A method for detection and reduction of stress using EEG

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 01 | Jan -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1598
A Method for Detection and Reduction of Stress using EEG
Ms. Neeta Baliram Patil1, Mr. Rahul P. Mirajkar2, Ms. Suchita Patil3, Ms. Pratibha Patil4
1Student, Department of Computer Science and Engineering, Bharati Vidypeeth’s College of Engineering, Kolhapur,
Maharashtra, India
2Assistant Professor, Department of Computer Science and Engineering, Bharati Vidypeeth’s College of
Engineering, Kolhapur, Maharashtra, India
Maharashtra, India
Maharashtra, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - World health organization says that,
nowadays people mental health problems and also physical
problems are because of stress. This stress is major problem
of our times. There are many stress detection methods are
available. To study the human behavior there are many
techniques are available which monitors the human brain.
However, there is less research available on reduction of
stress methods in terms of technology but studies are
available on stress detection methods. This research
proposes a method that uses the EEG signals for detecting
the stress in humans and introduces the stress reduction
techniques by adding interventions into the method. K-
means clustering method used in the research to measure
the stress which help in dividing the subjects into different
classes and detect the stress level. Product for human stress
reduction can be developed by using this method. The
success of implementation and development of this research
expected to help in reducing time consumed and human
power in determining best recommendation and solution for
stress management.
Key Words: Stress, Electroencephalography, Stress
measurement, Monitoring.
1. INTRODUCTION
Stress is a body’s method for reacting to a challenge.
Human stress can have an impact on a person’s mental
and physical well - being. Stress can lead to a change in
behavior and in physiology. Many people suffer from
stress in everyday life. Stress is related to human work in
one way or other. Originates of stress have different
sources such as time pressure while working in company,
responsibility, economic problem or physical factors such
as noise. Signs of stress are human fell tension, anxious,
angry, frustrated or irritated by things over which he has
no control.
Stress detection is an on-going research topic among
both psychologists and engineers. Wearable sensors and
bio signal processing technologies are developed for
detecting the human stress. There are various bio signal
processing technologies use for human stress detection
such as Electroencephalography (EEG),
Electrocardiography (ECG), Electromyography (EMG),
Blood Pressure (BP), Blood Volume Pulses (BVP), Galvanic
Skin Resistance (GSR), Respiration and Skin Temperature
(ST) etc. Also to measure the stress level human
physiological features are used for that human
physiological signal processing technology is used. There
is difference between Individuals physiological features.
Person physiological features are changes when he/she
response to stressful events. By considering various
physiological features occur in human while he/she is in
stress the estimation of stress are done by using cluster
based analysis method [1].
In proposed method EEG stress detection technology
is used. EEG is one of the most reliable sources to record
electrical activity of the brain along the scalp. To measure
the voltage fluctuation resulting from ionic current within
the neurons of the brain EEG is used. To diagnose the
epilepsy, coma, sleep disorders, encephalopathy and brain
death EEG is most often used. EEG is used as first line
method to identify stroke, focal brain disorder, and
tumors. Hardware cost of EEG is significantly lower than
other techniques. Near the patient’s bed patient bio signals
can be recorded by using EEG also EEG can monitor long
term sleep stages or epilepsy; this is the reason why EEG is
most preferable. Compare to other technique, EEG is most
superior choice because it is convenient tool for
physiological research. When subject perform some
behavioral or it is out of laboratory EEG can be used as
wearable sensor. EEG can track brain changes during
different phases of life without disturbing a patient for e.g.
EEG sleep analysis.
2. LITERATURE SURVEY
A literature review conducted over analysis of stress
using physiological signals and evaluation of stress level.

