Biomedical Engineering Reference
In-Depth Information
human head. However, because it is limited to only 21 scalp locations, alternatives
that providing for a larger number of channels have been proposed. In 1985, the
10-10 system for the placement of up to 74 electrodes was proposed [11].
Oostenveld and Praamstra [12] defined the 10-5 system to further promote the stan-
dardization of electrodes in high-resolution EEG studies. In the 10-5 system, a
nomenclature and coordinates for up to 345 locations are defined. The system pro-
vides great flexibility, because it allows the selection of a subset of homogeneously
distributed positions. The interelectrode distance (on a standard head with 58-cm
circumference) is typically between 53 and 74 mm for the 10-20 system, and
between 28 and 38 mm for a 61-channel montage following the 10-10 system. For a
homogenous 128-channel layout based on the 10-5 system, the interelectrode dis-
tance would further decrease to approximately 22 to 31 mm. Unfortunately,
because both the 10-10 and the 10-5 system are based on the original 10-20 system,
none of these systems features equal distances between electrodes. In addition to the
matter of interelectrode distance, an important issue is the distribution of spatial
sampling. If one simplifies the head as a sphere, the original 10-20 system spatially
covers only little more than half of the sphere. In contrast, both the 10-10 and 10-5
system extend the spatial coverage to approximately 64%. Both of these issues, a
sufficient electrode density and a maximum coverage of the head sphere, can be
considered beneficial for extracting spatial information from an EEG [13];
therefore, many
EEG laboratories and some manufacturers have
developed
equidistant and spatially extended channel montages.
Figure 2.5 shows three different electrode cap systems. Note the different
interelectrode distance between the caps as well as the different spatial sampling.
The equidistant 68-channel customized cap system features a relatively large
interelectrode distance of approximately 38 mm, but covers about 75% of the head
sphere and therefore provides a good basis for accurate source localizations [14].
Figure 2.5(c) shows a customized 256-channel cap developed at the Swartz Center
for Computational Neuroscience (San Diego, California). This cap provides both a
very dense array with 25 mm of distance between electrodes and a significantly
extended spatial sampling. In contrast to usual recording traditions, EEG signals
are recorded from the face as well, which can provide important additional infor-
mation [15].
To conclude, with the advent of multichannel EEG recordings, the choice and
design of the electrode cap used is a matter of great importance. Caps that extend
beyond the traditional 10-20 range can provide significant benefits, among them a
better and more comfortable fit, a more evenly distributed weight of electrode
cables, and, most importantly, more accurate spatial sampling of the scalp recorded
EEG. These are only some of the benefits of modern electrode caps, and we expect
further improvements to become commercially available over the next few years.
2.2.3 EEG Signal and Amplifier Characteristics
Probably the most important component in optimizing the SNR of EEG signals is
the amplification circuitry itself. It is here that noise present in the data can be
reduced or eliminated, and much of the reliability and validity of day-to-day EEG
research depends on the quality of the amplifier design used.
Search WWH ::
Custom Search