Application of swarm intelligence optimization in biomedical

Application of Swarm Intelligence Optimization in Biomedical

Asmaa Hamad Elsaied

Siminar,FCI,Cairo University (17-July-2016)

Pre-master seminar

للماجستير التسجيل محاضرة

Introduction

What is the role of CS in Bio-medical.

What is Electroencephalogram (EEG).

What is Swarm Algorithms.

What is machine learning algorithms.

2


Overview

Overview

Prediction epileptic seizure Problem.

Thesis Motivation. Proposed Model. Thesis Objectives. Literature Review.

3


Introduction What is the role of CS in Bio-medical.




4


Overview

What is the role of CS in Bio-medical

The use of computer in biology and clinical science has contributed to

improve life-quality and also to gather research results in shorter time.

Biomedical computing combines the diagnostic and investigative aspects of

biology and medical science with the power and problem-solving

capabilities of modern computing.

5


What is the role of CS in Biomedical Cont’d…

Biomedical computing develops computational methods that improve patient

lives and extend our knowledge of human medicine.

An accurate diagnosis and appropriate approach to treatment is crucial; it

improves patient outcome, avoids exposing patients to potentially harmful

treatment, and promotes efficient use of health-care resources.

6


What is the role of CS in Biomedical Cont’d…

A number of diagnostic tests such as Electroencephalogram

(EEG), Computed Tomography (CT), Magnetic Resonance

Imaging (MRI) and PET (Positron Emission Tomography) are

existed to diagnosis and to identify the disease.

7


Introduction What is the role of CS in Bio-medical. What is Electroencephalogram (EEG). What is machine learning algorithms. What is Swarm Algorithms.

8


Overview

What is Electroencephalogram (EEG). The EEG signal is usually used for the purpose of recording the electrical

activities of the brain signal that typically arises in the human brain.

The recording of the electrical activity is basically done by placing

electrodes on the scalp, which measures the voltage fluctuations in the

brain.

9


What is Electroencephalogram (EEG). The EEG signals are commonly decomposed into five EEG

sub-bands: delta, theta, alpha, beta and gamma.

10


What is Electroencephalogram (EEG) Cont’d…

The greatest advantage of EEG is speed. Complex patterns of

neural activity can be recorded occurring within fractions of a

second after a stimulus has been administered. EEG can

determine the relative strengths and positions of electrical

activity in different brain regions.

11






12


Overview

13

What is Meant by Swarm?


14

Swarm-based algorithms have recently

emerged as a family of nature-inspired

metaheuristic algorithms, population-based

algorithms that are capable of producing low

cost, fast, and robust solutions to several

complex problems.

What is Meant by Swarm?


15

(Biology) Swarm Intelligence (SI) can be defined as collective behaviour of a group of animals , social insects such as ants, bees, and termites, that are each following very basic rules.

(Computer Science) Swarm Intelligence (SI) can be defined as a relatively new branch of Artificial Intelligence that is used to problem solving using

algorithms based on the self-organized collective behaviour of social social swarms in nature.

What is SI Means?


16

Properties of SI System

Swarm intelligence system characterized by: It is composed of many agents. The interactions among the agents are based on simple behavioral. The agents are either all identical or belong to a few typologies. The overall behavior of the system results from the interactions of agents

with each other and with their environment.


The main advantages of the swarm intelligence approach compared

with a classical approach are the following: Scalability: SI systems are highly scalable the control mechanisms used

in SI systems are not too dependent on swarm size, as long as it is not too small.

Adaptability: the group can quickly adapt to a changing environment.

Robustness: even when one ore more individuals fails, the group can still perform its tasks.

17

SI Advantages


It covers chicken swarm optimization, particle swarm

optimization (PSO) algorithm, ant colony optimization

algorithm, bee colony optimization algorithm, bacterial

foraging optimization algorithm, cat swarm optimization

algorithm, harmony search algorithm, etc.

