Cognitive Neuroscience and qEEG for Educational Resilience

Griselda Cortés

, Abraham J. Jimenez

and Mercedes Flores Flores

Tecnológico de Estudios Superiores de Ecatepec, Av. Tecnológico S/N C.P. 55210 Col. Valle de Anáhuac,

Ecatepec de Morelos, Estado de México, Mexico

Keywords: Stimulator App, Neurocogitive, Reading, Writing, Object-Oriented Programming.

Abstract: The Program for International Student Assessment examines educational achievement, and the results show

that several countries are below average. The OECD average with a low level of competence in mathematics

is 24%, for Mexico it is 56%, 45% showed growth mindsets. Mexico and the Ministry of Public Education

have made every effort to make education relevant. Today the contingency of COVID-19 encourages the

search for educational alternatives to strengthen teaching-learning methods. The article converges on a

neurocognitive stimulator-based architecture general, protocol and tools that contributes to the functional

student’s regeneration nerve cells, and sensory stimulates the reinforcement of mathematical epistemic

thinking in cognitive abilities (memory, attention, and perception) through the use techniques and artificial

intelligence; contrasted with brain mapping. This tool will support specialists: psychologists, neurologists,

among others; in the interpretation of brain neuroplasticity. The results were contrasted with different ANN’s,

obtaining better performance from the Dendral Processing Network. The tests show that memory, attention,

and perception skills in children increased. When the children use Neurostimulator reinforce their ability.

Finally, the qEEG shows the region with more brain activity during cognitive processing tasks.

1 INTRODUCTION

Education at its different educational levels was

affected by the arrival of the COVID-19

pandemic, the world had to evolve and face new

challenges, drastically changing the way of

educating. The OECD Program for International

Student Assessment (PISA) assesses knowledge

and skills for full participation in the knowledge

society. Unfortunately, some countries are far

below the average in all three subject areas:

science, mathematics, and reading ( Programme

for International Student Assessment, 2021)

This is how the need arises a guide for

pedagogical practice that is a reference on educational

training. Since 2016 begins an evaluation and update

of the National Educational Model in Mexico and

receives the name of "New Mexican Family" (SEP,

2016). Instances such as Accreditation Council for

Engineering Education, A.C. (CACEI) and,

https://orcid.org/0000-0002-1159-0769

https://orcid.org/0000-0003-3058-9082

https://orcid.org/0000-0002-9435-4496

Council for Accreditation of Higher Education

A.C. (COPAES), encourage innovative

teaching-learning practices in higher

education

and consider it relevant to improve the curricular plan

by competencies to strengthen and promote the use of

technological advances, improve teaching, teaching

practice and to improve the overall educational

quality (COPAES, 2021) (UNESCO, 2013).

On the other hand, researchers address problems

with cognitive deficiencies in adults (Boyd , Synnott,

Nugent, Elliott , & Kelly , 2017), (Shi, 2020),

cognitive abilities in the early development of the

infant, and others the emerging role for the adequate

training of the educational neuroscientist.

Unfortunately, these are worked in isolation, on the

one hand, through basic level apps, based on

cognitive theory to help students to learn (Yi, Ruan,

Gao, & Zhang, 2020) (Aboalela, 2016) and reinforce

their cognitive skills (Shi, 2020). Also, using

techniques of Artificial Intelligence (AI) (Pritchard &

al, 2021) and independently using neurofeedback

CortÃl’s, G., Jimenez, A. and Flores, M.

Cognitive Neuroscience and qEEG for Educational Resilience.

DOI: 10.5220/0011919400003612

In Proceedings of the 3rd International Symposium on Automation, Information and Computing (ISAIC 2022), pages 245-251

ISBN: 978-989-758-622-4; ISSN: 2975-9463

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

245

techniques to predict cognitive decline (Giridhar,

Long, & and Mircea, 2020) (Cerda, Pérez, Romera,

Ortega-Ruiz, & Casas, 2017).

This paper establishes a collaborative work by

venturing into AI with Artificial Neural Networks

(ANN), mobile technology, educational cognitive

neuroscience (neurofeedback) and the national basic-

level curriculum proposed by the current Pedagogical

Guide of the SEP. The aim is to present a general

architecture to build or use a stimulating tool that

allows the student to interact and perform brain

mapping or quantitative electroencephalogram

(qEEG) of the cortical modules and identify the

region with the greatest brain activity. Furthermore, a

repository trained with RNA's will be generated,

which allows measuring the level of mathematical

knowledge based on the cognitive abilities (attention,

perception, and memory) of infants.

