Biomedical Engineering Reference
In-Depth Information
2
4
6
8
10
12
14
16
18
20
2
4
6
8
10
12
14
16
18
20
FIGURE 4.6:
Examples of different kinds of artifacts. Upper panel, 20 s epoch of
EEG contaminated with ECG artifact around seconds 1 and 3, and the eye move-
ment artifact around second 16. Lower panel, muscle and movement artifact causing
overflow around second 13.
[Gevins and Morgan, 1986, Mammone and Morabito, 2008] or more subjects [van de
Velde et al., 1998, Moretti et al., 2003, LeVan et al., 2006, Jiang et al., 2007]. The
methods aiming at elimination of different kinds of artifacts in the framework of the
same system are relatively few. Among them we may distinguish semi-automatic
[Nakamura et al., 1996, Schlogl et al., 1999, Delorme and Makeig, 2004] and auto-
matic methods [Durka et al., 2003, Klekowicz et al., 2009].
The different approaches to identification of artifacts are described in the review
by [Anderer et al., 1999]. The simplest methods were based on amplitude thresh-
olds (overflow check), with different scenarios for setting the thresholds. Further
works used time-varying autoregressive modeling and slope detection [van de Velde
et al., 1999], wavelets [Jiang et al., 2007], and also artificial neural networks [Schal-
tenbrand et al., 1993]. There were also attempts to separate the artifacts based on
the modeling of the sources of extra-cerebral activity. These methods usually mod-
eled EOG generators as dipoles and estimated the influence of their activity on EEG.
Search WWH ::
Custom Search