DIGITAL SIGNAL PROCESSING APPROACH TO EXTRACT HEART RATE VARIABILITY FROM QRS WAVE DERIVED FROM PATIENT BODY

yakzephyrΤεχνίτη Νοημοσύνη και Ρομποτική

24 Νοε 2013 (πριν από 3 χρόνια και 4 μήνες)

71 εμφανίσεις

JOURNAL OF INFORMATION, KNOWLEDGE AND RESEARCH IN

BIOMEDICAL ENGINEERING

ISSN: 0975


6752
|
NOV
10

TO OCT
11

|

Volume 1, Issue
2


Page
12



DIGITAL SIGNAL PROCESSING APPROACH TO
EXTRACT HEART RATE VARIABILITY FROM QRS
WAVE DERIVED FROM PATIENT BODY


PROF. MEHUL K.SHAH


Assistant Professor, IC Deptt.,Govt. Engg. College, Rajkot.


mkshah2000@yahoo.com


ABSTRACT

: The analysis of HRV (Heart Rate Variability) has become a non
-
invasive tool for assessing the
activities of ANS (Autonomic Nervous System). This paper suggests the new estimation technique of
determining the Heart Rate Variability with the
use of correlation (Auto
-
correlation and cross
-
correlation of the
sequence) obtained from the patients ECG. The ECG waveform comprises of VLF (0.0033
-
0.04 Hz), LF(0.04
-
0.15Hz) and HF(0.15
-
0.5 Hz) components in the QRS wave. In R wave detection method which

requires the
sampling frequency Fs≥100 Hz. This method suggested may be used to extract the instantaneous information
like Normal to Normal (NN) intervals, intervals between consecutive pulses. The correlation of pulses in
practical applications can be u
sed to identify periodicities in the ECG waveforms. Due to the presence of noise
picked up, the problem becomes worse. Even with the SNR of 5 dB, correlation can help to determine periodicity
of ECG.



1.
INTRODUCTION


:

The human body is like a mini pow
er generator. ECG
Electrocardiography deals with the electrical activity
of heart. With the help of placing limb electrodes on
patient’s body, these electrical activities can be
picked up. Electrocardiogram is a record of origin
and propagation of the elec
tric potential through
cardiac muscles
[6]
.


In ones life, the legs and feets are under
tremendous pressure of the body’s weight for
approximately about two thirds of time. The feet are
at the farthest distance from the heart. Consequently,
the blood whi
ch is pumped from the heart to feet and
circulated back to the heart will have an increase in
of blood difficulty in its ability to circulate. This may
lead to various ailments in the legs and feet. The
reflexology will allow the repeated movement of
muscl
es to produce a very prominent overall
pressuring action on the walls of the blood vessels in
the lower extremities. These blood vessels will have
more strength to contract and expand and will
enhance circulation of blood back to the heart. The
heart will
in turn have a greater supply of blood to
nourish the body. Hence, there will be more variation
in the heart rate and becomes more chaotic.


Last two decades have witnessed the
recognition of significant relationship between the
autonomic nervous system a
nd cardiovascular
mortality
[6]
, including sudden cardiac death. HRV
represents one of the most promising such marker.
The apparently easy derivation of this measure has
popularized its use. For the diagnosis, cardiologists
are provided with simple tool in
most modern
instruments. The phenomena of HRV are oscillations




between consecutive instantaneous heart beats.
“Heart Rate Variability” has become the most
accepted term to describe variation of both
instantaneous heart rate and RR interval
[8]
. Other
t
erms used to describe the HRV are cycle length
variability, heart period variability, RR variability
and RR interval Tachogram
[13]
. These parameter
more emphasize on the fact that it is the interval
between consecutive beats that is being analyzed
rather t
han the heart rate per se.

2.
HEART RATE VARIABILITY


HRV is a measurement of the interaction between
sympathetic and parasympathetic activity in
autonomic functioning. The main two

approaches
used for HRV are time domain which includes
standard deviation
of normal to normal intervals
(SDNN)
[13]
; and frequency domain analysis of Power
Spectrum Density (PSD). The PSD technique
provides low frequency (sympathetic activity) and
high frequency (parasympathetic activity) and
sympathetic / parasympathetic balance

values (total
power)

[12]
. Spectral analysis is the most popular
linear technique used in the high frequency band
which reflects respiratory sinus arrhythmia
[2]

and thus
cardiac vagal activity. Low frequency power is
related to baroreceptor control and m
ediated by both
vagal and sympathetic systems
[10]
. Very low
frequency power is related to thermo
-
regulatory and
vascular mechanisms, and rennin
-
angio tension
system
[1] [2]
.


3.
METHODS FOR MEASUREMENT OF HRV



TIME DOMAIN METHOD

The variation in the
hear
t rate may be calculated by number of methods.
Time domain method is the simplest of them. In time
JOURNAL OF INFORMATION, KNOWLEDGE AND RESEARCH IN

BIOMEDICAL ENGINEERING

ISSN: 0975


6752
|
NOV
10

TO OCT
11

|

Volume 1, Issue
2


Page
13



domain analysis, either the heart rate at any point in
time or the intervals between consecutive QRS pulses
is detected and the so called NN intervals. Simpl
e
time domain variables can be calculated i.e.
instantaneous heart rate, mean NN interval, mean
heart rate, difference between the longest and the
shortest NN interval, difference between the night
and day heart beats etc. These differences can be
describe
d as either differences in heart rate or cycle
length.

FREQUENCY DOMAIN METHOD

The spectral
methods for analysis of the ECG have been used
more widely. PSD method provides basic information
of how power (variance) distributes as a function of
frequency. U
sing proper mathematical algorithm, the
estimation about PSD can be derived.