2 PROPOSED SOLUTION

ARCHITECTURE

To achieve the aim, the Figure 1 shows the solution

where it integrates the strategic techniques of

cognitivism in the curriculum mapping, relying on

mobile technology, to reinforce the Teaching-

Learning Process (TLP) of the basic level exact

sciences. The results of the qEEG Cartography are

contrasted with AI techniques through ANN.

Figure 1: Solution architecture general

Now, Figure 2 shows the standardized protocol

for any subject applicable to students who need to

measure the teaching-learning process at any

educational level, theme, curricular map, including

any cognitive ability.

Figure 2: Protocol for obtaining and processing qEEG

signals

2.1 Development of the Neurocognitive

Stimulator General

The implementation of the architecture (Figure 1)

and using the protocol of Figure 2, a stimulator can

be developed where neurocognitive strategies:

Perception (Per), Attention (Aten), Memory (Mem)

related to the functioning of the human brain and its

biological mechanisms of different areas or themes:

Mathematical (M), Reading (R), Writing (W), and

Object-Oriented Programming (OOP) of any

educational level are implemented, considering

criteria of the Table 1.

Table 1: Thematical to proof architecture general

Thematical M R W OOP

Sample

290 267

204 285

Grade Basic level

Superior

level

Modality

Presential

Virtual

Axis

Sense, Form space and measure, Logical,

Problems resolution

Tools

Application web, Mobil, Batteries,

evaluation test, graphics, images,

audios, videos, etc.

Cognitive

abilities

Attention, Perception, Memory

The neuro stimulator can be from a web application,

mobile, presentations to a series of batteries where

playful activities oriented to cognitive neuroscience

and the curricular map of the selected topic are

implemented. These will be used to train, motivate

and over time the user manages to think critically and

reflectively autonomously, reaching brain

neuroplasticity (retaining as much information as

possible).

ISAIC 2022 - International Symposium on Automation, Information and Computing

246

2.2 Repository Construction

Once the theme, educational level, and cognitive

skills to be analyzed have been selected, see Table 1;

The process of selecting users and noise-free

workspace with Internet begins. In addition, it will be

necessary to collect the personal data of the students

to label records. For

the collection of brain wave

readings, it is necessary: Prepare the Emotiv

headband with sufficient charge, moisturize sensors.

Prepare the Emotiv headband with sufficient

charge, moisturize sensors.

Place students in the classroom or

workspace free of noise.

Put on headband Emotiv Insight, configure,

and connect via Bluetooth or USB.

Identify channel position according to

system 10-20 and brain region (Figure 3),

these are selected according to cognitive

ability, see Table 1.

Start training of selected topic (Object-

Oriented Programming).

At the same time, starting

bandpowerlogger.exe software to collect

data, this API (EMOTIV, 202O) and

Documentation (Emotiv, 2020) is available

on the Emotiv developer’s page.

Figure 3: System 10-20, the highlighted sensors are from

the Emotiv headband a) Epoc X-14 channels, and b)

Insight-5 channels (Emotiv, 2020).

La Table 2 muestra los datos obtenidos con

Emotiv Insight. La primera columna representa la

hora en formato Marca de tiempo, las siguientes

columnas representan los canales AF3, AF4, T7, T8,

Pz, (Figura 3a) con diferentes ondas cerebrales Theta,

Alfa, Beta Baja, Beta Alta y Gamma.

El total es de 26 columnas, datos representados en

microvoltios. Los datos obtenidos de la Table 2 hacen

referencia a un de 10 estudiantes de nivel superior que

fueron instruidos para reforzar la temática 4 en dos

modalidades, ver Table 1. Capacitación realizada en

dos tiempos diferentes (presencial y virtual). El

repositorio que puede obtener en

https://drive.google.com/drive/folders/1JepmwMsvE

uGM7Oknwm3sxzRbyReF07Zd?usp=sharing.

