AUTOMATIC EPILEPTIC SEIZURE

DETECTION WITH MACHINE

LEARNING

Submitted by
Suma Gunda
Roll number:164264

Keertana Upparapalli
Roll number:164167
NIT-WARANGAL

1
 INDEX

Topic Pg.No.

Title 01

Index 02

Abstract 03

Introduction 03

Wavelet decomposition 05

Feature Extraction 06

Machine learning classification algorithm 07

Experimental Results 08

Conclusion 09

References 09

2
Abstract:

Traditional method of epileptic seizure detection could not avoid the

process of manually selecting the features.This project deals with a
method of the seizure detection based on EEG signal using multilevel
wavelet decomposition and classification using KNN machine learning
algorithm.And the work verifies the feasibility of this method through the
actual research data and parameter setting.Then, the comparison with
the classical method using the support vector machine(SVM) has been
done, which shows that the approach presented in this project is better.
And the experimental result on single channel is that the average
accuracy is 88%.

Introduction:

Epilepsy can be potentially life threatening with brain failure, heart and
lung failure, head trauma due to accidents and sudden unexpected
death. Even understated epileptic can cause insignificant harm in the
brain. Long-term problems such as fall in intelligence quotient (IQ),
depression, suicide, Social problems may lead to reduce the quality of
life. So the diagnosis of epilepsy is the most important in the existing
development of research. The main challenge of this work is to detect
the epilepsy seizure in order to maintain independency in patient’s life
and also to help person with epilepsy to lead full and productive life.
Epilepsy is one of the common neurological disorder that affects the
nervous system and is characterised by the transient and unexpected

3
electrical disturbance of the brain. All brain functions including feeling,
seeing, thinking and moving muscles depend on electrical signals
passed between nerve cells in the brain. Epilepsy is also known as a
seizure disorder. A seizure occurs when too many nerve cells in the
brain “fire” too quickly causing an “electrical storm”.
According to the WHO , there are about 50 million people worldwide who
are suffering from epilepsy. In about 70% of people with epilepsy, the
cause is not known and in 30% of the people most common causes are
due to head trauma, infection of brain tissue, brain tumour and stroke,
heredity and prenatal disturbance of brain development. The
electroencephalogram (EEG) is the most influential technique in the
detection of epileptic seizures. EEG records the electric activity of the
cerebral cortex in the brain and it is used to measure or diagnose brain
diseases. EEG monitoring is an important factor when identifying
conditions, particularly of patients with epilepsy. The recording data of
the human EEGs are carried out by placing the electrodes on the scalp,
voltage range of the scalp EEG lie between 10 and 100 lV. The main
concern in EEG analysis is how to differentiate seizure from normal EEG
signal. Conventionally, seizure activities are visually inspected from EEG
signals by trained physicians. But this is a time-consuming and
cumbersome in case of long EEG recordings. So, a rapid detection and
classification of seizure activity would give a great support to quantitative
analysis and interpretation. In modern years, different methods have
been developed to automatically classify the normal or abnormal EEG
signals without spending hours for visual inspection.
The process of automatic classification is achieved using pattern
recognition. The various stages in pattern recognition are feature
extraction and classification. Feature extraction is a technique for
extracting the hidden characteristics of the signal such as amplitude,
frequency, mean value, etc., to interpret the signal. These features are
fed as input to the classifier to differentiate normal and epileptic
seizures.. However, the classification performance is mostly dependent
on the features that are being used to characterise the original EEG.
Therefore, feature extraction is one of the most significant constituents of
pattern recognition. It suggestively contributes to the performance of the

4
classifier and reduces data size without losing its characteristic power.
Thus, it is important to propose an effective feature extraction method
instead of designing the structure of a complex classifier.

