Transcript Seminar Slides - CSE, IIT Bombay

Neural Networks in
ECG classification
Under the guidance of
Prof. P. Bhattacharya
Nishant Chandra
Mrigen Negi
Meru A Patil
Layout
 History
of Neural networks in medical
 Need for accurate processing
 Applications of ANN in medical
 What is ECG?
 ANN in classification of Arrhythmias
and Ischemia
 Conclusion
History of Neural Networks in
Medical
 Pioneering
work of neural network
has started since 1943 by McCulloch
and Pitts.
 Pattern recognition problem was
introduced by Rosenblatt (1958)
Need for accurate processing




One of the major goals of observational
studies in medicine is to identify patterns in
complex data sets.
Correct classification of heart beats is
fundamental to ECG monitoring systems such
as an intensive care etc.
Computers are used to automate signal
processing.
ANNs can detect patterns and make
distinctions between different patterns that
may not be apparent to human analysis.
Applications of ANN in medical
It has been successfully applied to various
areas of medicine to solve non-linear
problems.
 Applications include prediction of diagnosis
such as:

–
–
–
–

Cancer
the onset of diabetes mellitus
survival prediction in AIDS
eating disorders etc
Applications in signal processing and
interpretation involve ECGs or
electrocardiograms
Motivation
 Cardiovascular
Diseases contribute
29.3% of total deaths in world.
 Online ECG monitoring in ICUs/CCUs.
 Acting Specialist in emergency cases.
 Each component (P,QRS,T waves)
has different frequencies.
 Each individual is different.
 Learning by experience.
What is Electrocardiogram
(ECG) ?
 ECG
is the graphic recording of electric
potentials generated by the heart.
 12 lead ECG
 3 bipolar limb leads – I, II, III
 3 unipolar augmented limb leads - AVF, AVR,
AVL
 6 unipolar chest leads – V1 to V6.
Anatomy of Heart and ECG signal
Normal ECG signal
Conducting System of Heart
Posterior
The 12 Views of the Heart
Anterior
Limb leads orientation with
respect to heart
Chest leads orientation with
respect to heart
12 Lead Normal ECG
6 Chest leads
6 Limb leads
RR
ECG and diseases
Some of the diseases diagnosed by
ECG are:






Myocardial Ischemia/Infarction.
Arrhythmias.
Hypertrophy and enlargement of heart.
Conduction Blocks.
Preexcitation Syndromes.
Other cardiac disorders.
Did you know !!
 In
heart Transplant Acute heart
rejection is more likely to happen
when the heart donor was female
regardless of recipient sex.
 Every
34 seconds, a person dies
from Heart Diseases in the United
States.
Myocardial Ischemia
 Due
to lack of adequate blood flow to
the myocardium.
 Ischemia is reversible.
 Changes in ECG:
 T wave peaking
 Symmetric T wave inversion
 ST segment elevation
Myocardial Ischemia cont..
Different ECG Signals
Normal Signal
ECG with T wave inversion
ST segment elevated signal
ECG Signal with peak T waves
Arrhythmias
 It
refers to any disturbance in the
rate, regularity, site of origin, or
conduction of cardiac electrical
impulse.
 Broadly two types:
 Tachycardia – Heart Rate beyond 100
bits/minute.
 Bradycardia – Heart Rate below 60
bits/minute.
Arrhythmias cont ..
Different ECG Signals
Normal ECG Signal
ECG signal of Bradycardia patient
ECG signal of Tachycardia patient
Sensitivity (SE) and Specificity (SP)
Helps us to explore the relationship
between a diagnostic test and the (true)
presence or absence of disease.
 A test which is very sensitive will rarely
miss people with the disease.
 A specific test will have few false positive
results - it will rarely misclassify people
without the disease as being diseased.
 Classification Rate:
CC = 100×(TP+TN)/(TN+TP+FN+FP)]

