Abstract:

To accurately characterize and evaluate the topology of brain functional networks, how to define the nodes and edges of network is critical. Brain functional networks are constructed based on four strategies for defining nodes and two processing methods for negative connections and applied to test-retest reliability analysis of different time intervals. The comparison between absolute and positive networks aims to find out the influence of different processing methods for negative connections on the reliability of brain functional networks. The results show that absolute networks containing negative connections information exhibit relatively low long-term reliability and are limited by strategies for defining nodes. The difference of test-retest reliability among different processing methods for negative connections provides methodological suggestions for the study of brain functional networks.

Key words: brain functional networks, negative connections, test-retest reliability

摘要：

为了准确刻画和评估脑功能网络的拓扑结构，如何定义网络的节点和边缘至关重要。基于四种节点定义策略和两种负连接处理方法来构建脑功能网络，并对其进行不同时间间隔的重测信度分析，通过绝对值与正相关网络的比较旨在发现负连接对脑功能网络可靠性的影响。实验结果表明，包含负连接信息的绝对值网络表现出相对较低的长期可靠性，并受到节点定义策略的限制。不同负连接处理方法的重测信度差异为人脑的功能网络研究提供了方法学建议。

关键词: 脑功能网络, 负连接, 重测信度