2.1 Related Research
Previous work on stress measurement has been focused
on the collection and analysis of physiological data and the
identification of the correlation between perceived stress
and multiple physiological features.
Qianli Xu, Tin Lay Nwe, and Cuntai Guan (2015) [1],
Proposes a novel Cluster-Based Analysis method to
measure stress using physiological signals, which accounts
for intersubject differences. This research uses the
clustering process that assigns the subject into subgroups,
so as to exploit the inherent homogeneity of subject’s
stress response within the cluster. Thus the intersubject
differences are automatically accommodated, and the
overall accuracy of stress evaluation is improved.
Chee-Keong Alfred Lim and Wai Chong Chia (2015) [2],
Focuses on evaluating to what extent a single electrode
EEG headset – NeuroSky MindWave is able to classify
brainwave in terms of subject’s stressor level. In this study
they use the MATLAB environment for processing the EEG
signals. By reducing the number of electrodes needed, it
also means cheaper EEG headset can be used to diagnose
various mental disorders.
Tong Chen, Peter Yuen, Mark Richardson, Guangyuan
Liu, and Zhishun She (2014) [3], Present a method to
detect psychological stress in a non-contact manner using
a human physiological response. In this paper they use a
Hyper Spectral Imaging (HSI) camera to obtain tissue
oxygen saturation (StO2) data as a feature for detecting
human stress.
Cornelia Setz, Bert Arnrich, Johannes Schumm,
Roberto La Marca, Gerhard Tr¨oster, and Ulrike Ehlert
(2010) [4], Developed a personal health system for
detecting the stress. They use the discriminative power of
electro dermal activity (EDA). This EDA power is use for
distinguishing stress from office work load. They do the
analysis on EDA data and evaluate the information about
stress level of a person.
Awanis Romli, Arnidcha Peri Cha (2009) [5], Suggested
a system which deals with stress management. They
change the manual system into computerized one. This
system does the stress management for particular
individual based on their activity interest and provides a
solution for stress management. This system using
combination of rule based technique and Holland’s Self
Directed Search Model to determine the best solution for
managing the stress. This system use Rule-Based
technique. In this technique user is given a test to develop
the system’s knowledge-based. From this test system can
determine the user’s interest, behavior and through their
thinking, the system gives the best solution to manage
user’s stress according to their interest.
Jennifer A. Healey and Rosalind W. Picard (2005) [6],
Shows how physiological data is useful during real-world
driving task to determine a driver’s relative stress level. In
this study they show how physiological sensor can be used
to obtain electrical signals that can be processed
automatically by an on-board computer to give dynamic
indications of a driver’s natural state under natural driving
conditions. The experiment was design to monitor the
driver’s physiological reaction during real-world driving
situation under normal condition.
2.2 Existing System
Existing system proposes a novel cluster based
analysis method to measure perceived stress using
physiological signals, which accounts for the intersubject
differences [1].
Fig -1: Architecture of Existing System
The system contains three stages Data Gathering, Data
Preprocessing and Data Classification. Data is collected by
using wireless sensor. Wireless sensor was used to collect
a subject’s physiological signals. Data preprocessing
discover the useful information. This is responsible for
collecting raw data and converting it into information
useful for decision making by users. It does the feature
extraction on data means redundancy in the dataset is
reduced by this process. In Data classification mode
feature extracted data are used to classified subjects into
different classes.
2.3 Limitations of the Existing System
There are various researches are going on human
stress detection and estimation of stress. For detecting
human stress various methods is use such as EEG
(Electroencephalography), EMG (Electromyography), GSR
(Galvanic Skin Response), BVP (Blood Volume Pulses), BP
(Blood Pressure), ST (Skin Temperature) etc. Some
researchers estimate the stress by using various
classification methods but not implement the reduction
technique yet. There are studies which also done on stress
management. They change the manual system into
computerized one. This system does the stress

