Department of Biology, University of Pennsylvania, Philadelphia, PA, USA.
, West Hartford, CT, USA.
Theory Biosci. 2021 Nov;140(4):361-377. doi: 10.1007/s12064-020-00311-9. Epub 2020 Mar 23.
From fish schools and bird flocks to biofilms and neural networks, collective systems in nature are made up of many mutually influencing individuals that interact locally to produce large-scale coordinated behavior. Although coordination is central to what it means to behave collectively, measures of large-scale coordination in these systems are ad hoc and system specific. The lack of a common quantitative scale makes broad cross-system comparisons difficult. Here we identify a system-independent measure of coordination based on an information-theoretic measure of multivariate dependence and show it can be used in practice to give a new view of even classic, well-studied collective systems. Moreover, we use this measure to derive a novel method for finding the most coordinated components within a system and demonstrate how this can be used in practice to reveal intrasystem organizational structure.
从鱼群和鸟群到生物膜和神经网络，自然界中的集体系统由许多相互影响的个体组成，这些个体通过局部相互作用产生大规模的协调行为。尽管协调是集体行为的核心，但这些系统中的大规模协调措施是特定于系统的，缺乏通用的定量尺度使得进行广泛的跨系统比较变得困难。在这里，我们基于多元依赖的信息论度量方法，确定了一种独立于系统的协调度量方法，并展示了它如何在实践中被用来对即使是经典的、研究充分的集体系统提供新的见解。此外，我们使用这个度量方法来得出一种新的方法，用于在系统中找到最协调的组件，并展示了如何在实践中使用它来揭示系统内的组织结构。
