You are on page 1of 48

Department of Electronics and Telecommunication Engineering

Shri Guru Gobind Singhji Institute of Engineering and Technology Nanded

Attention training and its effect on cognitive workload


using physiological signal

Supervisor : -
Dr.R.R.Manthalkar Presented by :-
Professor Krashana D. Nirde
E&TC Engg.Dept. Reg.No.:2019PEC902
SGGSIE&T, Nanded E&TC Engg.Dept.
SGGSIE&T, Nanded

02/11/2024 1
OUTLINE

 Problem Statement
 Objectives
 Introduction
 Literature Survey
 Contribution 1
 Contribution 2
 Contribution 3
 References

02/11/2024 2
PROBLEM STATEMENT

 To design robust and optimistic system which is used to measure the cognitive load
while doing the arithmetic operation and logical reasoning operation.

02/11/2024 3
OBJECTIVES

 Quantification of cognitive load while doing arithmetic’s operations using


physiological signal.

 Quantification of cognitive load while doing logical reasoning operations using


physiological signal.

 Analysing synchrony among various brain regions while doing arithmetic’s


operations & logical reasoning operations.

 Quantifying verbal cognitive load developing proper tests.

02/11/2024 4
INTRODUCTION

 Electroencephalography (EEG) is a method to record an electrogram of the spontaneous


electrical activity of the brain.

 It is typically non-invasive, with the EEG electrodes placed along the scalp (commonly
called "scalp EEG") using the International 10-20 system or variations of it.

Fig.1: 10-20 International system of EEG 19 channel Electrode placement

A healthy human EEG will show certain patterns of activity which correlate with how awake a
person is.

02/11/2024 5
 The range of frequencies one observes are between 1 to 30 Hz and amplitudes will vary
between 20 to 100 μV.

 The observed frequencies are subdivided in various groups, these are alpha (8-13 Hz), beta
(13-30 Hz), delta (0.5-4 Hz) and theta (4-7 Hz).

 Alpha waves are observed when a person is in a state of relaxed wakefulness and are mostly
prominent over parietal and occipital sites.

 During intense mental activity beta waves are more prominent in frontal areas as well as
other regions. If a relaxed person is told to open their eyes one observes alpha activity
decreasing and an increase in beta activity.

 Theta and delta waves are not seen in wakefulness and if they are it is a sign of brain
dysfunction.

02/11/2024 6
 An emerging body of literature suggests that attention training activates regions of the brain
involved in the monitoring and regulation of emotion, attention, and autonomic body
functions.

 In student life, the cognitive loads increase due to cognitive demands during examination
like memorization and mathematical calculations. In order to reduce the anxiety and fear
during critical situations, the physical and mental stability are important.

 Attention training affects many cognitive abilities. The stress and anxiety are reduced by the
attention training practice.

 Attention training has long been used for the maintenance of well-being and its gradual.
disassociation from religious practice has allowed it to be subjected to scientific inquiry.

02/11/2024 7
LITERATURE SURVEY

 H. Nagendra, Vinod Kumar, and S Mukherjee. Hindawi Publishing Corporation


Computational and Mathematical Methods in Medicine Volume 2015, Article ID821061,
13 pages http://dx.doi.org/10.1155/2015/821061.
The attention training practising group showed increased 𝛼 , 𝛽 , and 𝛿 EEG band powers
and significant reduction in 𝜃 and 𝛾 band powers. The increased 𝛼 and 𝛽 power can
represent enhanced cognitive functions such as memory and concentration, and that of 𝛿
signifies synchronization of brain activity.
The heart rate index 𝜃/𝛼 decreased, neural activity 𝛽/𝜃 increased, attention resource
index 𝛽/(𝛼 + 𝜃) increased, executive load index (𝛿 + 𝜃)/𝛼 decreased, and the ratio (𝛿
+ 𝜃)/(𝛼 + 𝛽) decreased.

