[Eeglablist] Merging ICA components

[Makoto Miyakoshi]

Dear Jason,

> Is there a way to 1) merge components

No, I have never heard of a solution.

> or 2) overcome this problem in a different manner?

Unfortunately I do not know any effective workaround.

Basically, if ICA does not return what you want to study, then you can't
study it. That is a limitation of this approach, at least as far as I know.

Makoto

On Wed, Jan 20, 2016 at 3:33 AM, Jenson, Dave <djenson1 at uthsc.edu> wrote:

> All,
>
> Does anyone know how to merge components from an ICA decomposition into a
> single component?  I am running a study on the sensorimotor mu rhythm, and
> I have a few subjects for whom ICA produces multiple potential components.
> At the moment I am choosing the best component per subject based on scalp
> map, spectra, and dipole location.  However, I still have a few subjects
> for whom there is no clear winner.  I don't want to lose statistical power
> by excluding a component that represents the sensorimotor mu, but also
> don't want to disproportionately weight some subjects over others at the
> study level by including multiple components.  Is there a way to 1) merge
> components or 2) overcome this problem in a different manner?
>
> -Dave Jenson
>
>
> _______________________________________________
> 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
Swartz Center for Computational Neuroscience
Institute for Neural Computation, University of California San Diego
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20160317/de1e1c47/attachment.html>