Initial Experimental EEG Signal Topography
Mahfuzah Mustafa
1
, Zunairah Hj. Murat
2
, and Mohd Nasir Taib
3
12
Faculty of Electrical Engineering, Universiti Teknologi MARA Malaysia, Shah Alam, Selangor.
3
Nondestructive Biomedical and Pharmaceutical Research Centre, Universiti Teknologi MARA Malaysia, Shah Alam, Selangor.
mahfuzah@ump.edu.my, zunairahh@yahoo.com, dr.nasir@ieee.org
Abstract—Recently computer hardware and signal processing
have made possible the use of Electroencephalographic (EEG) to
measure brainwaves which gives information about brain
activity. This paper presents an initial investigation of EEG
analysis using topography. The proposed method look at the
possibility of imaging the EEG signals to provide various colors
or contrast representing different amplitude measurements for
the delta, theta, alpha and beta bands. After that, interpretations
of the EEG signals could be carried out using color images rather
than amplitude.
Keywords- EEG; brain topography
I. INTRODUCTION
EEG is used to measure the electrical activity of the brain.
This activity is generated by billions of nerve cells called
neurons [1].The EEG waves are characterized by amplitude
(voltage) and frequency. The EEG wave band frequency is
divided into delta, theta, alpha and beta [2]. The frequency
varies in each band, delta range within 1 to 4 Hz, theta range
from 4 to 8 Hz, alpha range from 8 to13 Hz and beta range
from 13 to 20 Hz.
Brain topography may be used to map brain activity in
color, for example red color represents higher EEG amplitude,
while blue color represents low EEG amplitude. In [3], the
author described a method to visualize 2D and 3D topographic
brain mapping. The technique read the array of EEG data and
transforms the time domain to frequency domain and separates
to sub frequency bands such as delta, theta, alpha and beta.
Then, the array is converted to color.
Saad et al [4], used features obtained from time-frequency
domain electrocardiogram (ECG) spectrogram images
produced through Short Time Fourier Transform to classify
heart abnormalities. The features was fed into Multi-Layer,
Back-Propagation trained Artificial Neural Network in order
to classify heart abnormalities.
Various software or toolboxes are available to map the
brain image from EEG data. For example EEGLAB [5] which
can be embedded with Matlab and for processing continuous
and event-related EEG and MEG. The method in [6] used low
resolution brain electromagnetic tomography (LORETA) to
map the brain of mild Alzheimer patients. In [7], the author
using Brainstorm to visualized scalp distribution of EEG and
MEG signals.
In practice, EEG signals are in time domain data format.
The time domain plot presents a time-amplitude representation
of signal. However, frequency domain signal may produce
some extra information or unique features. For example, in
EEG signals we only concentrate from 1Hz to 40Hz which is
in frequency domain. Usually, Fourier Transform(FT) is used
to find frequency component in signals. The Fourier equation
as described in Equation 1.
...(1)
where, t stands for time, f stand for frequency, and x denotes
the signal. The Fourier Transform gives good frequency
information but hides timing information. To produce two
dimensional data from EEG signal we need time-frequency
information. Therefore, time-frequency analysis technique in
this work using Short Time Fourier Transform (STFT). The
STFT is described in Equation 2.
...(2)
where, is the signal, is the window function, and *
is the complex conjugate. The changes of signal that vary in
time are performed by STFT. The STFT is to perform a FT on
only a small section (window) of data at a time, thus mapping
the signal into a two-dimensional function of frequency and
time. This is clearly shown in Equation 2, the STFT of signal
is the FT of the signal mulplied by a window function.
In this paper, we describe two ways for EEG topography.
In the beginning of the experiment, we used Brainstorm
Matlab toolbox [8] to visualize EEG data into brain image.
Next, we used STFT to visualize EEG data into spectrogram.
II. METHODS
This project uses EEG data obtained from horizontal
rotation for brain wave balancing research [9]. In this project,
the EEG data were collected with scalp electrodes by using
standard gold disc electrodes, at a sampling rate of 256Hz
from the International 10-20 system. The “10” and “20” refer
to 10% and 20% inter-electrode distance [10]. EEG data
acquisition was performed in specialized room, where each
subject was comfortable in bed.
Proceedings of 2009 Student Conference on Research and Development (SCOReD 2009),
16-18 Nov. 2009, UPM Serdang, Malaysia
978-1-4244-5187-6/09/$26.00 ©2009 IEEE
157