02/11/2024 8
Attention training practising group showed significant improvement in various cognitive
functions, such as performance enhancement, neural activity, and executive function.
The effect of the cognitive load has been observed at the frontal and temporal lobe. The
reaction time has been increased as load level increases

02/11/2024 9
 N. Jadhav, R. Manthalkar and Y. Joshi Department of Electronics and
Telecommunication Engg.SGGSIE&T,Nanded-India.B.Iyer,S.Nalbalwar and R.Pawade
(Eds.) ICCASP/ICMMD- 2016. Advances in Intelligent Systems Research. Vol. 137,
Pp. 433-439. © 2017- The authors. Published by Atlantis Press.
In this paper, the effect of attention training on the seven load arithmetic tasks
using Electroencephalogram (EEG) is investigated. The focused attention on breathing
is taught to 11 healthy subjects as a simple attention training technique. EEG signals
have been recorded at the beginning of attention training and after four weeks of regular
attention training (daily 20 minutes) using EMOTIV device.
The higher order crossings (HOC) features are derived from EEG signals for
analysis of seven cognitive loads. The pre and post-experiment results are compared
based on reaction time and effect on brain lobe. The K-Nearest Neighbors (K-NN)
classifier is used for classification of seven loads and achieved 83.57% accuracy in the
post and 89.61 % accuracy in pre-experiment.
02/11/2024 10
Table 1: Analysis of different task level

N. Jadhav, R. Manthalkar and Y. Joshi Department of Electronics and Telecommunication


Engg.SGGSIE&T,Nanded-India.B.Iyer,S.Nalbalwar and R.Pawade (Eds.)
ICCASP/ICMMD- 2016. Advances in Intelligent Systems Research. Vol. 137, Pp. 433-
439. © 2017- The authors. Published by Atlantis Press.
In this paper, the effect of attention training on the seven load arithmetic tasks using
Electroencephalogram (EEG) is investigated. The focused attention on breathing is
taught to 11 healthy subjects as a simple attention training technique. EEG signals have
been recorded at the beginning of attention training and after four weeks of regular
attention training (daily 20 minutes) using EMOTIV device. The higher order crossings
(HOC) features are derived from EEG signals for analysis of seven cognitive loads.
The pre and post-experiment results are compared based on reaction time and effect on
brain lobe. The K-Nearest Neighbors (K- NN) classifier is used for classification of
seven
02/11/2024 11
 Anders Nesvold, Morten W Fagerland, Svend Davanger, Øyvind Ellingsen, Erik E
Solberg, Are Holen, Knut Sevre and Dan Atar. European Journal of Preventive
Cardiology 19(4) 773–780! The European Society of Cardiology 2011 .DOI:
10.1177/17418267.11414625.
Attention training practices are in use for relaxation and stress reduction. Some studies
indicate beneficial cardiovascular health effects of attention training. The effects on the
autonomous nervous system seem to vary among techniques.
The purpose of the present study was to identify autonomic nerve activity changes
during nondirective attention. There is an increased parasympathetic and reduced
sympathetic nerve activity and increased overall HRV, while practising the technique.
Hence, nondirective attention by the middle aged may contribute towards a reduction
of cardiovascular risk.

02/11/2024 12
 Radhika Desai, Anisha Tailor, Tanvi Bhatt Department of Physical
Therapy,UniversityofIllinoisatChicago,Chicago,USA.http://dx.doi.org/10.1016/j.ctc
p.2015.02.0021744-3881/©2015Elsevier Ltd.

It was concluded that breathing, attention training, increased overall brain wave activity.
Attention training practice may be an effective adjunctive treatment for a clinical and
healthy aging population.

Attention training seems to have positive effects on brainwave activity in terms of


stimulating the activation of alpha, beta, and theta brainwaves, which have been
associated with improvements in cognition, memory, mood, and anxiety. Attention training
training has been correlated with decreased amygdala activation and decreased negative
emotion in response to emotional distracter images.

02/11/2024 13
Attention training also seems to have a constructive effect on the anatomy of the brain.
The implementation of attention training into the clinical treatment of certain neurological
and psychosocial disorders may be beneficial to these populations because of its
neuroplastic effects.

02/11/2024 14
CONTRIBUTION 1

Title: - Attention and classification of arithmetic operation tasks by using machine


learning algorithm.(Accepted (Book Chapter) in ICSDP 2022 VIT Bhopal, India).

02/11/2024 15
DATASET EXPLANATION

 EEG data collection and validation of dataset is time consuming and tedious
task .however , several existing datasets are available in the literature for cognitive
load classification. Also it is the best practice to use existing dataset to prove
effectiveness of the proposed method. going through literature we found best suited
dataset for cognitive load classification for arithmetic operation.
In Electroencephalogram dataset for Mental Arithmetic Task is benchmark
dataset for EEG cognitive load classification. It includes EEG signals for cognitive load
condition such as, a) Normal Task, b) Difficult Task. Data collected from 36 healthy
humans includes 9 males and 27 females whose age lies between 17 to 26 year old.

