Role of causality in the evaluation of coherence and transfer function between heart period and systolic pressure in humans

To elicit the effects of considering causality in the study of the interactions between RR interval and systolic pressure (SP) variability, the traditional noncausal cross-spectral analysis was compared with a causal method able to separate the two arms of the RR-SP regulatory loop. Estimates of coherence (K) and causal coherences from SP to RR (Ksr) and from RR to SP (Krs), and of noncausal (G) and causal (Gsr) baroreflex gain were evaluated at 0.1 Hz in 10 healthy young subjects in the supine position and after head-up tilt. While K was high in both conditions, at rest Ksr was significantly lower than Krs. After tilt, Ksr increased and Krs decreased significantly. With respect to G, Gsr was significantly lower at rest and comparable after tilt. Thus, in the supine position the prevalence of non-baroreflex coupling mechanisms led the traditional approach to overestimate the baroreflex gain, while the tilt-induced activation of the baroreflex control made similar the gain estimates. © 2004 IEEE.