Comparison between Characteristics of EEG Signal Generated from Dyslexic and Normal Children

C.W.N.F. Che Wan Fadzal1, W. Mansor1, Khuan Y. Lee1, S. Mohamad2, N. Mohamad1, S. Amirin3

1Faculty of Electrical Engineering Universiti Teknologi MARA,

40450 Shah Alam, Selangor, Malaysia

2Universiti Kuala Lumpur British Malaysian Institute, Malaysia 3Dyslexia Association Centre, Malaysia

wahidah231@salam.uitm.edu.my

Abstract—Electroencephalogram (EEG) is one of the methods to detect dyslexia in children. Dyslexia has to be detected at an early stage to help the children to excel in their study and later be successful in life. In this study, the EEG signals generated from dyslexic and normal children during relax and writing words were processed, analysed and compared. Four electrodes; C3, C4, P3 and P4 were used in the recording of the EEG signals. The recorded EEG signals were filtered using a band pass filter with frequency range of 8 - 30 Hz. The signal was then analyzed using Fast Fourier Transform. Analysis of EEG signals showed that the range of frequency of EEG signals during writing for dyslexic was greater than that of normal children for each electrode placement at beta sub band frequency. The range of frequency of EEG signals for dyslexics is 22 – 28 Hz whereas for normal children is 14 – 22 Hz.

Keywords-component; Electroencephalogram; dyslexia, writing, Fast Fourier Transform

I. INTRODUCTION Dyslexic children encounter several difficulties throughout

the learning process in one or more activities such as reading, spelling and writing. The weaknesses may fall into the category of speed of processing, short- term memory, sequencing and organization, auditory and/or visual perception, spoken language and motor skills. They have difficulties in mastering and using written language, which may include alphabetic, numeric, and musical notation [1].

There are two areas in human brain called left and right hemispheres. The left hemisphere is an area where logical thinking, abstract analysis, and verbal dexterity are localized. It is also associated with language processing. The right hemisphere represents localization of awareness moments since sensory information resides here. The normal process of reading takes parts in the left hemisphere more than in the right hemisphere. The pathway of reading for dyslexics is different from non-dyslexics in the language processing centre [2]. Most researchers believe that dyslexia results from the impairment in the language processing centre which is located at the left hemisphere in the brain [2]. However, S.Casarotto et al[3] reported that there were abnormalities in the right hemisphere in language-impaired subjects which indicate that dyslexia is

spread disorder, the impairment is not limited to the classical brain area that normally involved in linguistic abilities [3].

The activities in the brain can be extracted from electroencephalogram (EEG) by processing and analyzing the signal. Electrode setting called 10-20 system is usually used to record the EEG signal. This system was designed by dividing the head into proportional distances from prominent skull landmark (nasion and inion). The electrode placement is labeled as F(frontal), C (central), T (temporal) P (posterior) and O (occipital). Each position indicates specific activity. Some researchers found that F7 was for rational activities, F8 for emotional impulses, C3, C4, Cz for sensory and motor functions, P3, P4, Pz for perception and differentiation, T for memory function and finally, O for primary visual area [4, 5].

Even though a few researches have been carried out to study the EEG signals from dyslexic patients, large number of electrodes and bipolar connections were used in the recording of the signals [6]. Furthermore, most researchers focused on reading disabilities in dyslexic children [7]. Dyslexic children who have reading difficulties also have writing problems. This paper describes the comparison of EEG signal characteristics obtained from dyslexic and normal children during writing. The EEG signals was analysed and compared in terms of its frequency pattern and time domain representation.

