[Eeglablist] ICA: Mutually exclusive conditions only?

[Makoto Miyakoshi]

Dear Helen,

> Should we be certain that all our conditions are mutually exclusive

Not necessarily. Seeing alphabet A and B is quite redundant for the brain
but not for psychologists :-)

> with no overlapping trials?

If the overlap is partial, it may bias ICA (for example, you feed 100 sec
of Condition A and 200 sec of Condition B to ICA, the results are more
biased to Condition B in theory). If the overlap is complete e.g. exact
repetition of the data, ICA will not be biased.

Makoto



On Mon, Jul 6, 2015 at 10:14 AM, Helen L. Wieffering <hwieffer at bowdoin.edu>
wrote:

>  Hi All,
>
> When using ICA, we are concatenating our conditions into a single dataset
> and running them jointly.
>
> Should we be certain that all our conditions are mutually exclusive, with
> no overlapping trials? Or is this a non-issue? Looking for an explanation
> on how (or whether) ICA operates differently if certain trials are
> represented more than others.
>
> Thanks for any and all help.
>
> Helen Wieffering
> Bowdoin College
>



-- 
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/20150710/d74b0f12/attachment.html>