Continue…

02/11/2024 16
 Neurocom monopolar EEG 23 channel instrument XAI-MEDICA is used for recorded
EEG signal for mental arithmetic task. Signals are recorded using the international
10/20 technique. The metalic electrodes were deployed at symmetrical anterior-
frontal, frontal, Central, parietal, and temporal of skull (Fp1, Fp2, F3, F4, Fz, F7, F8,
C3, C4, Cz, P3, P4, Pz, T3, T4, T5, T6 respectively).

Continue…

Fig.2: The EEG data was collected during the experiment.

02/11/2024 17
CLASSIFICATION

 In this study, three classification methods were used: KNN (K Nearest


Neighbours), SVM (Support Vector Machine), and Decision Tree for mixed EEG
dataset.
1) KNN: - K-nearest neighbours is a technique for non-parametric classification
and regression. It's one of the most fundamental machine learning methods.
The main idea behind KNN is to look about your neighbourhood, make the
assumption that the test datapoint is equivalent to them, and then calculate the
outcome.

Fig.3: Classification of two different arithmetic task by using KNN

02/11/2024 18
2) SVM: The SVM is a common Supervised Learning technique for both cases
such as classification and regression related problems . but in most of the time it is
used to solve the classification related problems.
The main focus of the SVM is to detect the most accurate line or decision
boundry for classifying n-dimensional space into classes so that subsequent
datapoints can be placed in the most suitable class. and the best potential choice
boundary is a hyperplane.

Fig.4: Classification of two different arithmetic task by using SVM

02/11/2024 19
3) Decision Tree: - A tree-based technique called a decision tree is used to tackle
regression and classification problems. For dependent variables with discrete values,
classification trees are utilised.

Fig.5: Working principle of Decision Tree

02/11/2024 20
 Evaluation Metrices Parameters : -

Accuracy = (TP+TN) / (TP+FN+TN+FP) 1

Precision = (TP) / (TP+FP) 2

Recall = (TP) / (TP+FN) 3

F1 – Score = (2*Precision*Recall) / (Precision+ Recall) 4

Where TP = True Positive ,


TN = True Negative ,
FN = False Negative ,
FP = False Positive ,

02/11/2024 21
RESULTS

Table 2: Classification Results for mixed EEG Dataset

Classifier Name and Min/Max Scaler Standard Scaler Without Scaler


Accuracy%

KNN (%) 63 44 53
SVM (%) 63 63 63
Decision Tree (%) 59 62 62

Table 3: Evaluation metrics comparison of KNN classifier

Name of the Scaler Classifier: KNN


(Acc%) (Pre) (Rec) (F1)
Min/Max Scaler 63 0.63 0.99 0.77
Standard Scaler 44 0.56 0.51 0.53
Without Scaler 53 0.60 0.73 0.66

02/11/2024 22
Table 4: Evaluation metrics comparison of SVM classifier
Name of the Scaler Classifier: SVM

(Acc%) (Pre) (Rec) (F1)

Min/Max Scaler 59 0.61 0.92 0.74

Standard Scaler 63 0.63 1.00 0.77

Without Scaler 63 0.63 1.00 0.77

Table 5: Evaluation metrics comparison of Decision Tree classifier


Name of the Scaler Classifier: Decision Tree

(Acc%) (Pre) (Rec) (F1)

Min/Max Scaler 59 0.37 0.96 0.53

Standard Scaler 62 0.62 0.99 0.76

Without Scaler 62 0.62 0.99 0.76

02/11/2024 23
CONCLUSION

 In this research work, the EEG signals of 36 subjects were recorded in two different
states, such as: a) Normal Task, b) Difficult Task. We used a new protocol that was
based on mathematical operations.
 The EEG signals were recorded using 21 channels, The frontal, central, parietal,
occipital, and temporal portions of the brain are all included.
 Three distinct classifiers, including KNN (K=3), SVM, and Decision Tree, were
individually applied to the mixed EEG data with variable scaling, such as 1)
Min/Max Scaler, 2) Standard Scaler 3) With and without the Scaler, the results were
compared.
 SVM is superior because it yields 63 percent more results in all scaling approaches.
The next step in this research will be to apply alternative classifiers, such as CNN,
ANN, and deep learning models, to improve the classifier's accuracy.

