生物反馈回路是生物网络中广泛存在的"动态模体"，在众多生命过程中发挥着关键调控作用。本项目针对生物神经网络中反馈回路模体辨识问题，提出一种新的辨识方法-多步格兰特因果性检验辨识法，该方法能够仅利用多电极阵列获得的多通道电脉冲序列数据，辨识出网络中功能性反馈连接。其基本思想为：运用Wald检验，为反馈辨识提供一个便利准则，在统计分析下，若两网络节点间存在显著的双向多步格兰特因果性，则表明该对节点间具有反馈回路。由于借鉴了多变量时间序列分析法，该方法克服了以往生物反馈辨识法大都依赖于两变量分析的局限性，不再受具体支路连接形式的限制，特别适用于大规模生物网络分析。最后，将多步格兰特因果性检验辨识法应用于体外培养的大鼠脑皮质神经元网络产生的脉冲序列数据，辨识其中反馈回路模体的存在情况、性质与分布特点，从系统水平阐明生物神经网络中支配其特定网络行为（如，同步脉冲振荡和过激振荡等）的反馈连接结构特点。

Biological feedback loops are network motifs that are ubiquitously found in many biological networks and induce plenty of crucial regulatory and control functions. We shall propose a novel method, called MGTIM (Multi-step Granger causality test identification method), by which the feedback loops embedded in biological neural networks can be identified using the spike train data measured by MEA (Multi-electrode array) devices. The basic idea of MGTIM is to employ a modified Wald test for inferring the existence of multi-step Granger causality between a pair of network nodes. If significant bi-directional multi-step Granger causalities are found between two nodes, this implies the existence of a feedback loop between the two nodes. Because of avoiding the limitations of traditional two-variant analysis methods, MGTIM is free of network connective forms and is especially suitable for large-scale network analysis. Finally, the proposed identification method will be applied to in vitro cultured rat cortical neural networks for elucidating relationship between their specific network behaviors (for examples, synchronized bursting activities and hyper-excitable activities) and existence, property, and distribution of underlying feedback loops.