Parametric and non
-
parametric are the two methods
generally classified for the calculation of PSD
[5] [9]
. In
most instances, both methods provide comparable
results. The advantag
es of non
-
parametric method are
simple FFT (Fast Fourier Transform) algorithm and
high processing speed. Parametric method has
advantages of smoother spectral components that be
distinguished independent of pre selected frequency
bands and easy post proce
ssing of the spectrum with
an automatic calculation of low and high frequency
power components with easy identification of the
central frequency of each component. With later
method, the accurate estimation even with lower
samples can be obtained. The basi
c disadvantages
with parametric methods are the need of verification
of the suitability of the chosen model and of its
complexities due to order of the model.

4.
CORRELATION
-

A DSP APPROACH TO
THE HRV

The problem of correlation is similar to the
convoluti
on except in case of correlation the sole
objective is to calculate the similarities or the
differences between two sequences. In this paper
emphasis is laid down on the heart rate variability
which is nothing but the differences between
consecutive pulse
train dynamics from the heart. The
ECG waveform derived from the heart are sampled
and quantized with sampler and a quantizer. The
sequence thus obtained may be used for comparison
with the upcoming sequences of derived ECG pulses.
The comparison of the sa
me sequence with the
delayed or advanced version of the same signal is
known as auto
-
correlation whereas with different
sequence is known as cross
-
correlation.


In notations auto
-
correlation of sequence
x(n)

with itself is referred to as r
xx

(auto correla
tion)
and with y(n) is r
xy

(cross correlation)
[4]
. If the
present ECG sequence is compared with either the
previous or the next, differences and similarities can
be calculated which is nothing but the HRV. The
correlation term is defined as
[4]
.











Where the index
l

is the time shift (lag) parameter
and the subscript
xy

on the cross correlation sequence
r
xy
(
l
) indicate sequence being correlated. The
xcorr
(x,y)

function in MATLAB is used to compute
the cross correlation of two sequences x(n)
and y(n).
To compute auto correlation of sequence
xcorr
(x)

is
used.

5.
CONCLUSION

As conventionally the use of correlation is to find out
the periodicities of the signals. Due to the white noise
or signal pick up problems, the ECG signals of very
low amp
litude is corrupted and looses its
periodicities. The author has tested this using
MATLAB. Random Gaussian noise signal is
generated and mixed with a periodic signal and using
correlation properties, periodicity has been extracted.
But the actual testing o
n ECG signal is yet to be
performed as ECG signal is not readily available.

6. REFERENCES

[1] GD Clifford, Azuaje F, and McSharry PE,
Advanced methods and tools for ECG analysis,
Artech Publishing House, London, 2006. pp
-
384

[2] Task force of European Soci
ety of Cardiology
and North American Society of Pacing
Electrophysiology. Heart rate variability: standards of
measurement, physiological interpretation and
clinical use. Circulation 1996;93: pp 1043
-
1065.

[3] Rangraj M. Rangayyan, Biomedical Signal Pres
a
nalysis, IEEE press, Wiley interscience, 2002
-
594p.

[4] John G. Proakis and Dimitris G. Manolkis,
“Digital Signal Processing, Principles, Algorithms
and Application”, Prentice Hall of India, 2004.

[5] AN Kalinichenko, MI Nilicheva, SV Khaseva,
OD Yurieva,
and OV Mamontov, Signal stationary
assessment for the Heart Rate Variability Spectral
Analysis, 2008; pp 965
-
968

[6] Natalia Trayanova, In the spotlight:
Cardiovascular Engineering, IEEE

[7] Lown B, Verrier RL. Neural and ventricular
fibrillation. N Engl J

Med, 1976;294:1165
-
1170

[8] Levy MN, Schwartz PJ, eds, Vagal Control of
heart: Experimental Basis and Clinical Implications,
Armonk, NY, Futura,1994.

[9] Sayers BM, Analysis of heart rate variability.
Ergonomics, 1973;16: pp17
-
32.

[10] Akselrod S, Gordon
D, Ubel FA, Shannon DC,
Cohen RJ., Power Spectrum analysis of heart rate
fluctuation: a quantitative probe of beat to beat
cardiovascular control. Science, 1981;213:220
-
222.

[11] Pomenraz M, Macaulay RJB, Caudill MA, Kurtz
I, Adam D, Gordon D, Kilborn KM,
Berger AC,
Shannon DC, Cohen RJ, Benson M, Assessment of
autonomic functions in humans by heart rate spectral
analysis. Am J Physiol, 1985;248:H151
-
H153.

n=
-


r
xy
(l)

=




x(n) y(n
-
l
)

n=
-



n=
-


r
x
x
(l)

=



x(n)
x
(n
-
l
)

n=
-



JOURNAL OF INFORMATION, KNOWLEDGE AND RESEARCH IN

BIOMEDICAL ENGINEERING

ISSN: 0975


6752
|
NOV
10

TO OCT
11

|

Volume 1, Issue
2


Page
14



[12] Luczak H, Lauring WJ. An analysis of heart rate
variability. Ergonomics. 1973;16:85
-
97.

[13]Scher
er P, Ohler JP, Hirche H, Hopp H
-
W.
Definitions of beat to beat parameter of heart rate
variability. PACE pacing Clin Electophysiol.
1993;16:939.

[14] Berger RD, Akselrod S, Gordon D, Cohen RJ.
An efficient algorithm for spectral analysis of heart
rate var
iability. IEEE Transaction Biomed Engg.,
1986;33;900
-
904.

[15] Malik M, Camm AJ. Components of heart rate
variability: what they really mean and what we really
measure. AM J Cardiol, 1993;72:821
-
822.