02/11/2024 24
CONTRIBUTION 2

Title: -EEG mental arithmetic task levels classification using machine learning and
deep learning algorithms.(Submitted to ICAISP 2023 VIT-AP, University, India).

02/11/2024 25
CLASSIFICATION

 In this study, three machine learning classification methods were used: KNN (K
Nearest Neighbours), SVM (Support Vector Machine), and Decision Tree for mixed
EEG dataset. And two deep learning methods were used : (ANN, and LSTM).

02/11/2024 26
RESULTS

Table 6: Evaluation metrics comparison of different machine learning algorithm


classifier with our EEG Dataset

Proposed Classifiers Accuracy Precision Recall F1-Score


method

Decision Tree 65 73 69 69
Algorithm_1 KNN 91 95 96 95
SVM 89 92 94 91

Table 7: Hyper parameters for ANN and LSTM architecture

Proposed DL model Optimizer Epoch Batch size Loss


method function
ANN rmsprop 500 8 MSE,
Algorithm__2 categorical
cross entropy
LSTM adam, SGD 110 8 MSE,
categorical
cross entropy

02/11/2024 27
Table 8: Architecture using ANN and LSTM

Proposed DL model No. of Layers Activation Dropout Learning


method function used ratio

ANN 2 Relu, Softmax 0 0.0001-0.3


Algorithm_
_2
LSTM 2 Relu, Softmax 0.2 0.0000-0.2

Table 9: Evaluation metrics comparison of different deep


learning algorithm classifier with our EEG Dataset

Proposed method Classifiers Accuracy Precision Recall F1-


Score

ANN 96.80 97 97 97
Algorithm_2
LSTM 94 1.00 96 98

02/11/2024 28
Table 10: A comparative report for our proposed approach with the existing
methods
. Different predicted
Authors Database Classifier used Obtained Accuracy
levels

Two different of
(Djamal et al., 2016) EEG SVM 77.00-83.00
attention

Three different states of


(Myrden & Chau, 2017) EEG SVM 71.6-84.8
attention

(Alirezaei & Sardouie, Two different of


EEG SVM 92.80
2017) attention

EEG engagement Two different states of


(Nuamah & Seong, 2018) SVM 93.33 ± 8.16
index attention

(Acı, Kaya & Mishchenko, Three different levels of 96.70 (best)


EEG SVM
2019) concentration 91.72 (avg.)

Decision Tree 65

KNN 91
Three different levels of
Proposed approach for task (i.e., no task,
Only EEG SVM 89
Algorithm_1 and_2 normal task, difficult
task)
ANN 96.80

LSTM 94

02/11/2024 29
CONFUSION MATRIX

Fig.6.a) Decision tree Fig.6.b) KNN

Fig.6.c) SVM Fig.6.d) LSTM

02/11/2024 30
TRAINING VS VALIDATION ACCURACY

Fig.7.a) ANN Fig.7.b) LSTM

02/11/2024 31
CONCLUSION
 In this research work, the EEG signals of 36 subjects were recorded in three
different states, such as: a) no task, b) normal task, c) difficult task. The signals were
recorded using 21 channels, The frontal, central, parietal, occipital, and temporal
portions of the brain are all included.
 Three machine learning classifiers, including KNN, SVM, and Decision Tree, were
individually applied to the mixed EEG dataset with different features extraction such
as skewness, kurtosis, mean, standard deviation, and variance.
 The results were compared with existing one using one classifier i.e., SVM. SVM
obtained the highest accuracy of 96.70%(best) and 91.72(avg).
 But here we obtained a maximum accuracy in deep learning models as compared to
machine learning classifiers. ANN is superior because it yields 96.80% accuracy
than LSTM and machine learning classifiers.

02/11/2024 32
CONTRIBUTION 3

Title: -Assessing the Effect on Cognitive Workload Index, EEG Band Ratio’s, Band
Frequencies Using Band Power and Implementing ML Classification (Accepted
(Book Chapter )in ICEBEHI 2022 Surabaya , Indonesia.

