The Design of the Fuzzy Inference System for the Determination of Attention

doi:10.4236/eng.2013.55B012

Paper Menu >>

Journal Menu >>

Engineering, 2013, 5, 58-62

doi:10.4236/eng.2013.55B012 Published Online May 2013 (http://www.scirp.org/journal/eng)

The Design of the Fuzzy Inference System for the

Determination of Attention

Hye Jin Kim1, Sun K. Yoo2*

1Graduate School of Biomedical Engineering, Yonsei University, Seoul, Korea

2Department of Medical Engineering, College of Medicine, Yonsei University, Seoul, Korea

Email: hjkim[email protected], *[email protected]

Received 2013

ABSTRACT

In this study, by using the response speed and the number of errors resulting from the children’s concentration test

through the fuzzy inferenc e system and comparing it to the theta which is one of the EEG’s parameter to find the level

of concentration. Targeting 21(Male 12, Female 9) healthy children between the ages of 10 - 14, the test was conducted

one time with a duration of 14 minutes. For the first 5 minutes the children were listening to the Bach’s Air on a G

string having a steady state and the next 9 minutes the children were subjected to the external stimuli audiogenic stimu-

lation that induces attention co ncen tration. When the number 3 wa s heard, ch ildren w ere subjected to press down on the

spacebar to check the response speed and the number of errors. By conducting computerized neurocognitive function

test to compare the theta wave related to the concentration with th e response speed and the number of errors that deter-

mines the attention con centration through the fuzzy system, the data from 15 children out of 21 have shown the results

for the concentration. In order to check the concentration level, a fuzzy inference system which was designed by the

user could be used.

Keywords: Attention; Computerized Neuro-physiological Test; Fuzzy system

1. Introduction

The modern society that is represented by knowledge •

Information shows an interest on the importance of

learning and learning for school aged children is growing

more than anything else. The factors that affect the

learning of children this time is the basic cognitive abili-

ties to focus attention on a particular object or targ et. The

children's ability to learn is in correlation between the

attitude and the attention concentration and the restric-

tions and the lack of the attention concentration is the

primary cause of poor learning ability [1,2].

The attention con centration is an ability [3 ] to focu s on

the consciousness from an internal • external stimuli, the

process that focus selectively[4] on relevant stimuli ac-

cording to the info rmation of the sensory receptors th at is

required for learning will select its input and filtrative

function [5,6]. Without the attentio n concentration, better

information processing is not possible and decides what

to focus on for ourselves and what information to transfer

in meaningful form that can be stored from the sensory

memory [7].

Children lacking the attention concentration have de-

clining attention for the surroundings in daily life and

their attention concentration becomes distracted and dis-

turbed even with a small stimuli showing problems such

as hyperactivity, unstable emotions or poor self-esteem

[8].

Also, they can be focused on the things of their interest

but monotonous and repetitive, and when performing a

tedious task it seems difficult to concentrate con tinuously

[9,10]. They cannot sit still du ring a school hour but run-

ning around with friends or simply doing other behaviors

can be cautioned but only for a little while and even tually

they will again show an erratic behavior [11]. Therefore

attention concentration is a mandatory basic skill that is

needed for the school aged children and teenager’s

learning, acquiring new information and adap tin g to lo cal

communities [12].

Nowadays, computerized neuropsychological tools are

commonly used to measure the attention concentration

and such inspection tools contributes to more objective

scoring results and interpretation, Compared to such tra-

ditional rating scale, expenses can be reduced for the

time it takes for the implementation of the inspection,

grading and reporting results, with an advantage of ques-

tions of the presenting time and inspection, controlling of

conducting time and it is easy to standardize the implan-

tation for the inspection [13].

*Corresponding author.

H. J. KIM, S. K. YOO 59

In recent years, using computerized neuropsychological

tests not only help to improve children's concentration

but ADHD group and normal children, and may be used

not for ADHD but neurocognitive function appearance

for a group of pediatric metal disord ers [13]. In addition,

it is useful to be clinically used for patients with brain

damage which i s widely used i n va rious fields.

However, the attention concentration test using a

computerized neurocognitive program cannot be clear

about how much it had induced the child to be focused

on the test. To have more definitive interpretation many

studies have been conducted together with a bio-signal

measurement. Especially an analysis for concentration

using an EEG is being conducted and the status of con-

centration from a stimuli induced events can be seen

through the alpha and theta wave in a median line of the

brain [14], and also during a short-term concentration

experimental test due to external stimuli and based on

previous research results that have indicated an increase

in the theta rhyth m, in order to find the level o f attention

concentration, an EEG such as theta can be an important

indicator [15].

