Nonlinear Method to Assess Autonomic Modulation During Controlled Breathing

This paper addresses the problem of characterizing cardiovascular signals describing the autonomic nervous system (ANS), considering the nonlinearity and complexity of its dynamics. We propose three approaches to reveal the dynamics of the ANS: joint symbolic analysis (JSA), phase-rectified signal averaging (PRSA), and Fuzzy Entropy (FuzzyEn). The analyzed dataset consisted of 34 young volunteers, in whom arterial blood pressure was measured non-invasively during controlled breathing sessions (6,10, and 15 breaths $/ \mathrm{min}$). The results show that quasiperiodicities in the tachogram, quantified using PRSA as heart acceleration and deceleration capacity, decreased with breathing frequency ($p(\lt)0.001$ for both). Cardiovascular interaction, quantified by JSA as the relative frequency of symmetric word types, decreased with breathing frequency ($\mathbf{p}\lt \mathbf{0. 0 0 1}$). FuzzyEn increased with breathing frequency but did not correlate with PRSA and JSA metrics. Our results demonstrate that JSA, PRSA, and FuzzyEn are sensitive to ANS modulation during breathing frequency challenges, providing complementary information about ANS modulation.