Sensitivity (SE) and Specificity (SP) Cont…
Approach
 Variable
attributes considered to
affect the training and generalization
of the ANNs were identified as
follows:
– Number of nodes in the hidden layer
– Feature Selection method employed
– Number of files in training set
– Size of input feature vector
– Number of epochs
Case Study
Feature Extraction:
 Fourier
Transform
 Principal component analysis (PCA)
– widely used in signal processing,
statistics, and neural computing.
– basic goal is to reduce the dimension of
the data.
 Linear
Prediction Coding (LPC)
Fourier Transform
 QRS
complex is extracted by
applying a window of some time
duration (say 250 ms).
 Each QRS complex is Fourier
transformed and then the power
spectrum is calculated.
 The components generated along
with the temporal vectors give the
feature vector.
QRS spectra of a normal beat
QRS spectra of a Arrhythmia beat
PCA
Step 1: Get some data
 Step 2: Subtract the mean
 Step 3: Calculate the covariance
matrix
 Step 4: Calculate the eigenvectors
and eigenvalues of the covariance
matrix
 Step 5: Choosing components and
forming a feature vector
 Step 6: Deriving the new data set

Linear Prediction Coding (LPC)
The basic idea of this technique is that
sampled QRS segment can be
approximated as a linear combination of
the past QRS samples.
 a is the i th linear prediction coefficient,
and p is the order of the predictor.
 LPC coefficients can be extracted using
various methods viz Burg’s Method.

Training the NN
Number of neurons in the input layer is
determined by the number of elements in
the input feature vector.
 The output layer is determined by the
number of classes desired.
 The number of neurons in the hidden layer
varies according to the specific recognition
task and is determined by the complexity
and amount of training data available.

Neural network classifier
architecture
Performance Analysis
 The
performance of the neural
classifiers is evaluated by computing
the percentages of:
– sensitivity (SE),
– specificity (SP) and
– correct classification (CC)
Results
Neural
Classifier
Input Layer
Hidden Layer
1
12
5
2
10
3
3
5
2
Results Cont.
Neural
Correct
Classifier classification
%
(Avrg.)
1
94.83
Sensitivity Specificity
%
%
86.63
94.42
2
91.34
81.33
91.92
3
88.25
76.17
88.95
Results Cont.
 How
does ANN based classification
compare with:
– Other ECG widely used interpretation
program?
 Neural
networks were 15.5% more sensitive
– Expert cardiologist
 10.5%
more sensitive than the cardiologist
Conclusion
 Performance
of the neural network
strategy has shown higher
performance than other classical
methods (Cox regression models) in
predicting clinical outcomes of the
risk of coronary artery disease.
References



[1] M. A. Chikh, F. Bereksi Reguig. Application of
artificial neural networks to identify the
premature ventricular contraction (PVC)
beats,2004
[2] Costas Papaloukasa, Dimitrios I. Fotiadisb,
Aristidis Likasb, Lampros K. Michalis. An ischemia
detection method based on artificial neural
networks,2002
[3] C.D. Nugent, J.A.C. Webb, N.D. Black, G.T.H.
Wright, M. McIntyre. An intelligent framework for
the classification of the 12-lead ECG, 1999.



Introduction to Neural Networks in Healthcare,
Open Clinic, 2002.
[4] M.S. Thaler, The Only EKG Book You’ll Ever
Need 3rd Edition, Lippincott Williams & Wilkins.
P.J Mehta, Understanding ECG, 5th Edition, The
National Book Depot.
Believe it or NOT !!

How much blood does your heart pump?
– An average heart pumps 2.4 ounces (70
milliliters) per heartbeat. An average heartbeat
is 72 beats per minute. Therefore an average
heart pumps 1.3 gallons (5 Liters) per minute.
In other words it pumps 1,900 gallons (7,200
Liters) per day, almost 700,000 gallons
(2,628,000 Liters) per year, or 48 million
gallons (184,086,000 liters) by the time
someone is 70 years old. That's not bad for a
10 ounce pump!

Men suffer heart attacks about 10 years
earlier in life than women do.