SPECTRAL ANALYSIS DETAILS


DEVICES AND METHODS

Data collection

- ECG is recorded with a standard apparatus that provides a dc analogue output
- Blood pressure is recorded non-invasively by finger photo-pletysmography (Finapres, Ohmeda 2300®).
- All signals are fed into an MS-DOS computer (12 bit A/D conversion) at a sampling rate of 1000 Hz, with the resolution of 1m V for ECG and 0,1 mmHg for Finapres

Signal processing

Off line beat to beat analysis are performed on the stored signals and time series of successive values of RR interval, systolic (SAP) and diastolic (DAP) arterial pressure. The time series are checked for ectopic beats, the values of which are substituted by linear interpolation of adjacent beats. In addition, significant trends are removed by subtracting from the time series the best-fitting regression line. An autoregressive monovariate model are fitted to each time series (Bartoli et al, 1985) and the power and central frequency associated to each spectral peak are automatically quantified by computation of the residuals (Johnsen and Andersen, 1978). Thus, we obtain, both for heart period and blood pressure, low frequency powers (LF) and high frequency powers (HF). Transfer function analysis are performed by fitting to the time series a bivariate autoregressive model, to quantify the frequency-related squared coherence, the phase shift and transfer function gain between RR and SAP. Discrete values of phase shift and transfer function gain between RR and SAP in the LF frequency region (TFG_LF) were taken at the frequency corresponding to the highest coherence value, where the estimate error is at minimum (Kay, 1991). If coherence is <0.5 and the phase shift is negative, TFG_LF may be used as an index of baroreflex sensitivity (Robbe et al. 1987).

REFERENCES

- Bartoli F, Baselli G, Cerutti S. Ar identification and spectral estimate to the R-R interval measurements. Int. J Bio-Medical Computing 16 : 201-215 (1985)
- Johnsen SJ, Andersen N. On power estimation in maximum entropy spectral analysis. Geophysics 43 : 681-690 (1978)
- Kay SM. Modern spectral estimation: theory and application. Prentice-Hall Inc., Englewood Cliffs, New Jersey (1991)
- RobbeHWJ, Mulder LJM, Ruddel H, Langewitz WA, Veldman JBP, Mulder G. Assessment of baroreceptor reflex sensitivity by means of spectral analysis. Hypertesion 15 : 538-543 (1987)

For further details see also:

- Grasso R, Schena F, Gulli G, Cevese A. Does low frequency variability of heart period reflect a specific parasympathetic mechanism? J Autonom Nerv Syst 63 : 30-38 (1997)
- Terziotti P., Schena F., Gulli G., Cevese A Post-exercise recovery of autonomic cardiovascular control: a study by spectral and cross-spectral analysis in humans. Eur J Appl Physiol 84: 187-194 (2001)

back