Therefore, in this study using computerized neurocog-

nitive program to perform a concentration test for the

child who requires the attention concentration and by

designing the Fuzzy Inference System that determines

the level of concentration obtained from the response

speed and the number of errors, in order to assess the

system checking the level of concen tration during the test

using definit e parameters theta wave was co nd uct e d .

2. Methods

2.1. Participants

This test was conducted using health y students, 12 males

and 9 females with total of 21 persons without a history

of cardiovascular and neur ological disorders. An average

age for the children was 11 with the range of 10-14 years

of age and included elementary 3rd graders to freshmen

of junior high school. The subjects were recruited through

the Electronics and Telecommunications Research Insti-

tute (ETRI) bulletin board. All subjects have performed

the concentration inducing test using the external stimuli

[16].

2.2. Experimental Design

An external stimuli that induces the attention concentra-

tion, CNT4.0 (Computerized Neuro-physiological Test:

Computerized Neurocognitive tests) commonly used in

the hospital program was used [17]. The audiometry test

used within the CNT program for this study was Audi-

tory Continuous Performance Test which is one kind of

CPT (Continuous Performance Test: Continuous Per-

formance Test). This test was performed by having the

subjects to listen to the target stimulation of the number

‘3’ where the subjects have to press the response button

as fast as possible in order to measure the number of

correct response, the number of omission errors and the

number of false alarm errors. For the correct response is

randomly presented 15 times every 60 seconds, distur-

bance stimulation 5 times each and the response time was

measured in the units of 1/100 seconds.

The basic paradigm of CPT makes occasional target

stimuli or possible to measure selective attention to rele-

vant stimuli. Therefore, the CPT program performs the

inspection by rapidly presenting the target stimuli or tar-

get pattern while constantly changing and the time

needed to perform the test will vary with each inspection

but it was designed sufficiently to measure the attentive-

ness [13,17].

The screen shown in Figure 1 was an actual CNT in-

spection monitored using a CNT program and the resu lts

of the correct response, the number of omission errors

and the number of false alarm errors can be seen(Table 1 )

Figure 1. CNT hearing test results.

H. J. KIM, S. K. YOO

Table 1. The attention concentration, computer assignments

test results items.

Item test results Interpreting the results

Correct response Measuring the lev el of concentration by subject’s

correct response.

The number of

omission errors Measuring the area of non-response to the target

stimuli by the s ubje cts.

The number of

comission errors Measuring the impulse that occurs when subjects

in response to non-tar get stimuli and disinhibi t ion.

The test process was performed by attaching the EEG

electrodes to the subjects and before the testing of the

cognitive ability test was conducted, the subjects had

listened to the ‘Air on a G string; for 5 minutes with an

opened eye. This is to have the subjects maintain a steady

state before the external stimuli test is performed [16].

After a steady state, running the CNT program inducing

audiometry stimuli for 9 minutes measuring the EEG and

the attention con centration test was conducted.

2.3. Design of Fuzzy Inference System

It is known that each relationship between response rate

and concentration with the error rate and the concentra-

tion do exists. According to the study of Japan's National

Institute of Radiolog ical Science, results were drawn that

chewing gum can increase the concentration and during

such moment whether or not such concentration was

made were confirmed th rough the response speed.

And of computerized neurocognitive program having

low number of CPT in correct response and slow reaction

time, further determining the number of false alarm er-

rors and the number of omission errors, it was suggested

that abnormal cases had attention concentration problem.

However when comparing the each concentration test

result with the response speed and the number of errors,

and because there are many ambiguity to determine the

level of concentration, therefore using uncertain informa-

tion with set of membership fu nction to check th e degree

of the binary system logic of each object belongs or does

not belong to any gatherings, a fuzzy theory that articu-

lates mathematically was used.

The system to check the level of concentration is con-

trolled by the fuzzy theory. Two inputs are the response

speed and the number of errors and the output is the level

of concentration. Through fuzzification section, a fuzzy

set is structured having two inputs and one output. The

range of response speed is 0.6030 seconds which was the

fastest and 0.0776 seconds which was the slowest and

based on the maximum value of the number of errors it

has a range of 0 to 65. The range of response speed is

divided into 3 membership functions; fast, average and

slow. The number of errors is divided into 5 membership

functions; very low, low, average, many, extremely many.

Having 13 or less is regarded as very low errors and 65

or more is regarded as extremely many errors. The output

has 2 membership functions; have concentrated and not.