management for particular individual based on their
activity interest and provides a solution for stress
management. From this test system can determine the
user’s interest, behavior and through their thinking, the
system gives the best solution to manage user’s stress
according to their interest. But these studies are also time
consuming to manage the stress.
From the survey it has been seen that, none of the
proposal provide a method for human stress reduction
using technology. Till now work is done on detecting the
stress and evaluating it [2], [3], [4], [7], [10], [9], and [11].
Also some works are done on identifying the stress level
[1]. But only few studies have been done on stress
management [5]. However, it makes the process complex
and time consuming. Therefore to control the health
related issues generate from stress, stress reduction
technique is important in relation to technology. It would
be beneficial to propose a method for developing products
for human stress reduction.
3. PROPOSED SYSTEM
Proposed system design and develop a method that
analyze the stress level effectively and also tackle the
intersubject difference of human stress and provide
techniques for reducing the stress among individuals for
improving their performance in work. Fig -2 shows the
architecture of proposed system.
Fig -2: Architecture of Proposed System
In the proposed approach, the goal is to reduce
the human stress after detecting the stress by using the
EEG signals. The objective of this research is to accurately
estimate the human stress and diagnose the human stress
level. The estimation of stress can be done by analysing
the EEG features and the human stress level i.e. stress or
relaxed mode is shown by using clustering process [1].
The k-means clustering will be used for dividing the
subjects into subgroups for predicting the human stress
level [1]. The aim is to reduce stress by introducing the
interventions into the system if stress level of human is
high and do the statistical analysis for checking that stress
level is reduced or not. In the proposed approach, a
method will be implemented to reduce the human stress,
so that they can efficiently improve his/her performance
in the work.
Stress reduction is an objective of proposed approach.
To achieve this objective steps used in this study are
describe below.
1. Get the test data for checking the stress.
2. Do the data preprocessing on collected data.
3. Classifying data using clustered based analysis
method.
4. If stress is high then introducing the intervention
into the system.
5. Perform statistical analysis on data for getting the
results.
4. PROPOSED WORK
4.1 Data Collection
Data Collection module contains the data related to
patients. This patient data contains Name of patient,
doctor’s name who gives the treatment to that particular
patient, age of patient, gender, admit date of patient etc.
This information will be stored into the database for
further processing. Also data collection module stores the
patient’s EEG data which will be useful for detecting the
stress and analyzing it. The EEG data will be collected from
lab study by involving subjects. Eight channel EEG is used
here for collecting the brain signals. Once the signals are
collected by using EEG, then stored that signals into
dataset.
EEG can track brain changes during different phases
of life without disturbing a patient [13]. Eight channel EEG
is adopted in our study for collecting the brain signals. EEG
is used to capture the brain signal in order to detect the
stress and analyse it. Through EEG monitoring,
individual’s stress level can be detected and quantified in
an efficient manner. Once the signals are collected by
using EEG, then stored that signals into dataset.
4.2 Data Analysis
Analysis of data is a process of inspecting, cleaning,
transforming, and modeling data with the goal of
discovering useful information, suggesting conclusions,
and supporting decision making. Analysis refers to
breaking a whole into its separate components for
individual examination. Data analysis is a process for
obtaining raw data and converting it into information
useful for decision making by users. Data is collected and
analyzed to answer questions. This module will be
responsible for analyze the data for discovering useful
information. For analysis phase, there are two sub phases
a) Stress Indices and b) Feature Extraction.

