[FieldTrip] Why does applying fastica change the form of EEG data?

Elif Tülay e.eliftulay at gmail.com
Fri Aug 5 12:34:50 CEST 2022

Dear Fieldtrippers,

Now, I work on EEG data which is exported from EEGlab as segmented data. I
try to remove eye blinks from the data and I wonder something about
ft_componentanalysis function. When I performed fastica to remove the eye
blink component, I realized that it changes the form of EEG in addition to
removing the related component (please see the attached files, before and
after; scales are the same). I cannot understand the reason. It is not
because demean application, I checked it.

Do you have any other idea? By the way, I used averaged referencing on
EEGlab. Also please find the script that I used below.

cfg            = [];

cfg.method     = 'fastica';

cfg.channel       = {'all', '-VEOG'};

comp           = ft_componentanalysis(cfg, data);

cfg = [];

cfg.component = [5];

data = ft_rejectcomponent(cfg, comp, data)

Also I wonder one more thing, what is the mission of cfg.updatesens
parameter in ft_ componentanalysis function?

Many thanks in advance,
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20220805/86a934ed/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: after_fastica.jpg
Type: image/jpeg
Size: 174680 bytes
Desc: not available
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20220805/86a934ed/attachment-0002.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: before_fastica.jpg
Type: image/jpeg
Size: 179226 bytes
Desc: not available
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20220805/86a934ed/attachment-0003.jpg>

More information about the fieldtrip mailing list