Biomedical Engineering Reference
In-Depth Information
present in the common average approach or Hjorth method introduce additional cor-
relations between signals, which disturbs the original correlation structure.
4.1.4 MEG measurement, sensor systems
The MEG measurement is technically much more demanding than the EEG mea-
surement. This stems mainly from the fact that the magnetic fields generated by
the brain are on the order of tenths to hundreds of femtotesla, which is orders of
magnitude weaker than the Earth's magnetic field and fields generated by, e.g., mov-
ing ferromagnetic objects in the environment. Partially the environmental artifacts
are removed by special spatial setup of measurement coils and partially by special
magnetic shielded chambers (walls made of a few layers of μ -metal with high mag-
netic permeability separated by pure highly conducting aluminum layers allowing
for generation of Eddy currents). The practical measurement of the extremely weak
fields of MEG is possible thanks to the superconducting quantum interference de-
vice (SQUID) which uses quantum effects in superconducting electrical circuits. The
SQUIDs are magnetically coupled to the sensor coils, which in principle can be of
three types: magnetometer, axial gradiometer, or planar gradiometer ( Figure 4.5) . In
a gradiometer inverse winding of the sensor coil eliminates the external magnetic
fields which are slowly changing in space.
The sensor configuration of modern whole-head MEG systems consists either of
magnetometers, or axial or planar gradiometers, or even a combination of them, cov-
ering the whole head in a helmet-like fashion.
Magnetometers are superior to gradiometers with regard to the possibility of
recording signals from deeper sources. Apart from their different sensitivity to noise,
these sensor configurations differ in the location of the signal's amplitude extremum
with respect to its source. Unlike planar gradiometers for which the maximum signal
is right above the source, magnetometers (as well as axial gradiometers) show ex-
trema on both sides of the underlying source [Hamalainen et al., 1993]. In the case
of signal analysis on the level of sensors, planar gradiometer representation is easier
in interpretation than that of axial gradiometers or magnetometers, since the maxi-
mum of the averaged signal directly indicates the location of the activated region.
4.1.5 Elimination of artifacts
An artifact in EEG can be defined as any potential difference due to an extra-
cerebral source. The careful identification and elimination of artifacts is of utmost
importance for EEG analysis. We can distinguish the artifacts of technical and bio-
logical origin. The first ones can be connected with: power supply, spurious electrical
noise from engines, elevators, tramway traction, etc., bad electrode contact, or its de-
tachment. Another kind of artifact has its origin within the subject's body. The most
common of them are connected with eye blinks, eye movements, muscle activity,
electrocardiogram (ECG). Eye movement generates an electric signal called an elec-
trooculogram (EOG) because of potential difference between cornea and the back
of the eye. Body and head movements may induce not only muscle electrical activ-
 
Search WWH ::




Custom Search