[Eeglablist] ICA/PCA removal of components filtered at different HZ

[Makoto Miyakoshi]

Dear Matthew,

Method 2 is correct.

> If one or both of the methods above are correct would you happen to know
why I am getting greatly attuned EEG waves, reduced FFT power across all
frequencies, and the appearance of large EEG waves induced by the component
removal?

If you are asking 'attenuated EEG waves' on scalp electrodes, that is
because you rejected IC1 which is mostly EOG according to your report.

> reduced FFT power across all frequencies,

Because the rejected component had broadband power--you can see it in the
original IC1 power spectral density plot.

> and the appearance of large EEG waves induced by the component removal?

Check the amplitude scale before and after the IC rejection. You should
find a smaller amplitude scale after rejection.

Makoto


On Wed, Apr 22, 2020 at 7:59 AM Gunn, Matthew P <mathewgunn1 at siu.edu> wrote:

> Hello,
>
> I had a question regarding the "What happens to the < 1 Hz data if ICA is
> calculated on > 1 Hz data and applied to 0.1 Hz data? (08/17/2018 Updated)"
> section on the EEGLAB wiki.
>
> Are the following methods the correct way of doing this?
>
> Set up: I have the same file filter at 1hz and 0.1hz. Also, component 1 is
> full of Eyeblinks.
>
> Method 1:
> 1) Run an ICA on the 1hz file
> 2) Extract component 1 and insert into file filtered 0.1hz (This is done
> by taking every ICA related variable in the EEG (EEG.icaact, EEG.icawinv,
> EEG.icaweights, etc)  for the 1hz file and copying them into the 0.1 Hz
> data)
> 3) Remove component 1 from the 0.1hz data
>
> Method 2:
> 1) Run an ICA on the 1hz file
> 2) With the 0.1 Hz file selected go to Edit->Dataset Info->  select "From
> other dataset" and set all options to the 1hz file.
> 3) Remove component 1
>
>
> If one or both of the methods above are correct would you happen to know
> why I am getting greatly attuned EEG waves, reduced FFT power across all
> frequencies, and the appearance of large EEG waves induced by the component
> removal?
>
> Thank you for the wiki and your time.
>
> Matt Gunn
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>