Application to biomedical signals - Practical Biomedical Signal Analysis Using MATLAB - page 180

4.2.2 Heart rate variability
The sequence of RR intervals (sometimes called normal-to-normal (NN)
intervals)—that is, all intervals between adjacent QRS complexes resulting from si-
nus node depolarizations—forms the RR interval time series or RR tachogram. The
sequence of successive times, t i , i , when the R wave appeared, e.g.,

identified by applying QRS detector to the ECG signal is:

∈{ 1 , 2 ,..., n }

RR i = t i − t i − 1 (4.40)

A corresponding sequence of instantaneous heart rate is defined as:

1
RR i

ff i = (4.41)

Heart rate variability (HRV) is a physiological phenomenon of the variation of the
time interval between heartbeats. For determination of HRV series ECG should be
appropriately sampled. A low sampling rate may produce a jitter in the estimation
of the R-wave fiducial points. The optimal recommended range is 250 to 500 Hz or
perhaps even higher, while a lower sampling rate (in any case

100 Hz) may behave
satisfactorily only if a properly chosen algorithm of interpolation is used to refine the
R-wave fiducial point.
Since each RR interval is related to the previous one, the RR tachogram is un-
evenly sampled; therefore it has to be resampled to produce evenly sampled time
series. Common resampling schemes involve linear or cubic spline interpolative sam-
pling with resampling frequencies between 2 and 10 Hz. This procedure may intro-
duce some bias in estimating power spectra. It was reported that the bias is lower for
the procedure based on cubic splines than in case of linear resampling [Clifford and
Tarassenko, 2004].
The Task Force of the European Society of Cardiology and the North American
Society of Pacing Electrophysiology [Malik, 1996] provided an extensive overview
of standards of measurement, physiological interpretation, and clinical use of HRV. It
is believed that HRV is an indicator of sympathovagal interaction after acute myocar-
dial infarction [Lombardi et al., 1987] and that it is helpful in diagnosis of congestive
heart failure and predicting survival in premature babies. HRV is also relevant to the
pathological conditions not necessarily connected with the heart action, e.g., diabetic
neuropathy and susceptibility to SIDS. Since HRV is influenced by the autonomous
nervous system the phenomena connected with its functioning have an impact on
HRV. An example may be the sleep-wake cycle. HRV series may be analyzed in
time or in the frequency domain.

≥

4.2.2.1 Time-domain methods of HRV analysis

Time domain statistics are usually calculated on RR intervals without resampling.
A review of the time domain methods of HRV analysis may be found in [Malik,
1996]. The basic indices involve: the standard deviation of RR intervals (which may
be calculated over a 24-hour period, or short, usually 5 minutes, periods), the square