For each of the membership functions, it was able to se-

lect a range of distinctly different and range could be

divided differently. This, as with all design activities it is

based on the experience of designers and system re-

quirements.

The total number of the rules in the rule-base is the

value multiplied by the number of each set of input

variables. The response speed has 3 membership func-

tions and the number of errors has 5 so there are total of

15 rules.

Mamdani type fuzzy reasoning rules were used.

By taking the maximum value from the goodness of fit

calculated by the fuzzy arithmetic, found the output

member value and in order to find the exact value related

to the level of concentration it must use defuzzification.

The output values were computed based on the center

of gravity method. The centroid method is multiply the

only value and the corresponding value of the output

member, divided by the sum of the belonging value.

The system is designed by using MATLAB R2009a as

shown in Figure 2.

Using the self designed Fuzzy Inference System to

obtain concentration result and changes in the Theta

wave result (Table 2) from the audiometry test result

(Table 3) show that 15 out of 21 person’s data had

matched each other’s results (Table 4).

3. Conclusions

By using the response speed and the number of errors

obtained through the concentration test for children, to

find out whether or not the subject has concentrated or

not by using the results and the Theta wave, one of EGG

parameter, from a fuzzy system design. As a result 15 out

of 21 person’s data had matched each other’s results.

H. J. KIM, S. K. YOO 61

Figure 2. Fuzzy infe rence system.

Table 2. Change of theta wave.

Change of Theta wave

Name Before experiment After experiment

Subject 1 KSM 0.1323 0.1533

Subject 2 KYK 0.1072 0.1349

0ubject 3 KYJ1 0.0999 0.0999

Subject 4 KYJ2 0.095 0.0927

Subject 5 KYH 0.1115 0.1389

Subject 6 KYJ 0.1108 0.1077

Subject 7 REH 0.112 0.077

Subject 8 PAH 0.0956 0.1178

Subject 9 SHN 0.0898 0.083

Subject 10 LSM 0.1085 0.1262

Subject 11 LYH 0.1499 0.0965

Subject 12 LJJ 0.0993 0.083

Subject 13 JYJ 0.1209 0.1142

Subject 14 JMH 0.0858 0.102

Subject 15 JES 0.128 0.1298

Subject 16 JJB 0.1038 0.1199

Subject 17 JJ 0.1495 0.1429

Subject 18 JHG 0.0972 0.0995

Subject 19 CHR 0.1142 0.1226

Subject 20 HJW1 0.1144 0.0911

Subject21 HJW2 0.1359 0.1249

Table 3. CNT hearing test results.

CNT hearing test results

Name Reaction time The number of Er rors

Subject 1 K S M 0.615 12

Subject 2 KYK 0.684 41

Subject 3 KYJ1 0.627 44

Subject 4 KYJ2 0.776 27

Subject 5 KYH 0.665 16

Subject 6 KYJ 0.646 42

Subject 7 REH 0.655 41

Subject 8 PAH 0.622 24

Subject 9 SHN 0.76 9

Subject 10LSM 0.68 12

Subject 11LYH 0.714 31

Subject 12LJJ 0.662 39

Subject 13JY J 0.664 43

Subject 14JMH 0.603 64

Subject 15JES 0.672 7

Subject 16JJ B 0.706 46

Subject 17JJ 0.667 36

Subject 18JHG 0.713 24

Subject 19CHR 0.682 30

Subject 20HJW1 0.71 16

Subject21HJW2 0.672 28

Table 4. Comparison of results matching the results of fuzzy

Inference system and theta waves of change.

Name check the concentratio n l ev el

Subject 1 KSM O

Subject 2 KYK X

0ubject 3 KYJ1 O

Subject 4 KYJ2 O

Subject 5 KYH O

Subject 6 KYJ X

Subject 7 REH O

Subject 8 PAH O

Subject 9 SHN O

Subject 10 LSM O

Subject 11 LYH O

Subject 12 LJJ O

Subject 13 JYJ X

Subject 14 JMH X

Subject 15 JES O

Subject 16 JJB O

Subject 17 JJ O

Subject 18 JHG O

Subject 19 CHR O

Subject 20 HJW1 X

Subject21 HJW2 X

In order to secure more data from many subjects and

Genetic algorithm and more sophisticated fuzzy system

through a combination of different algorithms, future

research goal is to improve the performance of the sys-

tem presented in this study.