02/11/2024 33
 In this paper, the dataset collected from publicly available online sources and
consists of 36 subjects performing mental arithmetic tasks (series of subtraction).
Based on their performance on the subtraction task subjects are divided into two
groups: “Bad” and “Good”.
 Each group is divided into two parts: before the mental arithmetic task of 180
seconds (resting state), and while doing the mental arithmetic task of 60 seconds
(response).

02/11/2024 34
METHODS
 The nature of EEG signal has a time varying property, in order to convert the time
domain signal to frequency domain we used wavelet transform. Wavelet transform
can provide the time and frequency signals simultaneously.
 A wavelet family with mother wavelet (x) consists of functions a,b(x) of the form
5
 The continuous wavelet transform (CWT) of a function f(x), introduced by Morlet,
W f (a, b) = 6
 In discrete wavelet transform (DWT), signals are discrete, rather than
continuous.The coefficients are defined as before, except:
1. Only particular values of a and b are used
2. Due to the discrete representation of the signal, the integrals that define the
coefficients must be computed numerically.

02/11/2024 35
 DWT of a function f(x) is given by
W f (m, n) = = 7
Where
 In this paper, we used DWT to obtain different levels of frequency bands.

Fig.8: 1-D wavelet Decomposition

 By using wavedec function in MATLAB, given the sampling frequency of 500Hz


and obtained 5 sets of frequency bands like: delta, theta, alpha, beta & gamma. The
obtained frequency bands are analysed to acquire the band power.

02/11/2024 36
 Band Power consider the energy distribution in signal and computed as a sum of the
squares of the signal data points.
BP = 8

1) Individual Band Frequencies:


The subjects are divided into 2 groups: “Good” & “Bad”, the task involves before
mental arithmetic task, while doing mental arithmetic task. Individual band frequencies
used to assess the performance of the subject by considering the task in both, before
and while doing.

02/11/2024 37
Fig. 9: Analysis of “Good” subjects performing before and while doing task(a)Delta band, (b)Theta band,
(c)Alpha band, (d)Beta band, (e)Gamma band.

02/11/2024 38
Fig. 10: Analysis of “Bad” subjects performing before and while doing task(a)Delta band, (b)Theta band,
(c)Alpha band, (d)Beta band, (e)Gamma band.

2) Cognitive Load indices (EEG band ratios):


EEG cognitive load indices which represent the different EEG band ratios are analysed
using mean and standard deviation are shown in Table 11.
Mean = 9

where
n = number of values, = data set values

02/11/2024 39
Standard Deviation:
10
σ = standard deviation, N = size of elements, = each value from the element, μ = the
element mean
Table 11: Comparison of different EEG band ratios between Bad and Good subjects
performing before(3min) & while doing(1min) task using mean & standard deviation.

EEG band ratio Bad(3min) Bad(1min) Good(3min) Good(1min)


Activity Index
Before While doing Before While doing

Performance Enhancement 2.99±4.09 2.17±3.69 1.55±2.14 1.49±0.90

0.72±1.34 1.22±0.62 1.94±2.07 1.16±1.23


Arousal Index

2.17±5.51 2.65±2.31 3.02±4.45 1.74±1.12


Neural Activity

0.54±1.08 0.83±0.49 1.18±1.41 0.69±0.59


Engagement

0.33±0.24 0.46±0.27 0.64±0.46 0.66±1.10


Load Index

1.83±0.92 1.19±2.03 0.84±0.70 1.43±1.69


Alertness

1.21±2.19 1.64±1.03 2.08±2.44 1.32±1.12


CWI

02/11/2024 40
Fig.11: CWI for Subjects performing before, while doing task, (a) “Good” (b) “Bad”

 CWI performance for "Good" and "Bad" subjects with band power is shown in Fig.
11.(a)(b). CWI represents the load on subjects in which there are more (line above)
"Bad" subjects than "Good" subjects (line above). Due to the load, the "Bad"
subjects had difficulty in performing the task (while doing task), whereas the
"Good" subjects were more conscious, concentrated, and relaxed (while doing task).

02/11/2024 41
3) Prefrontal Electrode analysis :
The prefrontal electrodes used to assess the effect of load on the skull, we examined the
individual data of Fp1, Fp2 electrodes for the subjects under before and while doing
task shown in Fig. 12. Various cognitive skills can be analysed but we evaluated only
cognitive load on subjects performing the task. Band powers of Fp1 and Fp2 for
“Good” and “Bad” subjects are more in while doing task (line above), whereas less in
before the task (line below).

