abstract: Intrinsic dynamical noise is a fundamental component of neurophysiological systems, which are characterized by underlying complex nonlinear dynamics. While informative, such noise may bias complexity assessment using real data. Moreover, estimating this noise is challenging without specific knowledge of system behavior. This talk introduces informative randomness in complex systems and showcase model-free methods to estimate physiological noise using nonlinear entropy profile, applicable without assumptions about system dynamics. The methodological framework is applied to real cardiovascular and brain data gathered from from healthy and pathological subjects.