Non-invasive model-based estimation of the sinus node dynamic properties from spontaneous cardiovascular variability series.

A non-invasive model-based approach to the estimation of sinus node dynamic properties is proposed. The model exploits the spontaneous beat-to-beat variability of heart period and systolic arterial pressure and the sampled respiration, thus surrogating the information from direct measures of neural activity. The residual heart period variability not related to baroreflex, to direct effects of respiration and to low frequency influences independent of baroreflex, is interpreted as the effect of the dynamic properties of the sinus node and modelled as a regression of the RR interval over its previous value. Therefore the sinus node transfer function is modelled by means of a filter with a real pole z = mu (and a zero in the origin). It was found that: first, in young healthy subjects the nodal tissue responded as a low-pass filter with mu = 0.76 +/- 0.12 (mean +/- SD); secondly, ageing did not significantly modify either its shape or gain at 0 Hz; thirdly, in heart transplant recipients, the dynamic transduction properties were lost (all-pass filter, p = 0.06 +/- 0.16, p < 0.001); fourthly, low-dose atropine left the sinus node dynamic properties unmodified; fifthly, high-dose atropine affected the dynamic transduction properties by increasing the gain at 0 Hz and rendering steeper its roll-off (the percent increase of mu with respect to baseline was 18.3 +/- 22.3, p < 0.05).