Fig.12: Fp1, Fp2 for subjects performing before and while doing task, (a)(b) “Good”, (c)(d) “Bad ”

02/11/2024 42
4) Binary Classification :
The binary classification is done by taking mean of individual (Fp1, Fp2, F3, F4, Fz,
F7, F8, C3, C4, Cz, P3, P4, Pz, O1, O2, T3, T4, T5, T6) electrodes & calculated overall
mean of 19 electrodes for one subject, this process repeated for all subjects and finally
obtained mean results are divided into 2 parts: 1.“Bad” before mental arithmetic(3min)
and while doing mental arithmetic task(1min), 2.“Good” Subjects before mental
arithmetic(3min) and while doing mental arithmetic task(1min).The obtained
mean(using eq-9) of all the subjects considered & divided among “Bad” , “Good”.
The subjects mean merged (before, while doing mental arithmetic task of “Bad”) and
labelled as ‘0’(zero). The same repeated for “Good” subjects and labelled as ‘1’(one).
Different Classifier Outputs
80
73 73 73
70 67

60 60
60

50
Percentage

40

30

20

10

0
SVM KNN DT RT GNB LR
Classifier

Fig.13: Different classifier outputs-SVM, KNN, DT, RT, Gaussian NB, LR


02/11/2024 43
CONCLUSION

 In this paper, we presented a framework in four experiments, in first experiment we


have analysed the individual band frequencies with band power, “Good” subjects
showed better performance in while doing task of alpha and beta band, while “Bad”
subjects showed better performance in delta and beta band,

 Second experiment includes the performance of EEG band ratios obtained by mean and
standard deviation of “Bad” and “Good” subjects (before mental arithmetic(3min) and
while doing mental arithmetic task(1min) The variations among the bands resulted to
increase and decrease in different activity indices,

 Third experiment we considered prefrontal electrodes to know the effect of load on the
task,

 Fourth includes binary classification with mean using different classifiers.

02/11/2024 44
REFERENCES
[1] H. Nagendra, Vinod Kumar, and S Mukherjee. Hindawi Publishing Corporation Computational and Mathematical Methods in
Medicine Volume 2015, Article ID821061, 13 pages http://dx.doi.org/10.1155/2015/821061.

[2] N. Jadhav, R. Manthalkar and Y. Joshi Department of Electronics and Telecommunication Engg.SGGSIE&T,Nanded-
India.B.Iyer,S.Nalbalwar and R.Pawade (Eds.) ICCASP/ICMMD- 2016. Advances in Intelligent Systems Research. Vol. 137, Pp. 433-
439. © 2017- The authors. Published by Atlantis Press.

[3] Anders Nesvold, Morten W Fagerland, Svend Davanger, Øyvind Ellingsen, Erik E Solberg, Are Holen, Knut Sevre and Dan Atar.
European Journal of Preventive Cardiology 19(4) 773–780! The European Society of Cardiology 2011 .DOI:
10.1177/17418267.11414625.

[4] Radhika Desai, Anisha Tailor, Tanvi Bhatt Department of Physical Therapy, University of Illinois at Chicago,
Chicago,USA.http://dx.doi.org/10.1016/j.ctcp.2015.02.0021744-3881/©2015Elsevier Ltd.

[5] Igor Zyma, Sergii Tukaev, Ivan Seleznov, Ken Kiyono, Anton Popov, Mariia Chernykh, and Oleksii Shpenkov (2019, January)
“Electroencephalograms during mental arithmetic task performance”. Data 2019, 4, 14; doi:10.3390/data4010014

[6] Sun, H., Bi, L., Chen, B., & Guo, Y. (2015). EEG-based Safety Driving Performance Estimation and Alertness Using Support Vector
Machine. International Journal of Security and Its Applications, 9(6), 125-134.

[7] Wan, Z., He, J., & Voisine, A. (2013, October). An attention level monitoring and alarming system for the driver fatigue in the
pervasive environment. In International Conference on Brain and Health Informatics (pp. 287-296). Springer International Publishing.

