Fourier-domain transfer entropy spectrum

We propose the Fourier-domain transfer entropy spectrum, a generalization of entropy, as model-free metric causality. For arbitrary systems, this approach systematically quantifies causality among their different system components rather than merely analyzing systems entireties. The generated spectrum offers rich-information representation time-varying latent causal relations, efficiently dealing with nonstationary processes and high-dimensional conditions. demonstrate its validity in aspects parameter dependence, statistical significance tests, sensibility. An open-source multiplatform implementation is developed computationally applied on neuroscience data sets diffusively coupled logistic oscillators.Received 12 October 2021Accepted 30 November 2021DOI:https://doi.org/10.1103/PhysRevResearch.3.L042040Published by American Physical Society under terms Creative Commons Attribution 4.0 International license. Further distribution work must maintain attribution to author(s) published article's title, journal citation, DOI.Published SocietyPhysics Subject Headings (PhySH)Research AreasComplex systemsSynchronizationTechniquesInformation theoryProbability theoryTime series analysisInterdisciplinary PhysicsNonlinear DynamicsStatistical PhysicsNetworks