Computational Intelligence in Electroencephalogram Analysis - Computational Intelligence in Biomedical Engineering - page 283

find features that could be used for recognition. Moreover, ANNs and GA
were implemented to find dominant features prior to this information being
used as a data source for an ANN.
6.2.2 Evoked Potentials
Evoked potentials (EVP), also referred to as evoked responses, record the
brain's responses to various stimuli, such as sound, touch, and light. These
tests often help in evaluating neurological conditions. For each test, electrodes
are attached to the patient's scalp and, in some cases, to the earlobes, neck,
shoulders, and back.
Auditory evoked potential . This test records the brain's electrical
potential through EEG signals in response to stimulation of the audi-
tory sensory mechanism (Hoppe et al., 2001; Davey et al., 2006). Such
tests have many clinical applications, such as obtaining an objective
audiogram for auditory organ evaluation (Hoppe et al., 2001). Pattern-
recognition techniques such as NNs can be applied to classify the fea-
ture vector extracted from the EEG signal.
Somatosensory evoked potential . The aim of this test is to monitor
through responses reflected in EEG signals when mild electrical cur-
rents are passed through body through electrodes. During the test,
electrodes are located on the scalp, neck, shoulders, wrist, and ankle
and applied electrical currents cause a mild twitch of the thumb or
big toe (Davey et al., 2006).
Visual evoked potential . This test involves EEG response analysis
while visual cues are presented, for example, the patient may be
instructed to sit in a chair and watch a television screen displaying
a checkerboard pattern while recording his/her brain EEG potentials
(Davey et al., 2006).
6.3 Brain-Computer Interfaces
A BCI uses noninvasively recorded EEG signals from the brain to analyze
an individual's intended action and then translates the information into an
external action. Such a system can provide disabled persons with a means
of communicating with the physical environment such as for controlling a
wheelchair, a computer mouse, and other everyday appliances. A BCI sys-
tem includes EEG signal acquisition, signal-processing, feature extraction and
selection, and finally, pattern recognition. Pattern recognition is utilized to
categorize or classify the intention from the signal characteristics embedded
within a segment of the EEG waveform. There has been much recent interest
in the implementation of BCI for a variety of tasks, and to help classify the