Closed-Loop Dynamic Modeling of the Heart-Rate Reflex to Concurrent Spontaneous Changes of Arterial Blood Pressure and CO2 Tension: Quantification of the Effects of Mild Cognitive Impairment

Objective: To extend closed-loop modeling of the heart-rate reflex (HRR) by including the dynamic effects of concurrent changes in blood CO2 tension. This extended dynamic model can be used to generate physio-markers of 'baroreflex gain' (BRG) and 'chemoreflex gain' (CRG) that allow quantitative assessment of the possible impact of pathologies upon HRR. Mild Cognitive Impairment (MCI) is used as an example. Methods: The proposed data-based closed-loop modeling methodology estimates the forward and reverse dynamic components of the model via Laguerre kernel expansions of two open-loop models using spontaneous time-series data collected in 45 MCI patients and 15 controls. The BRG and CRG physio-markers are subsequently computed for each subject via simulation of the obtained closed-loop model for unit-step change of arterial pressure or blood CO2 tension, respectively. Results: Both open-loop and closed-loop HRR modeling revealed that MCI patients exhibit significantly smaller CRG relative to controls (p<0.001), but not significantly different BRG. Furthermore, the closed-loop model captured the dynamic effect of sympathetic activity as resonant peak around 0.1 Hz (Mayer wave) in the chemoreflex and baroreflex transfer functions (not captured via open-loop modeling). This may prove valuable in advancing our understanding of how sympathetic activity impacts HRR in various pathologies. Conclusion: The extended HRR model, incorporating the dynamic effects of concurrent changes of blood CO2 tension, revealed significantly reduced chemoreflex gain (but not baroreflex gain) in MCI patients. Furthermore, the closed-loop model captured the sympathetic influence around 0.1 Hz. Significance: Multivariate closed-loop dynamic modeling is valuable for understanding physiological autoregulation.