Cristín Javier, Bartumeus Frederic, Méndez Vicenç, Campos Daniel
Grup de Física Estadística, Departament de Física, Facultat de Ciències), Universitat Autònoma de Barcelona, 08193 Bellaterra, Barcelona, Spain.
Centre d'Estudis Avançats de Blanes (CEAB-CSIC), Carrer Cala Sant Francesc 14 17300 Girona, Spain.
R Soc Open Sci. 2020 Dec 16;7(12):201250. doi: 10.1098/rsos.201250. eCollection 2020 Dec.
Emergence of collective, as well as superorganism-like, behaviour in biological populations requires the existence of rules of communication, either direct or indirect, between organisms. Because reaching an understanding of such rules at the individual level can be often difficult, approaches carried out at higher, or effective, levels of description can represent a useful alternative. In the present work, we show how a spin-glass approach characteristic of statistical physics can be used as a tool to characterize the properties of the spatial occupancy patterns of a biological population. We exploit the presence of pairwise interactions in spin-glass models for detecting correlations between occupancies at different sites in the media. Such correlations, we claim, represent a proxy to the existence of planned and/or social strategies in the spatial organization of the population. Our spin-glass approach does not only identify those correlations but produces a statistical of the system (at the level of occupancy patterns) that can be subsequently used for testing alternative conditions/hypothesis. Here, this methodology is presented and illustrated for a particular case of study: we analyse occupancy patterns of ants during foraging through a simplified environment consisting of a discrete (tree-like) artificial lattice. Our spin-glass approach consistently reproduces the experimental occupancy patterns across time, and besides, an intuitive biological interpretation of the parameters is attainable. Likewise, we prove that pairwise correlations are important for reproducing these dynamics by showing how a null model, where such correlations are neglected, would perform much worse; this provides a solid evidence to the existence of superorganism-like strategies in the colony.
生物群体中集体行为以及类似超个体行为的出现，需要生物体之间存在直接或间接的交流规则。由于在个体层面理解这些规则往往很困难，在更高或有效描述层面进行的研究方法可能是一种有用的替代方法。在本研究中，我们展示了统计物理学特有的自旋玻璃方法如何用作表征生物群体空间占用模式属性的工具。我们利用自旋玻璃模型中的成对相互作用来检测介质中不同位置占用情况之间的相关性。我们认为，这种相关性代表了群体空间组织中存在计划和/或社会策略的一种指标。我们的自旋玻璃方法不仅能识别这些相关性，还能生成系统的统计量（在占用模式层面），随后可用于测试替代条件/假设。在此，针对一个特定的研究案例介绍并说明了这种方法：我们通过一个由离散（树状）人工晶格组成的简化环境，分析蚂蚁觅食过程中的占用模式。我们的自旋玻璃方法始终能重现随时间变化的实验占用模式，此外，还能对参数进行直观的生物学解释。同样，我们通过展示忽略这种相关性的零模型表现要差得多，证明了成对相关性对于重现这些动态至关重要；这为蚁群中存在类似超个体策略提供了确凿证据。