EEG is an electronic record of the oscillations in the
human brain, recorded from multiple electrodes attached
to the scalp. Depending on the individual’s state of
relaxation, EEG can vary in shapes [14]. In healthy adults,
the amplitudes and frequencies of such signals change
from one state of a human to another, such as wakefulness
and sleep. The characteristics of the waves also change
with age [15]. Therefore, In Stress Indices phase, subject’s
threshold value will be calculated. This value will be used
as subject’s normalized stress indices value. This stress
indices value computed as using formula
Where is, normalized stress indices of subject i.
Is, Average value of subject i.
is smallest value of the subject.
is largest value of the subject.
Human’s EEG signal is a kind of weak electrical signal.
Without processing, we will get EEG signals with a variety
of noise interference, especially ocular artifacts [15]. Eye-
blinks and movement of the eyeballs produce electrical
signals that are collectively known as Ocular Artifacts (OA)
[16]. Signals recorded from electrodes will be mixed with
noises such as: blinks, eye movement artifact etc.
Therefore, in order to meet the requirements of data
analysis and processing, we should do noise reduction
processing to original EEG signal at first [8], [15]. In order
to clean EEG signal data, algorithm based on discrete
wavelet transformation (DWT) is designed to identify and
remove ocular artifacts from EEG signal [14]. Denoising
EEG using this algorithm yields better results, in terms of
ocular artifact reduction [17], [18]. In Feature Extraction
phase, data preprocessing will be done on EEG dataset.
This process will be useful for avoiding over fitting and
improving the model performance. This process provide
us noise-filtering feature.
4.3 Data Classification
Classification of data is the process of sorting and
categorizing data into various types, forms or any other
distinct class. Data classification enables the separation
and classification of data according to data set
requirements for various objectives. The Data
Classification Module will be performed to show whether
the subject is in stress or relaxed mode. Data classification
is diverse process that involves various methods and
criteria for storing data. In this module clustering process
is use for doing the data classification. The features
extracted data are used for classified subjects into
different classes. Clustering process is used to divide
subjects into a set of subgroups. K-means clustering
process is used for dividing the subjects into subgroups
[12].
4.4 Statistical Analysis
Statistics is a term used to summarize a process that
an analyst uses to characterize a data set. Statistical
analysis involves the process of gathering and evaluating
data and then summarizing the data. Statistics studies
methodologies to gather, review, analyze and draw
conclusions from data. Statistical analysis is a component
of data analytics. Statistical analysis involves collecting
and scrutinizing every data sample in a set of items from
which samples can be drawn.
After doing the cluster based analysis the perceived
stress effectively reflects the changes in subject’s stress
level. By getting the subject’s stress, statistical analysis is
done on individual subject’s stress level. This module
implemented as first by introducing interventions into the
module for getting low stress. After that taking new EEG
data of patient and then calculating subject’s new stress
indices value. New stress indices value will compare with
old stress indices value for showing that now stress of
patient is low or high.
5. SNAPSHOTS
Fig -3: Add Patient Information
Fig 3. shows the form which is useful to add new
patient information into the database. In this form user
have to fill Name of particular patient, patient’ Doctor
name, Patient’s Address, Patient Admission Date, it’s
Gender and Age and also have to upload patient’s collected
Physical data and Cognitive data. Submit button saves
patient information into the database. From this form user
can save individual patient information into the database
for future use.

Fig -4: Add Patient New Data
Fig 4 shows add patient new data window. When user
clicks on Add Data button which then this window opens
for adding new patient data after going through
interventions phase. This window shows patient ID and
Name and contains browse button, Submit button and
Cancel button. When user clicks on browse button it
shows the window. From that window user select patient
new data for further processing and click on submit
button which saves data into system and user can now use
that data for further processing.
6. RESULT ANALYSIS
6.1. Stress Level
The first process within the system is to calculate the
stress indices value of subject. After collection of data first
process is to calculate stress indices value of cognitive
data and physical data. This subject’s stress indices value
is used as threshold value of that subject. EEG data of
subject can vary depending on subject’s behaviour.
Therefore, its need to calculate individuals threshold value
which used for calculating individuals stress level.
We calculate individuals stress indices value for
establishing stress level. We analyse collected data and
calculate this value. By this analysis we get following
graph. This graph indicates individual subject’s stress
indices value.
Chart -1: Stress Indices in two stages
This graph reflects the change into the stress indices
value. This graph showed a stress indices value before task
and after the task which are cognitive SI and physical SI.
Stress indices value after the task was higher than before
task load.
6.2. Data Classification
To improve the overall performance clustering
procedure were used. Subjects EEG data were pre-
processed using DWT with the aim to reduced noise from
data. Based on subjects stress classification was done
using K-means clustering. We analyze the reduced data for
establishing stress level of subject. By this analysis we get
following graph which shows stress level of subjects.
Chart -2: Data Classification outcome
This graph showed two clusters of subject’s high
stress subject’s cluster and other one is low stress subjects
cluster. From this classification we can easily differentiate
subjects between high stress and low stress.