Table 2: Lectura de la diadema EMOTIV INSIGHT

Time AF3/𝜽 AF3/∝ T7/𝜷 Pz/𝜸 T8/𝜽 T8/∝

1633403866 9.895 3.427 4.579 6.576 0.653 0.43

1633403867 14.071 2.845 6.196 6 0.336 0.247

1633403868 10.04 2.875 4.529 7.216 0.308 0.195

1633403869 6.68 3.958 2.572 6.511 0.659 0.334

1633403870 5.597 4.125 2.512 6.572 0.63 0.307

1633403871 108.232 10.331 2.827 8.585 0.581 1.229

1633403872 56.869 9.811 3.408 12.347 0.513 1

1633403873 7.865 5.807 5.371 6.942 0.471 0.633

1633403874 12.491 8.026 7.07 6.851 0.466 0.58

1633403875 128.967 24.681 6.767 10.361 2.117 3.302

1633403876 249.941 32.937 7.259 10.627 2.418 3.917

1633403877 8.081 6.379 6.286 12.085 0.427 0.418

2.3 Generate Brain Mapping or qEEG

This stage performs a treatment of the data in Table 2

to normalize them, standardize the length, the type of

data required and select the most representative

characteristic fields or attributes. To select these last

features, the functions of the channels with respect to

the headband used (Figure 4) and brain waves were

considered, see Table 3.

Figure 4: Brain waves in typical EEG (Emotiv, 2020).

Cognitive Neuroscience and qEEG for Educational Resilience

247

With the data from Table 3 matrix was generated

with the readings of the sensors (AF3, AF4, T7, T8,

Pz, and each with the Beta High and Gamma wave).

The Beta wave is engaged in a task and the Gamma

wave is related to several tasks of high cognitive

processing: the way of learning, the ability to learn

new information and the process of simultaneous

information (Donoghue, Schaworonkow, & Voytek,

2021).

The same process, we used the EMOTIV EPOC+

headset with 14 sensors. For brain mapping (qEEG),

see Figure 7 y Figure the International System 10-20

and sensor configuration were used (AF3, F7, F3,

FC5, T7, P7, O1, O2, P8, T8, FC6, F4, F8, AF4).

Table 3: Electrodes EEG, Regions of the cortex cerebral

associated with brain functions (Bitbrain Technologies,

2018)

Channel

Cerebral

Cortex

Brain Functions related

with brain lobes

AF3,

AF4, F7,

F3, F4,

F8, FC5,

FC6

Frontal

Central

Reasoning, speech and

movement control, emotions

and problem solving.

Sensorimotor

T7, T8

Temporal

Memory, meaning and

interpretation, and

processing of auditory

stimuli

Attention, perception, and

processing of stimuli related

to the senses (temperature,

touch,

ressure,

ain)

Pz, P7, P8

Parental

O1, O2

Occipital Vision

The qEEG is a useful tool to evaluate the

neurophysiological characteristics of individuals with

various disorders and consists of graphically

digitizing the reading of brain waves through colours:

red, green, blue, black, and white. The headband is

connected and configured via Bluetooth to a PC and

through the EMOTIVPRO license, the qEEG data is

obtained in .edf format; This file and through the

execution of EEGLAB functions in MATLAB,

allows neuroimaging.

2.4 Data Analysis and Interpretation

The Criteria for Selecting ANN for Classification by

Level of Knowledge are: The ANN has the 10

characteristic attributes (AF3, AF4, T7, T8, Pz, and

each with the Beta High and Gamma wave), see Table

2, and one represents the class that determines the

level of knowledge (5-10). Official Gazette of the

Federation, 2018 establishes 4 classifications that

indicates expected learnings, see Table 4.

Table 4: Class for training and test

Class

Learning

description

Value or range

Insufficient Menor o igual a 5

NII

Basic

Mayor que 5 y

menor o i

ual

ue 7

NIII

Satisfactor

Mayor que 7 y

menor o igual que 9

NIV

outstandin

ue 9

With the previous data, the repository was

normalized, to train and test its operation using

different classifier methods like: Multilayer

Perceptron (MLP), Nearest Neighbor

(LinearNNSearch), Radial Base Functions (RBF),

Support Vector Machines (SMO) and Neural

Network with Processing Dendritic (DMNN), each

with criteria different, see Table 5. Likewise, identify

the ANN with the best performance.

Table 5: The criteria used for each classifier method

Classifier

Methods

Learning algorithm

Learning rate by

Thematical

M R W OOP

DMNN Hyper boxes 57 59 55 60

MLP

Backpropagation

(Hidden layers)

9 8 10 9

SVM

Hyperplane

(Polynomial

nucleus)

1 1 1 1

KNN

Euclidean distance

(K)

1 1 1 1

RBF

Hybrid learning

(Cluster groups)

2 2 2 2

2.5 Results and Discussions

The tests performed in the First and second phase we

used 1046 instances in four thematical to basic and

superior level in times different. Table 6 shows, the

results obtained of the two stages in the 3 cognitive

skills implementing the different themes of Table 1,

performing 2 tests a) manual or face-to-face, b)

Application or virtual session.

