Spectral decomposition of irreversibility reveals structure of nonequilibrium activity in biological systems

Biological systems, such as cytoskeletal networks, exhibit stochastic mechanical fluctuations on mesoscopic scales which can violate detailed balance. The spatiotemporal structure of nonequilibrium activity on these scales remains unexplored, due to a lack of methods able to reliably quantify irreversibility. To probe activity in both space and time, we image spatially extended single-walled carbon nanotubes (SWNTs) embedded in actin-intact Xenopus cytoplasmic extract. Using normal mode decomposition of filament shape fluctuations, we infer the structure of the actomyosin-driven mechanical fluctuations. Metrics for irreversibility based on normal mode correlation functions quantify the spatiotemporal extent of nonequilibrium activity. To estimate the noise floor of our analysis, we compare our results to the fluctuations of SWNTs in an equilibrium, entangled F-actin gel. By altering network architecture and generating chemostatted ATP reservoirs, we probe the response of nonequilibrium activity to distinct perturbations. Our analysis quantifies the spatiotemporal structure of irreversibility on mesoscopic scales and shows it is affected by network mechanics and its coupling to the ATP chemical reservoir.