6.3. Statistical Analysis
After data classification system is responsible to
calculate new stress indices value of high stress subject.
This stress indices value was used for comparing with
previous stress indices value.
Chart -3: Stress Indices in three stages
This graph reflects the change into the stress indices
value. This graph showed a stress indices value into three
stages before task load, after the task load and after the
recovery. This graph showed comparison between three
stages and gives the results.
7. CONCLUSIONS
We have made survey of different literatures and
studied different techniques for establishing stress. For
analyzing the human stress level there are different
techniques are used. In the various researches, it uses data
pre-processing technique and proper classification
technique to identify stress level in human. But there are
fewer researches which provide the stress reduction
technique.
However, this system consist stress reduction
techniques by introducing interventions into the system.
In this work first we calculate the stress indices value
which is used as threshold value for comparison. This
work improves the accuracy of data by analyzing data
DWT technique which reduces the noise from data. Also
data classification technique is used for identifying the
stress level of subject. Clustering is used for classifying the
subjects. After classification we get the stress level of each
subject. After that interventions are added into the system
and new data is collected for processing and we get the
results that stress is reduced or not.
Form our experiments we found that our technique is
better than other techniques. In this system we identify
stress level of human and also this system provides stress
reduction technique. This system provides easy process of
data and it is not time consuming.
Thus, the proposed system is helpful for analyze the
stress level effectively and provide techniques for
reducing stress among individuals for improving their
performance.
8. FUTURE WORK
As future work, first we can implement this same
system by using other stress detection technologies like,
ECG, EMG, and GSR etc. For further research we can go
through additional reference paper which can helpful to
look on advanced research like. Advance research on
adding detail into the data which can be helpful for
obtaining the results. It is needed that design the database
more detail so that detail analysis on data is easier and it is
understandable to the developer to find the effective
solution for stress reduction. Second we can increase the
classification accuracy by adding other classification
techniques. Also other feature extraction technique can be
used instead of DWT like DCT. Third we can improve the
accuracy of the system by adding more subject data and
also can check the accuracy of data using alternative
detection techniques such as blood pressure or heart rate.
ACKNOWLEGMENT
There have been many contributors for this to take shape
and authors are thankful to each of them.
REFERENCES
[1] Qianli Xu, Member, IEEE, Tin Lay Nwe, Member, IEEE,
and Cuntai Guan, Senior Member, IEEE, “Cluster-Based
Analysis for Personalized Stress Evaluation Using
Physiological Signals”, IEEE Journal of Biomedical And
Health Informatics, Vol. 19, No. 1, January 2015.
[2] Chee-Keong Alfred Lim and Wai Chong Chia, “Analysis
of Single-Electrode EEG Rhythms Using MATLAB to Elicit
Correlation with Cognitive Stress”, International Journal of
Computer Theory and Engineering, Vol. 7, No. 2, April
2015.
[3] Tong Chen, Peter Yuen, Mark Richardson, Guangyuan
Liu, and Zhishun She, Senior Member, IEEE, “Detection of
Psychological Stress using a Hyper spectral Imaging
Technique”, IEEE Transactions on Affective Computing,
Vol. 5, No. 4, October-December 2014.
[4] Cornelia Setz, Bert Arnrich, Johannes Schumm, Roberto
La Marca, Gerhard Tr¨oster, Member, IEEE, and Ulrike
Ehlert, “Discriminating Stress from Cognitive Load Using a
Wearable EDA Device”, IEEE Transactions On Information
Technology In Biomedicine, Vol. 14, No. 2, March 2010