18

Examples of SI Algorithms






19


Overview


Machine learning is a subfield of computer science that

explores the study and construction of algorithms that

can learn from and make predictions on data. Such algorithms

operate by building a model from example inputs in order to

make data-driven predictions or decisions expressed as

outputs.

20


What is machine learning algorithms Cont’d….

Machine learning models like neural network (NN) and

support vector machine (SVM) have been successfully applied

to neuroimaging data to make predictions about behavioral and

cognitive states of interest.

21


While these multivariate methods have greatly advanced the field of neuroimaging, their application to electrophysiological data has been less common especially in the analysis of human intracranial electroencephalography (iEEG, also known as electrocorticography or ECoG) data, which contains a rich spectrum of signals recorded from a relatively high number of recording sites.

22


What is machine learning algorithms Cont’d….

Prediction epileptic seizure Problem

Proposed Model.

Thesis Motivation

Thesis Objectives.

Literature Review.

23


Overview


Epilepsy is a critical neurological disease stemming from temporary abnormal discharges of the brain electrical activity, leading to uncontrollable movements and trembling

Epilepsy is the second most common neurological condition seen in primary practice worldwide with an approximate prevalence of 5.8 per 1000 population in the developed world and between 10.3 per 1000 to 15.4 per 1000 in developing countries . Despite its prevalence, epilepsy can be very challenging to diagnose and treat.

24


Prediction epileptic seizure Problem Cont’d….

Clinically to predict and diagnose epileptic seizures, the brain

activities are to be monitored through EEG signals which

contain the markers of epilepsy. EEG signals of epileptic

patients exhibit two states of abnormal activities namely

interictal or seizure free (in-between epileptic seizures) and

ictal (in the course of an epileptic seizure) .

25


Generally, a clinician relies on identifying interictal (seizures

free) EEG signals for epilepsy prediction as the ictal segments

are obtained rarely. Thus, longer durations of EEG signals are

necessary to visually monitor and analyze in order to localize

the normal, interictal and ictal episodes for a patient.

26


Prediction epileptic seizure Problem Cont’d….

Overview


Thesis Motivation.

Proposed Model.

Thesis Objectives.

Literature Review.

27


Thesis Motivation

In the majority of cases, seizures occur unexpectedly, without a sign of warning to alert and prepare the person for an onset of seizure. Such abrupt and uncontrollable nature of the disease can cause physical injury. In addition to bodily harm, there is a feeling of helplessness associated with a lack of control over seizure and inability to anticipate and know when a seizure may strike. In order to adopt a seizure prediction algorithm in clinical practice, it must pass rigorous statistical validation using real EEG data. A system that can reliably predict a prospective seizure can have a significant impact on the patient's life. The study, characterization and implementation of such a model are the subject of this thesis.

28


Overview


Thesis Motivation

Proposed Model.

Thesis Objectives.

Literature Review.

29


Proposed Model.

Epilepsy can be detected by traditional methods by well-trained and experienced neurophysiologists by visual inspection of long durations of EEG signals. This is time – consuming, tedious and subjective. Hence, in order to overcome these limitations, a computer – aided detection of epileptic EEG signals can be used. And also there is a need to generate an efficient prediction model for making a correct diagnosis of epilepsy

30


Proposed Model Cont’d….

The following figure discuss the general framework for EEG

signal analysis, especially to identify the epileptic seizure

31


Proposed Model Cont’d….32

Feature Engineering Using Swarm Intelligence





Raw Data(no Feature

Engineering)


Proposed Model Cont’d….

This model consists of the following processes: EEG signal pre-processing: this is used to remove the noises from the signal Feature extraction: this is used to extract the EEG signal features from

decomposed signal. Feature selection: In this process the relevant features are selected from the

extracted features. Classification: In this process, the selected features are given as inputs to the

classification process. The classification method is mainly used to analyse the EEG signal and it classifies the signal into normal or abnormal.

35


Overview


Thesis Motivation.

Proposed Model.

Thesis Objectives.

Literature Review.