For mathematical thinking was the first test that

we report in (Cortes, Gutierrez, Avila, & Flores,

ISAIC 2022 - International Symposium on Automation, Information and Computing

248

2020). Now, we try with other thematical like

Reading, Writing, and Object-Oriented Programming

and we had good results.

Table 6: Average assessment in the two stages

Thematical

Instances

First phase

Second phase

Manual or

presential

Application or

virtual session

Mem Aten Per Mem Aten Per

M 290 6.6 8 7.4

7.5 8.3 8.1

R 267 7 8.2 7.9

8.1 8.4 8.9

W 204 6.9 7.8 8

7.1 8.5 8

OOP 285 7.2 8 7.5

8 8 7.9

Average evaluation 6.9 8 7.7 7.6 8.3 8.2

The average evaluation of the first stage when

using any of the tools of Table 1 and carried out

manually or face-to-face, reaches a 7.5 with respect

to the three cognitive skills.

Figure 5: Average rating of first phase by theme

For a second stage, training was carried out with

the application or through virtual sessions for two or

three weeks and despite the short training or

reinforcement time 8.0 is reached, see Figure 6.

Figure 6: Average rating of second phase by theme

Obtaining the arithmetic mean for the three

cognitive skills. Figure 7 shows, that there is a

relevant improvement in memory and perception.

Figure 7: Average results of the three cognitive skills in the

First and Second stages

The cartography obtained will serve to contrast

the results obtained in the evaluation stages when

using a neurocognitive stimulator (Junk Kyung, Hye

Youn, & al, 2021). Figure 7 shows the behavior of

each user. In addition, we can see graphically,

dynamic changes in a region of the brain during

cognitive processing tasks. Colors represent brain

activity, blue indicates a deficiency connection, and

red an excessive connection.

Figure 8: qEEG while the users used mathematical

Neurostimulator before stimulation.

For a second phase, parents were asked for their

support so that students could download the app and

use it for two weeks, and a second evaluation was

applied, the results are shown in Figure 6; in which it

is observed that the 1st, 2nd grade students improved

their memory capacity than the 3rd grade students.

3rd grade improved their ability to perceive. In figure

4a it is observed the students of all grades.

Finally, to measure the cognitive learning process

of the different topics during the training and testing

process, configuring the criteria selected from Table

5, a standarized repository was achieved,

implementing the different AI techniques.

Cognitive Neuroscience and qEEG for Educational Resilience

249

Figure 9: qEEG while the users used mathematical

Neurostimulator after stimulation.

The data of Table 7 were obtained using 100%,

80%, 70%, 50% of the samples for training and the

remaining percentage for testing, from the results

obtained the arithmetic mean was obtained reaching

a classification percentage of 99% (Table 4) when

working with DMNN proposed by (Humberto &

Elizabeth, 2014).

Table 7: Behaviour of classifiers methods

Thematical

Classifiers Methods

DMNN MLP SVM KNN RBF

Mathematical 99.757 99.25 96.25 96.75 99.50

Reading 99 100 99 98 97

Writing 98 97 96 95 96

OOP 99.2 99 96.5 95.7 97.7

Overall

avera

99 98.8 96.9 96.3 97.5

3 CONCLUSIONS AND FUTURE

PERSPECTIVES

In the current context of education according to PISA,

this article allows to attend to the education and

integral training of students, achieving the interaction

of Neuroscience, education, current technology and

AI in favor of mathematical cognitive development,

reading, writing, OOP in its 3 skills (attention,

perception and memory) of students from basic to

higher level, impacting the growth of our country and

society, facing the effects on education in Mexico due

to the lags caused by the COVID-19 pandemic.

The architecture, protocol and tools proposed in

this paper ensure that students of any subject

reinforce their cognitive skills and participate actively

and autonomously in their cognition, motivated under

the key competence of "learning to learn", ensuring

that the student reinforces what has been learned,

retaining it in the long term.

In addition, the proposal allowed to measure the

level of mastery of various topics, through the DMNN

due to its great performance. Finally, the data

obtained through a neurostimulator (manual, web,

mobile or virtual) allowed to obtain the brain

mapping or qEEG that could be a support tool and

interpreted by psychologists, neurologists, among

others; to achieve brain neuroplasticity in students of

different educational levels.

