Development of fast and accurate functional cluster/hub cell detection method in brain network

2022 Fiscal Year Final Research Report

• Project Area: Morphological features and gene expression patterns underlying hub neurons
• Project/Area Number: 20H05776
• Research Category: Grant-in-Aid for Transformative Research Areas (B)
• Allocation Type: Single-year Grants
• Review Section: Transformative Research Areas, Section (III)
• Research Institution: Ibaraki University
• Principal Investigator: Takeda Koujin 茨城大学, 理工学研究科(工学野), 准教授 (70397040)
• Project Period (FY): 2020-10-02 – 2023-03-31
• Keywords: 機能的神経クラスタ / 神経ネットワーク / 神経活動特徴量抽出 / ベイズ推定 / マルコフ連鎖モンテカルロ法 / 行列分解

Outline of Final Research Achievements

The main results of this research project are summarized as follows. (A) The estimation algorithm for neuronal ensemble based on Bayesian statistics was accelerated, generalized for application to continuous-valued data, and further extended to analyze non-stationary neuronal ensemble structure. (B) The algorithm to simultaneously estimate the position and spike time series of neurons in the picture of Ca2+ imaging was developed. By applying this algorithm, it was confirmed that this algorithm can estimate both the position and spike time series with high accuracy. (C) Various matrix factorization methods were applied to extract features in fMRI data of brain activity. As a consequence, it was found that the method incorporating sparsity can extract appropriate features. This result indicates that sparse coding is realized in information processing in the brain.

Free Research Field

統計物理学

Academic Significance and Societal Importance of the Research Achievements

神経科学では近年の計測技術の向上によりCa2+イメージング画像やfMRIデータ等の大規模な神経活動データが取得可能となった。しかし大規模データから神経活動の有意な情報を抽出するための数理的技術は発展途上であった。本研究の主要な成果により、神経科学においても大規模データの高精度・高速な処理が近い将来に可能になり、今後の新たな解析手法の開発も期待される。また大規模データに対する新手法を活用することで、生物の神経ネットワーク構造および各部位の機能に関する知見が得られ、最終的には神経科学の発展に大きく寄与できると考えられる。