[5] Awanis Romli, Arnidcha Peri Cha, “An Expert System
for Stress Management”, November 2009
[6] Jennifer A. Healey and Rosalind W. Picard, “Detecting
Stress during Real-World Driving Tasks using
Physiological Sensors”, IEEE Trans. Intell. Transp. Syst.,
vol. 6, no. 2, pp. 156–166, Jun. 2005.
[7] Prof. Shamla Mantri1,Vipul Patil2, Rachana Mitkar3,
“EEG Based Emotional Distress Analysis – A Survey”,
International Journal of Engineering Research and
Development, Volume 4, Issue 6 (October 2012).
[8] Yvan Saeys, Inaki Inza and Pedro Larranaga, “A review
of feature selection techniques in bioinformatics”,
Bioinformatics, vol. 23, no. 19, pp. 2507–2517, 2007.
[9] Mariya Khan, Zoha Rizvi, Muhammad Zakir Shaikh,
Warda Kazmi, and Anum Shaikh, “Design and
Implementation of Intelligent Human Stress Monitoring
System”, International Journal of Innovation and Scientific
Research, ISSN 2351-8014 Vol. 10 No. 1 Oct. 2014, pp.
179-190.
[10] Burcu Cinaz, Bert Arnrich, Roberto La Marca, Gerhard
Troster, “Monitoring of mental workload levels during an
everyday life office-work scenario”, Pers. Ubiquit. Comput.,
vol. 17, no. 2, pp. 229–239, 2013.
[11] Liza Varvogli, Christina Darviri, “Stress Management
Techniques: evidence-based procedures that reduce stress
and promote health”, Health Science Journal, Volume 5,
Issue 2 (2011).
[12] Tapas Kanungo, Senior Member, IEEE, David M.
Mount, Member, IEEE, Nathan S. Netanyahu, Member,
IEEE, Christine D. Piatko, Ruth Silverman, and Angela Y.
Wu, Senior Member, IEEE, “An Efficient k-Means
Clustering Algorithm: Analysis and Implementation”, IEEE
Transactions On Pattern Analysis And Machine
Intelligence, Vol. 24, No. 7, July 2002.
[13]http://en.wikipedia.org/wiki/Electroencephalograph
y>, (accessed on 12 August 2015).
[14] Bin Hu*, Hong Peng, Qinglin Zhao, Bo Hu, Dennis
Majoe, Fang Zheng, Philip Moore, “Signal Quality
Assessment Model for Wearable EEG Sensor on Prediction
of Mental Stress”, IEEE Transactions on NanoBioscience.
[15] Fei Wang, Shaonan Wang, Xihui Wang, Ying Peng,
Yiding Yang, “Design of Driving Fatigue Detection System
Based on Hybrid Measures Using Wavelet-packets
Transform”, IEEE International Conference on Robotics &
Automation (ICRA) Hong Kong Convention and Exhibition
Center, May 31 - June 7, 2014.
[16] V.Krishnaveni, S.Jayaraman , S.Aravind,
V.Hariharasudhan, K.Ramadoss, “Automatic Identification
and Removal of Ocular Artifacts from EEG using Wavelet
Transform”, MEASUREMENT SCIENCE REVIEW, Volume 6,
Section 2, No. 4, 2006
[17] Chuan-Yi Liu and Pei-Chen Lo, “Investigation of
Spatial Characteristics of Meditation EEG: Using Wavelet
and Fuzzy Classifier”, June 30, 2006.
[18] Dipalee Gupta, Siddhartha Choubey “Discrete Wavelet
Transform for Image Processing”, International Journal of
Emerging Technology and Advanced Engineering, Volume
4, Issue 3, March 2015.

A method for detection and reduction of stress using EEG

More Related Content

What's hot (20)

Similar to A method for detection and reduction of stress using EEG (20)

More from IRJET Journal (20)

Recently uploaded (20)

A method for detection and reduction of stress using EEG