02/11/2024 45
[8] Zyma, I., Tukaev, S., Seleznov, I., Kiyono, K., Popov, A., Chernykh, M., & Shpenkov, O. (2019). Electroencephalograms during
mental arithmetic task performance. Data, 4(1), 14

[9] Höll, M., & Kantz, H. (2015). The relationship between the detrendend fluctuation analysis and the autocorrelation function of a
signal. The European Physical Journal B, 88(12), 1-7.
[10] Kiyono, K. (2015). Establishing a direct connection between detrended fluctuation analysis and Fourier analysis. Physical Review
E, 92(4), 042925.
[11] Gorgolewski, K. J., Auer, T., Calhoun, V. D., Craddock, R. C., Das, S., Duff, E. P., ... & Poldrack, R. A. (2016). The brain imaging
data structure, a format for organizing and describing outputs of neuroimaging experiments. Scientific data, 3(1), 1-9.

[12] Farsi, L., Siuly, S., Kabir, E., & Wang, H. (2020). Classification of alcoholic EEG signals using a deep learning method. IEEE
Sensors Journal, 21(3), 3552-3560.
[13] Acı, Ç. İ., Kaya, M., & Mishchenko, Y. (2019). Distinguishing mental attention states of humans via an EEG-based passive BCI
using machine learning methods. Expert Systems with Applications, 134, 153-166.
[14] Alirezaei, M., & Sardouie, S. H. (2017, November). Detection of human attention using EEG signals. In 2017 24th National and 2nd
International Iranian Conference on Biomedical Engineering (ICBME) (pp. 1-5). IEEE.
[15] Ibrahim, S. W., Djemal, R., Alsuwailem, A., & Gannouni, S. (2017). Electroencephalography (EEG)-based epileptic seizure
prediction using entropy and K-nearest neighbor (KNN). Communications in Science and Technology, 2(1).
[16] Myrden, A., & Chau, T. (2017). A passive EEG-BCI for single-trial detection of changes in mental state. IEEE Transactions on neural
systems and rehabilitation Engineering, 25(4), 345-356.
[17] Nuamah, J. K., & Seong, Y. (2018). Support vector machine (SVM) classification of cognitive tasks based on
electroencephalography (EEG) engagement index. Brain-Computer Interfaces, 5(1), 1-12.
[18] Kim, Y., & Choi, A. (2020). EEG-based emotion classification using long short-term memory network with attention
mechanism. Sensors, 20(23), 6727.

02/11/2024 46
[19] Nagabushanam, P., Thomas George, S., & Radha, S. (2020). EEG signal classification using LSTM and improved neural network
algorithms. Soft Computing, 24(13), 9981-10003.

[20] Madhusmita, M., Mousumi, B., Narayan, P. D., & Kumar, M. S. (2019). A novel method for epileptic eeg classification using dwt,
mga, and anfis: A real time application to cardiac patients with epilepsy. In Cognitive Informatics and Soft Computing (pp. 525-534).
Springer, Singapore.

[21] F. E. Gunawan, K. Wanandi, B. Soewito, S. Candra, and N. Sekishita, ‘‘Detecting the early drop of attention using EEG signal,’’ in
Proc. 4th Int. Conf. Electr. Eng., Comput. Sci. Informat. (EECSI), Sep. 2017, pp. 1–6, doi: 10.1109/EECSI.2017.8239175.

[22] A. List, M. D. Rosenberg, A. Sherman, and M. Esterman, ‘‘Pattern classification of EEG signals reveals perceptual and attentional
states,’’ PLoS ONE, vol. 12, no. 4, pp. 1–23, 2017, doi: 10.1371/journal.pone.0176349.

[23] Ibrahim, S. W., Djemal, R., Alsuwailem, A., & Gannouni, S. (2017). Electroencephalography (EEG)-based epileptic seizure
prediction using entropy and K-nearest neighbor (KNN). Communications in Science and Technology, 2(1).

[24] C.-J. Peng, Y.-C. Chen, C.-C. Chen, S.-J. Chen, B. Cagneau, and L. Chassagne, ‘‘An EEG-based attentiveness recognition system
using Hilbert–Huang transform and support vector machine,’’ J. Med. Biol. Eng., vol. 40, no. 2, pp. 230–238, Apr. 2020, doi:
10.1007/s40846-019-00500-y

[25] J. Kim, J. Seo, and T. H. Laine, ‘‘Detecting boredom from eye gaze and EEG,’’ Biomed. Signal Process. Control, vol. 46, pp. 302–
313, Sep. 2018, doi: 10.1016/j.bspc.2018.05.034.

02/11/2024 47
Thank You
!!!

02/11/2024 48

You might also like