Cardiovascular and cardiorespiratory interactions are characterized by complex and intricate dynamics that lead to the emergence of redundant and synergistic behaviors among the involved physiological systems. The Partial Information Decomposition (PID) approach has been introduced to decompose the information shared among multiple interacting variables in non-negative measures describing the redundant, synergistic, and unique content of the overall information exchanged among the variables. Until now, PID has been defined only for discrete or continuous Gaussian variables, thus limiting its applicability to categorical data or to linear interactions. In this paper, a nearest-neighbor estimation approach was introduced for quantifying the PID terms when working with a discrete target variable, representing specific states for a process Y, and a pair of continuous source variables X1 and X2. The proposed method was first investigated on the Boolean logic XOR-gate obtained by looking at the sign of two simulated uniformly distributed processes, and then applied to heart period, systolic arterial pressure and respiratory variability series measured in healthy subjects in the supine resting position and during head-up tilt. Both on simulated and physiological data, the results demonstrated how the PID framework allows to distinguish for the role of each of the sources and the contributions emerging from their interaction, even when information shared variations remain hidden when looking at the whole system. Specifically, the analysis of the information mutually exchanged by the cardiac and vascular dynamics with inspiratory/expiratory breathing activity highlights an overall reduction of the information shared within the investigated physiological network in response to postural stress, evidencing the leading role of cardiorespiratory regulatory mechanisms, an opposite influence of the vasculo-respiratory dynamics, and an important redundant contribution of the interaction mechanisms involving both cardiac and vascular dynamics.