H. J. KIM, S. K. YOO

4. Acknowledgements

This work was supported by the National Research

Foundation of Korea (NRF) grant funded by the Korea

government(MEST) (No.2010-0026833)

REFERENCES

[1] Choi Chun Ok, “The Effects of Attention Concentration

Training Program on the ADHD Children’s Attention

Concentration and Impulsiveness,” Graduate School of

Education The Catholic University of Korea Seoul, Korea.

2000.

[2] P. Grossman, “Respiration, Stress and Cardiovascular

Function,” Psychology, Vol. 20, 1973, pp. 284-300.

[3] S. A. Kirk and J. C. Chalfant, “Academic and Develop-

mental Learning Disabilities: Attentional Disabili-

ties,” Columbus Ohio : Love Publishing Company, 1984.

[4] D. Reid and W. Hresko, “ A Cognitive Approach to

Learning Disabilities, New York : Mcgraw Hill, 1981.

[5] A. M. Wiliams and K. Davids, “Visual Search Strategy

Selective Attention, and Expertise in Soccer,” Research

Quartely of Exercise and Sport, Vol. 69, No. 2, 1998, pp.

111-1118. doi:10.1080/02701367.1998.10607677

[6] A. M. Wiliams, “Perceptual Skill in Soccer: Implications

for Talent Identification and Development,” Journal of

Sports Science, Vol. 18, 2000, pp. 737-750.

doi:10.1080/02640410050120113

[7] C. Christiansen, “Occupational Therapy: Intervention for

Life Performance,” In C. H. Christiansen, & C. M.

Baum(Eds), “Occupational Therapy : Overcoming Hu-

man Performance Deficits,” 1st edition, Thorofare, NJ:

SLACK Inc., 1991.

[8] E. J. S. Sonuga-Barke, K. Houlberg and M. Hall,

“When is ‘Impulsiveness’ Not Impulsive? The Case of

Hyperactive Children’s Cognitive Style,” Journal of child

psychology and psychiatry, Vol. 35, 1994, pp.1247-1253.

doi:10.1111/j.1469-7610.1994.tb01232.x

[9] R. A. Berkly, “Attention-Deficit/Hyperactivity Disorder:

A Handbook for Diagnosis and Treatment,” 2nd edition,

New York: Guilford, 1998.

[10] P. L. Frick, and B. B. Lancy, “The Nature and Character-

istics of Attention Deficit Hyperactivity Disorder,”

School Psychology Review, Vol. 20, 1991, pp.163-173.

[11] Park Ah Chung, Vol. 287, 1993, pp. 22-25.

[12] A. A. Rizzo and U. Buckwalter, “Psycho-Neuro-Physio-

Logical Assessment and Rehabilitation in Virtual Envi-

ronments: Cognition, Clinical, and Human Factors in

Advanced Human Computer Interaction,” Amsterdam:

IOS Press, 1997.

[13] B. H. Koo and D. S. Bai, “Diagnostic Availability of a

Computerized Neurocognitive Function Test for Children

with ADHD,” Journal of Korean Soc Biol Ther Psychia-

try, Vol. 12, No. 1, 2006, pp. 39-49.

[14] W. Klimesch, “EEG Alpha and Theta Oscillations Reflect

cognitive and Memory Performance: A Review and

Analysis,” Brain Research Reviews, Vol. 29, No. 2-3,

1999, pp. 169-195. doi:10.1016/S0165-0173(98)00056-3

[15] B. Anatoly, N. Meir, K. Boris David and a Potential

Method for the Evaluation of the Autonomic Nervous

system,” Physiological Measurement, Vol. 19, No. 1,

1998, pp. 93-102. doi:10.1088/0967-3334/19/1/008

[16] W.–J. Choi, C.-K. Lee and S.-K. Yoo, “An Analysis of

Correlation Between EEG Signal and HRV During At-

tentional Status with Children Under 15 years,” Koean

Journal of The Science of Emotion & Sensibility, Vol. 14,

No. 2, 2011, pp. 269-278.

[17] K. Jinsung, L. Jongbum, K. Wanseok, S. Hyounjin, B.

Daiseg and L. Hyelin “The Validity and Reliability of

Computerized Neuro Cognitive Function Test in the Ele-

mentary School Child,” Korean Psychosomatic society,

Vol. 11, No. 2, 2003, pp. 97-117.

[18] J.-M. Shim, H.-H. Kim and Y.-s. Lee, “Effects of Com-

puterized Neurocognitive Function Program Induced

Memory and Attention for Patients with Stroke,” J Kor

Soc Phys Ther, Vol. 19, No. 4, 2007, pp. 25-32.