Abstract:
Spectral parameters extracted from the heart rate variability (HRV) signal are obtained
on a beat-to-beat basis, following a procedure which uses two recursive algorithms.
In the first step of the procedure the set of the AR model coefficients is updated
each time a new RR value is available. Then from the estimated AR model parameters,
the new position of the poles of the model transfer function in the complex z-plane
is evaluated and, finally, through a residual calculation, it is possible to calculate
the spectral parameters which quantify the control of the autonomic nervous system
in assessing the cardiac frequency (i.e., power and frequency of LF and HF components).
The whole procedure has first been tested on a simulated time series, in order to
evaluate its performance in tracking the dynamic changes during different conditions;
next the algorithms were employed in the study of the HRV signal for continuous monitoring
of non-stationary conditions.
Keywords
Time-variant Spectral Analysis - Auto-Regressive Models - Heart Rate Variability -
Autonomic Nervous System
