Epilepsy Detection and Monitoring - Quantitative EEG Analysis Methods and Clinical Applications - page 145

In considering epilepsies as dynamic diseases of brain systems, Lopes da Silva
and colleagues proposed two scenarios of how a seizure could evolve [11]. The first
is that a seizure could be caused by a sudden and abrupt state transition, in which
case it would not be preceded by detectable dynamic changes in the EEG. Such a sce-
nario would be conceivable for the initiation of seizures in primary generalized epi-
lepsy. Alternatively, this transition could be a gradual change or a cascade of
changes in dynamics, which could in theory be detected and even anticipated.
In the sections that follow, we use these basic concepts of brain dynamics and
review the state-of-the-art seizure detection and seizure prediction methodologies
and give examples using real data from human and rat epileptic time series.

6.3 Seizure Detection and Prediction

The majority of the current state-of-the-art techniques used to detect or predict an
epileptic seizure involve linearly or nonlinearly transforming the signal using one of
several mathematical black boxes, and subsequently trying to predict or detect the
seizure based off the output of the black box. These black boxes include some
purely mathematical transformations, such as the Fourier transform, or some class
of machine learning techniques, such as artificial neural networks, or some combi-
nation of the two. In this section, we review some of the techniques for detection
and prediction of seizures that have been reported in the literature.
Many techniques have been used in an attempt to detect epileptic seizures in
the EEG. Historically, a visual confirmation was used to detect seizures. The onset
and duration of a seizure could be identified on the EEG by a qualified technician.
Figure 6.1 is an example of a typical spontaneous seizure in a laboratory animal
model. Recently much research has been put into trying to predict or detect a seizure
based off the EEG. The majority of these techniques use some kind of time-series
analysis method to detect seizures offline. Time-series analysis of an EEG in general
falls under one of the following two groups:

0 ~ 30 seconds

30 ~ 60 seconds

Seizure onset

60 ~ 90 seconds

90 ~ 120 seconds

120 ~ 150 seconds

1000 μ

1s

150 ~ 180 seconds

Figure 6.1 Three minutes of EEG (demonstrated by six sequential 30-second segments) data
recorded from the left hippocampus, showing a sample seizure from an epileptic rat.