Noninvasive subject-specific monitoring of autonomic-cardiac regulation.

This paper presents a feasibility study of a model-based approach to noninvasive and subject-specific monitoring of autonomic-cardiac regulation. The proposed approach is built upon individualizing a physiologically-based model by applying a parameter estimation method to routine clinical observations, thereby assuring physical transparency, computational efficiency, and clinical adaptability. To develop an efficient parameter estimation procedure, a parametric sensitivity analysis was performed on the autonomic-cardiac regulation model to identify high-sensitivity model parameters whose changes exert significant impacts on the system outputs. Then, a parameter estimation problem formulated as a nonlinear optimization was solved to estimate high-sensitivity model parameters associated with autonomic-cardiac regulation, whereas the remaining parameters were fixed at their nominal values. The proposed approach can potentially monitor temporal changes in autonomic-cardiac regulation by identifying time-varying changes in the autonomic-cardiac model parameters, including sympathetic and parasympathetic nerve activities on the heart (modulating heart rate), and sympathetic nerve activity on the arterial tree (modulating total peripheral resistance). The proof-of-concept for the proposed approach was tested using a number of experimental data from the MIMIC database and the orthostatic hypotension tests. Our finding shows that the proposed approach is able to provide low-variance estimates of the autonomic-cardiac model parameters, which are consistent with their anticipated behaviors inferred from the physiologic knowledge. An extensive comparison study must be conducted in the future to establish the clinical validity of the proposed approach.