Fluctuating Physiological Noise in Cortical Activity During Visual Stimulation

This study introduces a novel time-resolved identification of non-stationary physiological noise, specifically applied to complex biomedical signals such as electroencephalographic (EEG) data. By modeling physiological noise as a dynamical recursive realization of independent and identically distributed (IID) Gaussian random variables, we employ an information-theoretic quantifier, the Approximate Entropy, to estimate noise power using a sliding window approach. The method's accuracy is validated on synthetic time series with varying dynamic noise levels, demonstrating reliable performance even with short window lengths. Applying this approach to real EEG recordings during resting-state and visual stimulation reveals significant time-resolved fluctuations in physiological noise. Notably, we observe a reduction in noise levels during visual stimulation in both occipital and frontal cortical regions. This finding aligns with previous studies indicating decreased complexity alongside increased dynamic activity in these regions, which are associated with visual processing and emotional expression. This preliminary work highlights the robustness of the proposed method and its potential to improve our understanding of physiological noise dynamics, offering implications for more precise cortical signal analysis and possible clinical applications.Clinical relevancePhysiological noise may serve as a novel biomarker for characterizing various pathophysiological conditions, especially linked to brain and cardiovascular dynamics.