Information Technology Reference
In-Depth Information
seizures using EEG signals. The selection of a
reliable detection method is crucial to the success
of CRESH for remote detection and prediction of
epileptic seizures.
There have been many researches on the
automated computational methods for detecting
seizures based on Electroencephalographic (EEG)
signals received from Epileptic patients' brains.
EEG signals are the most commonly used for the
prediction of coming epileptic seizures (Suresh
and Shankara, 2005). Human EEG signals are re-
flected by multiple ictal patterns in which epileptic
seizures typically appear in a clearly characterized
format, usually rhythmic signals, often come with
or even happen before the earliest observable
changes in behavior. The detection at the earliest
onset of ictal patterns in the EEG can be used
for diagnosis procedures during seizures and to
differentiate epileptic seizures from other similar
symptoms (Meier, Dittrich, Schulze-Bonhage, &
Aertsen, 2008).
EEG signals were discovered by Hans Berger
in 1923 as electrical activity from the cerebrum.
These signals have distinct waves of different
amplitudes and frequencies that can be used to
differentiate different processes such as “sleep,
rest, wakefulness, and pathologies”. EEG patterns
show standards of normality and deviations from
the standard express abnormality. However, there
was an obstacle in interpreting time-serial EEG
signals due to the lack of a proper model of the
central nervous system that is consistent with these
observed states of different processes (Hively and
Protopopescu, 2003).
Hjorth (1970) established three parameters
which can be used to describe and quantify EEG
signals in the spatio-temporal domain. These
parameters include
activity, mobility, and com-
pleteness
. Since then EEG signals have become
the most commonly used tools for clinical diag-
nostic. Figure 1 illustrates an eight-second sample
of EEG data.
Figure 1. An eight-second sample of EEG signal
Search WWH ::
Custom Search