For future research, the proposal will be used as a

brain-computer interface using neurofeedback

techniques to stimulate the brain. In addition, we will

obtain your qEEG and automatically perform the

interpretation through the DMNN, without the need

to consult a specialist.

ACKNOWLEDGEMENTS

The autors thanks Tecnológico de Estudios

Superiores de Ecatepec and TecNM for the support to

undertake this investigation through of (CI-

01/2021A) project: re5nt0(10389). The authors

would like to express their gratitude to the students

for their insightful contributions to attain this

research.

REFERENCES

Cerda, G., Pérez, C., Romera, E. M., Ortega-Ruiz, R., &

Casas, J. A. (2017). Influencia de variables cognitivas y

motivacionales en el rendimiento académico en

matemáticas en estudiantes chilenos. Educación XXI, 20(2),

365-385.

COPAES. (03 de 2021). Consejo para la Acreditación de

la Educación Superior a.c. Obtenido de

https://www.copaes.org/noticia18marzo2021.html

Cortes, G., Gutierrez, A., Avila, J., & Flores, M. (2020).

App to Reinforce Mathematical Neurocognitive Skills

During and After the COVID-19 Pandemic. Research

in Computing Science, 37-48.

Programme for International Student Assessment. (22 de 05

de 2021). OECD. Obtenido de

https://www.oecd.org/pisa/data/

Aboalela, R. a. (2016). Model of Learning Assessment to

Measure Student Learning: Inferring of Concept State

of Cognitive Skill Level in Concept Space. 3rd

International Conference on Soft Computing &

Machine Intelligence (ISCMI), 189-195.

Bitbrain Technologies. (2018). Bitbrain. Recuperado el 22

de 01 de 2019, de

https://www.bitbrain.com/es/blog/colocacion-

electrodos-eeg

Boyd , A., Synnott, J., Nugent, C., Elliott , D., & Kelly , J.

(2017). Community-based trials of mobile solutions for

ISAIC 2022 - International Symposium on Automation, Information and Computing

250

the detection and management of cognitive decline.

Healthc Technol Lett, 4(3), 93-96.

Donoghue, T., Schaworonkow, N., & Voytek, B. (2021).

Methodological considerations for studying neural

oscillations . European Journal of Neuroscience,

55(11-12), 1-26.

Emotiv. (s.f.). Cortex API. Recuperado el 25 de Abril de

2020, de https://www.emotiv.com/developer/

Emotiv. (2020). Cortex API. Recuperado el 25 de Abril de

2020, de https://www.emotiv.com/developerx/

EMOTIV. (202O). Github.com. Recuperado el 25 de Abril

de 2020, de https://github.com/Emotiv/cortex-v2-

example

Giridhar, P. K., Long, S., & and Mircea, I. C. (2020). A

neurophysiological brain map: Spectral

parameterization of the human intracranial

electroencephalogram. Clinical Neurophysiology,

131(3), 665-675.

Humberto, S., & Elizabeth, G. (2014). Efficient training for

dendrite morphological neural networks.

Neurocomputing, 131, 132-142.

Junk Kyung, H., Hye Youn, p., & al, e. (2021). Longitudinal

qEEG changes correlate with clinical outcomes in

patients with somatic symtom disorder. Journal of

Psychocomatic Research.

Pritchard, M., & al, e. (2021). Dynamic Fusion of

Electromyographic and Electroencephalographic Data

towards Use in Robotic Prosthesis Control. IOP

Publishing.

SEP. (2016). gob.mx. Obtenido de

https://www.gob.mx/cms/uploads/attachment/file/118

382/El_Modelo_Educativo_2016.pdf

Shi, Y. X. (2020). Research and Design of APP for Primary

School Students' Safety Education Based on Embodied

Cognitive Theory. IEEE 2nd International Conference

on Computer Science and Educational Informatization

(CSEI), 1-4.

UNESCO. (2013). UNESCO. Obtenido de

http://unesdoc.unesco.org/images/0021/002196/21963

7s.pdf.

Yi, C., Ruan, F., Gao, Y., & Zhang, X. H. (2020). QiFei:

Assisting to Improve Cognitive Abilities for Autism

Children Using a Mobile APP. Information

Communication Technologies Conference (ICTC),

297-301.

Cognitive Neuroscience and qEEG for Educational Resilience

251