II. METHODOLOGY

A. Data Acquisition and Writing Activity Three Right-handed dyslexic children and three normal

children, aged between 8 – 12 years old took part in this study as subjects. Before recording commenced, each subject was seated comfortably on a chair. Six tasks were given to the subjects to be carried out. The tasks include relax, recognizing alphabets, sounding out alphabets, writing alphabets, spelling words and writing words. The tasks were designed based on the most popular and successful technique used in Malaysia to diagnose dyslexia [8]. In each task, the subject has to relax before performing the next tasks. The word to be written depends on level of assessment and subject ability. The assessments are divided into three levels; level 1for subjects aged 7 to 8, level 2 for aged 9 to 10 and level 3 for aged 11 to

978-1-4673-1666-8/12/$31.00 ©2012 IEEE

2012 IEEE EMBS International Conference on Biomedical Engineering and Sciences | Langkawi | 17th - 19th December 2012

843

Figure 2: EEG signal of dyslexic children during writing

Figure 3: EEG signal of normal children during writing

12. Examples of written word used for level 1 is ‘box’, for level 2 is ‘chair’ and for level 3 is ‘friend’. Even though six tasks were given to the subjects, this study only focuses on writing word for all levels.

In the first activity, subjects were asked to relax their mind, arms and hands for about 120 seconds. They were then asked to perform writing activities which took about 30 seconds. Each subject writes 6 different words and repeats writing each word three times. They take 30 seconds to complete one word. Therefore, for writing, there are 18 data for each subject. Here, Ag/AgCI electrodes were used and placed on the scalp based on International 10-20 Electrode Placement system to record the EEG signals. The EEG signals were recorded through electrode C3, C4, P3 and P4 using g.MOBIlab hardware with 8 monopolar channels. These electrode locations were selected since C3 and C4 deal with sensory and motor function meanwhile P3 and P4 are for recording perception and differentiation activities [9]. In addition, C3 and C4 can also be used to record the right and left finger movement related signals for BCI applications [3]. The reference electrode was placed on the right ear whereas the grounding electrode was on the forehead (Fpz). Figure 1 below shows one of the dyslexic children performed writing activity during EEG signal recording.

Figure 1: Writing activity performed during EEG signal recording

B. EEG Signal Processing and Analysis In order to examine the frequency of the signals generated

from writing words, frequency analysis was performed. The signal was first filtered to remove artifacts such as Electrocardiograms (ECGs), electroocculograms (EOG), eye blinks and others. A band pass FIR filter with frequency range between 8-30 Hz was used to extract the EEG signal that contains information on relax and writing. After filtering, Fast Fourier Transform was performed to analyze the frequency content of the signals. The frequency range of the signals was then measured.

III. RESULTS AND DISCUSSIONS

Figure 2 and 3 show the EEG signals of dyslexic and normal children. Note that EEG signal of dyslexic children has larger maximum amplitude compared to that of normal children.

2012 IEEE EMBS International Conference on Biomedical Engineering and Sciences | Langkawi | 17th - 19th December 2012

844

Figure 4: Frequency spectrum of EEG signal during relax for dyslexic

children

Figure 5: EEG spectrum of EEG signal during relax for normal children

Figure 6: Frequency spectrum of EEG signal during writing for dyslexic

children

Figure 7: Frequency spectrum of EEG signal during writing words for

normal children

Figure 4 and Figure 5 show the frequency of EEG signals obtained from each electrode placement; C3, C4, P3 and P4 for dyslexic and normal children during relax. It is obvious that the frequency pattern of EEG signal from dyslexic children in relax condition consists of two peaks compared to that of normal children. In the frequency pattern of EEG signal from dyslexic children during writing, the second peak can be seen

clearly (see Figure 6 and Figure 7). The frequency spectrum of EEG signal from dyslexic children have greater energy in beta band, range from 22 -28 Hz meanwhile for normal children, the energy is more focused in the beta sub band that is from 14 to 22 Hz

Table I and Table II show the frequency ranges of EEG signals obtained from each electrode placement, C3, C4, P3 and P4 for dyslexic and normal children during relax and writing respectively. The frequency ranges of EEG signal for dyslexic children are larger than those of normal children. For dyslexic children during relax, electrodes C3 and P3 produce EEG signals with frequency range of 9 to10 Hz meanwhile for electrodes C4 and P4, the frequency range is from 10 -12 Hz. For normal children, all electrode give EEG signal with frequency range of 9-10 Hz. During writing activities, the frequency range of EEG signal for dyslexic children in the beta band is higher than that of normal children. Alpha sub band give the same frequency range meanwhile beta sub band give prominent differences from those of normal children. The results obtained in this study are in agreement with Bernard Sklar et al [10]. They have reported that the EEG spectra of dyslexic children showed significant differences from that of normal children.

