[Eeglablist] Merging components time frequency

[Makoto Miyakoshi]

Dear Max (cc to Arno),

Arno, correct me if I'm wrong.

> I understand that activity is not comparable directly between
> components, so that the values found in EEG.icaact aren't comparable across
> components. Is that correct?

Incorrect. There is no problem at least for within subject comparison.
For between subject comparison, it is still comparable, unless you
have a serious doubt in making between-subject comparison with your
channel EEG data without individual normalization. Usually we don't do
this though. Do you agree?

> I reason that if my TF power data is baseline
> corrected (logtransformed and divided by baseline), then all the activity is
> in baseline "units", and so I should be allowed to simply average them. Is
> it correct?

Probably incorrect, because with the default parameters, it does not
normalize variance. However, this does not mean it prevents you from
performing between-subject comparison as I wrote above. For
normaliation, use the option 'basenorm', 'on' and you'll see std
instead of power in dB for ERSP.

> What if my data is not baseline corrected?

Again, baseline correction is just for mean subtraction and does not
normalize the data.

Makoto

2012/7/31 Maximilien Chaumon <maximilien.chaumon at gmail.com>:
> Hello eeglabers,
>
> I have a question about merging similar (clustered) components in an
> analysis. I understand that activity is not comparable directly between
> components, so that the values found in EEG.icaact aren't comparable across
> components. Is that correct?
> Then I have time frequency transforms from these components and would like
> to merge these somehow. I reason that if my TF power data is baseline
> corrected (logtransformed and divided by baseline), then all the activity is
> in baseline "units", and so I should be allowed to simply average them. Is
> it correct? What if my data is not baseline corrected?
>
> Note that at this moment I'm not using the STUDY features to look at the TF
> data so the time frequency data that I use is the output timefreq.m
> directly.
>
> I would like to be sure that I'm not doing things wrong.
>
> Thank you!
> Max
>
> _______________________________________________
> 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



-- 
Makoto Miyakoshi
JSPS Postdoctral Fellow for Research Abroad
Swartz Center for Computational Neuroscience
Institute for Neural Computation, University of California San Diego