36


Thesis Objectives

It is clear that detecting and controlling a seizure is not enough to make patients completely free of seizures. The objectives of this thesis are: The seizure needs to be predicted well in time so that actions can be

taken to avoid the upcoming seizure. We aim to construct a patient-specific predictors for interictal EEG

signals, i.e., aimed to find both the appropriate input set and also the appropriate classifier parameters that result in an improved prediction at low computational cost using swarm optimization technique.

37


Overview


Thesis Motivation.

Proposed Model.

Thesis Objectives.

Literature Review.

38


Author developed an automated system for the classification of brain abnormalities.

In this work the EEG signals are given as input to the pre processing. From the pre processing the discrete wavelet transform are used to remove noises and the EEG signal are decomposed into five sub-band signals. The non linear parameters (time and frequency) were extracted from each of the six EEG signals (original EEG, delta, theta, alpha, beta and gamma). A genetic algorithm was used to extract the best features from the extracted time and frequency domain features. Then the k-means classifier is used to classify the given EEG signal as normal or abnormal.

39

Literature Review


Kalaivani, M., V. Kalaivani, and V. Anusuya Devi. "Analysis of EEG Signal for the Detection of Brain Abnormalities." IJCA Proceedings on International Conference on Simulations in Computing Nexus. No. 2. Foundation of Computer Science (FCS), 2014.

Author presented a supervised machine learning approach that classifies

seizure and nonseizure records using an open dataset containing 342 records. the results show an improvement on existing studies by as much as 10% in most cases with a sensitivity of 93%, specificity of 94%, and area under the curve of 98% with a 6% global error using a k-class nearest neighbor classifier.

40

Literature Review


Fergus, Paul, et al. "Automatic epileptic seizure detection using scalp EEG and advanced artificial intelligence techniques." BioMed research international 2015 (2015).

Author proposed a method using subband nonlinear parameters and genetic algorithm for

automatic seizure detection in EEG. In the experiment, the discrete wavelet transform was

used to decompose EEG into five subband components. Nonlinear parameters were extracted

and employed as the features to train the support vector machine with linear kernel function

(SVML) and radial basis function kernel function (SVMRBF) classifiers. A genetic algorithm

(GA) was used for selecting the effective feature subset

41

Literature Review


Hsu, Kai-Cheng, and Sung-Nien Yu. "Detection of seizures in EEG using subband nonlinear parameters and genetic

algorithm." Computers in Biology and Medicine 40.10 (2010): 823-830

Author Used two-features to improve the performance of EEG signals.

Neural Network based techniques are applied to feature extraction of EEG signal. Extracting features based on Average method and Max & Min method of the data set. The Extracted Features are classified using Neural Network Temporal Pattern Recognition Technique. The two methods are compared and performance is analyzed based on the results obtained from the Neural Network classifier.

42

Literature Review


Nandish.M, Stafford Michahial, Hemanth Kumar P, Faizan Ahmed, “Feature Extraction and Classification of EEG Signal Using Neural Network Based Techniques”, International Journal of Engineering and Innovative Technology (IJEIT) Volume 2, Issue 4, October 2012.

Author developed an automated system for epileptic seizure

prediction from intracranial EEG signals based on Hilbert-Huang transform (HHT) and Bayesian classifiers.

43

Literature Review


Nilufer Ozdemir and Esen Yildirim , “Patient Specific Seizure Prediction System Using Hilbert Spectrum and Bayesian Networks Classifiers”, Computational and Mathematical Methods in Medicine, 2014.

The proposed model will imply the following steps:

1. Survey and identify major problems associated with Bio-medical (e.g., EEG and

Epileptic Seizures).

2. Survey some techniques for Epileptic Seizures Prediction.

3. Build a prediction model for Epileptic Seizures using Swarm Optimization Technique.

4. Test the developed model for Epileptic Seizures Prediction.

5. Conduct a performance analysis of the developed model with the existing ones.

6. Release a recommendation for a future work for Epileptic Seizures Prediction

44

Work Plans


45

Any Questions!?


Thanks and Acknowledgement46


Education

Application of swarm intelligence optimization in biomedical