In living systems, irreversible, yet stochastic, molecular interactions form multiscale structures (such as cytoskeletal networks), which mediate processes (such as cytokinesis and cellular motility) in a close relationship between the structure and function. However, owing to a lack of methods to quantify non-equilibrium activity, their dynamics remain poorly characterized. Here, by measuring the time-reversal asymmetry encoded in the conformational dynamics of filamentous single-walled carbon nanotubes embedded in the actomyosin network of Xenopus egg extract, we characterize the multiscale dynamics of non-equilibrium activity encoded in bending-mode amplitudes. Our method is sensitive to distinct perturbations to the actomyosin network and the concentration ratio of adenosine triphosphate to adenosine diphosphate. Thus, our method can dissect the functional coupling of microscopic dynamics to the emergence of larger scale non-equilibrium activity. We relate the spatiotemporal scales of non-equilibrium activity to the key physical parameters of a semiflexible filament embedded in a non-equilibrium viscoelastic environment. Our analysis provides a general tool to characterize steady-state non-equilibrium activity in high-dimensional spaces.

在生命系统中，不可逆但随机的分子相互作用形成多尺度结构（例如细胞骨架网络），其在结构和功能之间的密切关系中介导过程（例如胞质分裂和细胞运动）。然而，由于缺乏量化非平衡活动的方法，它们的动态特征仍然很差。在这里，通过测量嵌入爪蟾卵提取物肌动球蛋白网络中的丝状单壁碳纳米管构象动力学编码的时间反转不对称性，我们表征了弯曲模式振幅编码的非平衡活动的多尺度动力学。我们的方法对肌动球蛋白网络的明显扰动以及三磷酸腺苷与二磷酸腺苷的浓度比敏感。因此，我们的方法可以剖析微观动力学与更大规模非平衡活动的出现的功能耦合。我们将非平衡活动的时空尺度与嵌入非平衡粘弹性环境中的半柔性细丝的关键物理参数联系起来。我们的分析提供了表征高维空间中稳态非平衡活动的通用工具。