Information Technology Reference
In-Depth Information
trials as compared to correct trials in 80 % of the subjects (all t tests with
p
< 0.05).
The maximum number of digits to memorize for each subject was highly variable,
ranging from 4 to 10, with mean (sd) = 6.5 (1.37). These results demonstrate that
our presentation software succeeded in equalizing the cognitive load across sub-
jects, despite the great intersubject variability of digit memory span.
8.9.2 Sensor Space Analysis
The ERP in the error trials differed from the correct trials in three time windows
with different timing and/or electrode location (Fig.
8.3
). A signi
cant positivity for
errors was found at time window [320 ms 400 ms] at electrode Cz (p < 0.01), a
signi
cant negativity for errors at time window [450 ms 550 ms] at clustered
electrodes Fz, FCz, Cz (p < 0.01), and a signi
cant positivity for errors at time
[650 ms 775 ms] at clustered electrodes Fz, FCz (
p
= 0.025).
An ERS (power increase as compared to baseline) could be seen in the theta
band in both correct and error feedback at fronto-midline locations. This syn-
chronization unfolds from around 250
600 ms poststimulus. In some subject, it
goes up to more than 200 % of power increase for error trials. Albeit present in both
conditions, this ERS is signi
-
cantly more intense for error trials as compared to
correct ones (Fig.
8.4
) in the frequency band-pass region 5
8 Hz and time window
[350 ms 600 ms] poststimulus over the clustered electrodes Fz and FCz (
p
= 0.015).
-
Fig. 8.3
a Grand-average (
N
= 19) ERP for correct (
pointed line
) and error (
solid line
) trials.
Time windows where the difference in amplitude between the two conditions is significant (
gray
panels
) and b scalp topographies of
t
values computed within the three signi
cant windows.
White
disks
show the signi
cant clustered electrodes
