[Eeglablist] 'ncomps' option for runica

[Scott Makeig]

The 'ncomps' option does not work as originally thought - instead use 
'pca',N where N is the number of components desired. Note. however, that 
minimal use of PCA is preferred (depending on length of training data 
available, mainly).

Scott Makeig


Peter Bachman wrote:

>Hi everyone,
>
>I am interested in running an ICA decomposition on a 128-channel set of
>epoched EEG data, specifying the number of components I would like the
>solution to include (I can run the decomposition successfully to produce
>128 components - I would like to request fewer).
>
>Using the pop_runica GUI, I have attempted to use the ncomps option (and
>have reattempted to use ncomps using every modification of the syntax I
>could think of) but receive a variety of errors.  
>
>I'm wondering if anyone has any experience using the ncomps option for
>runica, and if so, could you suggest specific syntax.  Or if anyone has
>any suggestions as to how I could approach the problems differently, I'd
>sincerely appreciate the input.
>
>Thanks in advance for your time,
>Peter
>
>----------------------------------
>Peter Bachman
>PhD Candidate, Clinical Psychology
>UCLA Department of Psychology
>
>
>_______________________________________________
>eeglablist mailing list eeglablist at sccn.ucsd.edu
>http://sccn.ucsd.edu/mailman/listinfo/eeglablist
>Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
>To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
>  
>