorithms, and parallel computing technology, and released the software. The developed system, HOTARU, is capable of detecting spikes with high accuracy, even for cells that are difficult to detect using existing methods.

Free Research Field

計算論的神経科学

Academic Significance and Societal Importance of the Research Achievements

本研究で開発したシステム HOTARU は，恣意的なパラメータを持たずデータの性質に依存せず安定して動作し，発火率が低い，あるいは SN が小さい細胞についても検出することが可能である.この手法により，従来法では細胞強度のみで細胞の妥当性を判断していたのに対して，細胞の形状が安定していること，検出されるスパイクの数が妥当であることを元に改善が可能となった.また，システムの動作に必要なパラメータは必要最小限で恣意的な要素を排除している。そのため，神経回路の動作を実験データから検証する際に誤った結論を導く可能性を大幅に低減することができた。