ANALYSIS OF EEG SIGNAL FOR USING IN BIOMETRICAL
CLASSIFICATION
Roman Zak, Jaromir Svejda, Roman Senkerik and Roman Jasek
Department of Informatics and Artificial Intelligence
Tomas Bata University in Zlin
Nam T.G. Masaryka 5555, 760 01 Zlin, Czech Republic.
E-mail: {rzak, svejda, senkerik, jasek}@fai.utb.cz
KEYWORDS
Brain Computer Interface, EEG, Signal processing,
Neural network classification,
ABSTRACT
Aim of this article is to clarify the potential use of EEG
signal in modern information age. The basic principle of
Brain Computer Interface (BCI) lies in the connection
of brain waves with output device through some
interface. BCI technology represents a communication
interface between brain and computer. To sense electric
signal from the brain, it is usually used an equipment
based on the last results of scientific research on neuro-
technology. Communication is provided by wireless
transmission through which the signal is transmitted
from the equipment to personal computer. Then the
signal is analysed, processed and used for finding
appropriate classification parameters.
INTRODUCTION
Many scientific disciplines deal with the human brain;
for example numerical neuroscience, neuro-informatics,
informatics or medicine. All of them bring theories,
which could explain different brain activities.
Numerical neuroscience provides mathematical and
biophysical models, which are able to model basic
processes in neurons and neural networks. The main
goal of neuro-informatics is systematic development of
database intended to collect information such as brain
morphology, brain parts anatomy and their functional
connection, brain electrophysiology, brain states
obtained with magnetic resonance and their integration.
Further, it seeks to develop tools for modelling, where
the aim is the most accurate emulation of brain activity.
In Informatics, complex networks are highly suitable to
model a complex system among which the brain
includes. The contribution of medicine is undisputable
especially in brain anatomy research.
The human brain is a complex system, which is an
object of our research. It is regarded as the most
complex system in the universe. The modern science is
currently attempting to understand the complex
interconnection among individual parts of the brain
(Sporns et.al. 2005). There are many publications,
which deal with description of the brain (Adeli 2010;
Damasio 1995; Sporns et al. 2005).
Currently there are many known applications of BCI
technology, but not enough at each particular field of
study. Signal that is sensed from the brain is the key
element in the BCI model; therefore the design of an
appropriate algorithm for processing of the signal is the
most discussed part of BCI model structure (Schalk et
al. 2004).
Invasive methods of sensing the brain activity could
provide very accurate data, but it is not both technically
and user friendly; thus, it would not be further
mentioned in this article. On the other hand, more
accessible non - invasive methods can obtain relatively
weak signal with amplitude ranging from units to
hundreds of microvolts. Moreover, the signal is also
prone to noise elements. Another disadvantage of this
method is a summation of neuron signals; thus, obtained
data are referenced to a specific group of neurons. The
brain itself is composed of several parts, without which
his activity could not be possible. One of its basic
structural parts is a neuron. The neuronal cells are
characterized by the fact that electrical activity is
carried out in them. These cells communicate with each
other by electrical signals. According to the last
estimate, there are approximately 10
11
neurons in the
brain. Every one of them is connected with thousands of
other neurons. The main source of EEG signal is an
electric activity of synapse - dendrites membrane
located in the surface layer of the cortex. Each active
synapse dispatches electromagnetic pulse to the
environment during excitation. Due to the high number
of these pulses, it is difficult to locate their source by
means of multichannel sensor on the skin. This issue
could be compared to full amphitheatre, in which there
are chanting people and the task is to recognize from
outside, which specific group of fans shouts. A different
perspective on this issue may be such that the aim is to
identify uniqueness of the signal for each individual
subject. In the example shown above, it is as we would
like to recognize the type of the stadium by the mass of
chanting people. For example, there is noticeable
difference between hockey and tennis fans. The
biometric signatures are different for each creature on
the planet Earth.
METHODS
There are several approaches for sensing brain activity.
The most widely used is EEG technology, which
Proceedings 28th European Conference on Modelling and
Simulation ©ECMS Flaminio Squazzoni, Fabio Baronio,
Claudia Archetti, Marco Castellani (Editors)
ISBN: 978-0-9564944-8-1 / ISBN: 978-0-9564944-9-8 (CD)