Wavelet Decomposition:

The central idea of the wavelet transform analysis is to decompose a

signal into different levels of resolution (for this research six levels will be
used), this process is known as multiresolution . The multiresolution
representation provides a simple hierarchical framework for the
interpretation of the information in the time series. At different
resolutions, the details of a signal usually characterize different physical
structures of themselves.
One area in which the DWT has been particularly successful is the
epileptic seizure detection because it captures transient features and
localises them in both time and frequency content accurately. DWT
analyses the signal at different frequency bands, with different
resolutions by decomposing the signal into a coarse approximation and
detail information. DWT employs two sets of functions called scaling
functions and wavelet functions, which are related to low-pass and
high-pass filters, respectively. The decomposition of the signal into the
different frequency bands is merely obtained by consecutive high-pass
and low-pass filtering of the time domain signal. Each stage of this
scheme consists of two digital filters and two down-samplers by 2. The
first filter, h[.] is the discrete mother wavelet, highpass in nature, and the
second, g[.] is its mirror version, low-pass in nature. The down-sampled
outputs of first high-pass and low-pass filters provide the detail, D1 and
the approximation, A1, respectively. The first approximation, A1 is further
decomposed and this process is continued.Selection of suitable wavelet
and the number of decomposition levels is very important in analysis of
signals using the DWT. The number of decomposition levels is chosen
based on the dominant frequency components of the signal. The levels
are chosen such that those parts of the signal that correlates well with
5
the frequencies necessary for classification of the signal are retained in
the wavelet coefficients.In this present study,the number of
decomposition levels were chosen to be 6 because the EEG signals
were sampled at the rate of 256Hz and the mother wavelet,db4.

Feature Extraction:
The extracted wavelet coefficients provide a compact representation that
shows the energy distribution of the EEG signal in time and frequency.
The following statistical features were used to represent the time
frequency distribution of the EEG signals:
1) Maximum of the wavelet coefficients in each sub-band.
2) Minimum of the wavelet coefficients in each sub-band.
3) Standard deviation of the wavelet coefficients in each sub-band.
4)Energy of the entire signal.
The data was acquired from chbmit database.The DWT was performed
at 6 levels, and resulted in six sub-bands: d1-d5 and a6 (detail and
approximation coefficients respectively).For each of these sub-bands,
we extracted four measures of dispersion, yielding a total of 19 attributes
per sample window.

6
Machine learning classification algorithm:

Algorithm

A case is classified by a majority vote of its neighbors, with the

case being assigned to the class most common amongst its K
nearest neighbors measured by a distance function. If K = 1, then
the case is simply assigned to the class of its nearest neighbor.

It should also be noted that all three distance measures are only
valid for continuous variables. In the instance of categorical
variables the Hamming distance must be used. It also brings up
the issue of standardization of the numerical variables between 0
and 1 when there is a mixture of numerical and categorical
variables in the dataset.Since our classifiers use supervised
learning, we must also provide the outputs, which was simply a
class label.

7
Experimental Results:
The proposed diagnostic system consists of a preprocessing
/feature selection and one classifier subsystem. Duabechies
Wavelets order-4 with 6 levels have been used for pre-processing
in order to achieve the same dimensionality reduction of wavelet
​ he 19 inputs correspond to the four features times
coefficients.​ T
the number of wavelet decomposition (D1-D5 & A6). The results
using K-NN were encouraging. The classification accuracy for this
two-class problem found to be 88%.
From the result of experiments, the average of TPR is 91% that is
to say, nearly most of the EEG signal on seizure status can be
classified into seizure status.
The method using the support vector machine(SVM) is frequently
used for the traditional epileptic seizure recognition because SVM
could map the feature vector into the space with the higher
dimensions using the kernel function for the classification. To
compare with the K-NN algorithm mentioned in this project, the
method combined SVM with the manually selected features is
validated on the same dataset and the accuracy was found to be
76.1%.

8
CONCLUSION :
The result of experiments shows that the algorithm of single
channel proposed in the project has a high accuracy(88%), a good
TPR(91%). Meanwhile, the parameters that might influence the
performance of the algorithm are discussed, and the proper
parameters are chosen to match the model. Also, the comparison
with the method using the SVM combined with the manually
selected features shows that, the method mentioned using the
K-NN has better performance.
​ HE COMPARISON OF THE RESULT BETWEEN THE METHOD USING
T
THE SVM AND K-NN ALGORITHM.

Method Accuracy TPR

SVM 76.1% 87%

K-NN 88% 91%

​REFERENCES :

1.Epileptic seizure prediction using deep learning. Maarten

Larmuseau

2.Epileptic Seizure Auto-detection Using Deep Learning

Method.Yuzhen Cao, Yixiang Guo, Hui Yu School of Precision
Instrument and Opto-electronics Engineering Tianjin University
Tianjin, China .

9
10
11