2012 IEEE EMBS International Conference on Biomedical Engineering and Sciences | Langkawi | 17th - 19th December 2012

845

TABLE 1: FREQUENCY FREQUENCY RANGE (HZ) OF EEG FOR RELAX ACTIVITIES

Electrode Dyslexic Children Normal Children

C3 9-12 9-10 C4 10 -12 9-10 P3 9 - 12 9-10

P4 10 - 12 9-10

TABLE II: FREQUENCY FREQUENCY RANGE (HZ) OF EEG FOR WRITING ACTIVITIES

Electrode Dyslexic Children Normal Children

Alpha Sub Band

Beta Sub Band

Alpha Sub Band

Beta Sub Band

C3 9-10 23-27 9-10 15-22 C4 9-10 22-27 9-10 15-20 P3 9-10 23-26 9-10 14-18 P4 9-10 22-28 9-10 14-20

IV. CONCLUSIONS The analysis of EEG signals produced from dyslexic and

normal children based on writing some words has been discussed in this paper. Four electrodes were used to acquire the EEG signals. It was found that the frequency spectrum pattern and frequency ranges of EEG signals obtained during relax and writing for dyslexic and normal children are difference. Dyslexic children produce higher frequency of EEG signal either during relax or writing activities compared to normal children. The frequency range of EEG signal for normal children is 14 -22 Hz whereas for dyslexic children is 22 – 28Hz.

ACKNOWLEDGMENT This work was supported by Exploratory Research Grant

Scheme (ERGS), Malaysia (600-RMI/ERGS 5/3/ (19/2011)). The authors would like to express their gratitude to Ministry of Higher Education, Malaysia and Universiti Teknologi Mara, Malaysia, for financial support and providing equipment for this research and Dyslexia Association Centre, Malaysia for their support in data collection and suggestions.

REFERENCES [1] Ministry of Education. "Literature Review: An International Perspective

on Dyslexia," New Zealand Government, 2007. [2] C. W. N. F. C. W. Fadzal, W. Mansor, L. Y. Khuan, “Review of Brain

Computer Interface Application in Diagnosing Dyslexia” IEEE Control and System Graduate Research Colloquium, pp. 124-128, 2011.

[3] S.Casarotto,S.Cerutti,A.M.Bianchi,G.A.Chiarenza,” Dynamic Time Warping in the Study of ERPs in dyslexic children”, Proceeding of the 25th Annual International Conference of the IEEE EMBS,2003.

[4] M. Teplan, “Fundamental of EEG measurement”, Measurement science review, Vol.2, 2002.

[5] C.W.N.F. Che Wan Fadzal, W. Mansor, L. Y. Khuan, “An Analysis of EEG Signal Generated From Grasping and Writing”, In Proceedings of 2011 IEEE Conference on Computer Application & Industrial Electronics , pp. 535-537, 2011.

[6] C.W. Anderson, M. J. Kirby,” EEG Subspace Representations and Feature Selection for Brain-Computer Interfaces”, in Proc. 1st IEEE Conference on Computer Vision and Pattern Recognition Workshop for Human Computer Interaction (CVPRHCI'03), vol. 5, Madison, Wis, USA, June 2003.

[7] M.M. Showman,M.Ahissar, “Isolating The Impact Of Visual Perception On Dyslexics’ Reading Ability”, Vision Research, vol. 46, pp. 3514–3525, 2006.

[8] M.H.L.Abdullah,S.Hisham,S.Parumo “MyLexics: An Assistive Courseware for Dyslexic Children to Learn Basic Malay Language”, Sigaccess newsletter, Issue 95, 2009.

[9] S.Sanei,J.A.Chambers,”EEG Signal Processing”,John Wiley & Sons Ltd,England,2007.

[10] B.Sklar,J.Hanley,W.W.Simmons,” A Computer Analysis of EEG Spectral Signatures from Normal and Dyslexic Children”,IEEE Transaction on Biomedical Engineering,Vol.BME-20,No.1, Jan, 1973.

2012 IEEE EMBS International Conference on Biomedical Engineering and Sciences | Langkawi | 17th - 19th December 2012

846