Information Technology Reference
In-Depth Information
Fig. 8.6
Grand average (
N
= 19) of the ERS generated by the Ne source a and by the theta source
b for error trials
hereafter we will refer to as the
; although differences in amplitude
exist also for this latter source, they are not signi
“
theta source
”
cant.
On the other hand, the theta source power increase was significant in frequency
band-pass region [5 Hz 8 Hz] for time window [300 ms 600 ms] (
p
< 0.01). The
ERS generated by this source is shown in Fig.
8.6
b. In Fig.
8.6
a, it is displayed the
same ERS when computed using the spatial
filter of the Ne source instead; the ERS
in this case disappears. These results suggest that the Ne source and the theta source
correspond to separate phenomena generated by different brain structures with
different dynamics. The source responsible for the ERS (theta source) appears more
speci
c.
We can now illustrate the advantage brought upon from the BSS analysis with
these data. Compare Fig.
8.5
ato
8.3
and Fig.
8.6
to
8.4
.
Although in both cases,
results in the sensor space are computed for the optimal cluster of electrodes, in
both cases, it is clear that working in the source space allows a better sensitivity
and speci
city: In both cases, the difference between the error and correct trials is
highlighted.
8.9.4 Source Localization
The BSS source responsible for the ERP (Ne source) difference between correct and
error trials was localized by sLORETA in the anterior cingulate gyrus (BA 24). The
BSS source responsible for the ERS (theta source) was localized close to the sup-
plementary motor area (BA 6) (Fig.
8.7
). Keeping in mind the approximation of a
source localizationmethod applied on a standard headmodel, these anatomical results
are in line with results reported by previous studies (Gehring and Willoughby
2002
;
Herrmann et al.
2004
; Nieuwenhuis et al.
2003
).
