6.3 Signals of the Cardiovascular System

|

259

In addition to these two main clinical applications, HRV has also been studied in

relation to various cardiovascular conditions such as renal failure, exercise, occupa-

tional and psychosocial stress, gender, age, drugs, alcohol, smoking and sleep. The

term HRV analysis generally refers to changes in the heartbeat interval to investigate

precursors of disease.

The study of cardiac rhythm variability can be performed in both the time and

frequency domains, as well as in the time-frequency domain.

Time domain

Statistical methods are used for the investigation in the time domain, including i) the

determination of statistical parameters, such as the mean duration of all RR intervals

and their standard deviation, ii) the performance of analysis of variance , such as the

frequency of occurrence of different heart rates in the time domain. e.g. the frequency

of occurrence of different heart rates in a histogram, and iii) the use of correlation

methods such as the changes in heart rate HFn between successive RR intervals ac-

cording to Poincaré (see Figure 6.31).

Frequency Domain

When investigating in the frequency domain, for example, long and short time-Fourier

analysis as well as a wavelet-transformation can be applied to the course of the heart

rate in the time domain. The wavelet transformation has the advantage that the time

interval in which the transformation is to be carried out is adapted to the frequency

range, i.e. a large time interval is selected for the analysis of low frequencies and a

small time interval for the analysis of high frequencies (cf. section 2.4). For the short-

time-Fourier transform and the wavelet-transform, different "windows" can be used

to select the time range, which must be examined for optimal analysis behaviour de-

pending on the signal type.

For Analysis in the Time Domain

In addition to the basic parameters such as mean and standard deviation

E[HF] = ¹

N

N

∑

n=1

HFn ,

mean , N : number of frequency values

σ²[HF] = ¹

N

N

∑

n=1

(HFn −E[HF])² ,

standard deviation

(6.26)

the frequency of occurrence of different heart rate values is often also presented in the

form of a histogram (cf. Figure 6.32), the width of which is additionally a measure of

variability.

As an alternative to the histogram, the variability of the heart rate can also be seen

very well, as already mentioned, in a Poincaré-diagram, in which the heart rates are