Background and Aim: Nonlinear analysis of cardiovascular variability se-ries has been recognized as a valid tool for the assessment of health and dis-ease states. Recent modeling advances successfully derived time-varying es-timates of nonlinear heartbeat dynamics, whose quantifiers mainly rely on first-order moments (i.e., average in time). Nevertheless, while these metrics account for the information carried by the tonic (slow trend) nonlinear dy-namics, they fail to quantify potentially meaningful information nested in the superimposed phasic (high-frequency) activity of the series. In this study, we investigate whether new metrics derived from phasic activity of time-varying bispectra are able to track autonomic nervous system changes as elicited by a cold-pressure test (CPT). Method: Instantaneous bispectral measures are derived from nonlinear point-process modeling of heartbeat dynamics, fitted on ECG series gathered from 22 healthy volunteers undergoing sympathetic elicitation through the CPT. Instantaneous phasic bispectral activity is derived using wavelet de-composition, and quantified using the area under the curve (AUC) and vari-ance (VAR) as a second order moment. Results: Results show that phasic components of low-frequency (LL) in-stantaneous bispectral measures significantly change between resting and CPT states. The calculated AUC and VAR of the phasic phenomenon during resting state were (6.04±5.62)10^10 and (4.22±4.03)10^8, respectively, whereas they were (2.25±1.64)10^10 and (1.17±1.02)10^8 during CPT, demonstrating statistically significant discrimination between the two before and after stimulus phases (p-values of 0.026 for AUC, and 0.045 for VAR, gathered from Wilcoxon non-parametric test for paired data). Conclusion: Phasic activations of bispectral estimates carry meaningful in-formation for the nonlinear assessment of sympatho-vagal regulation to the heart. This study poses a foundation for a novel signal processing framework.