[Eeglablist] The number of removed components

[Arnaud Delorme]

Dear Bin,

I personally rarely remove more than 3 to 5 components (even with 256 channel montage), only the ones I am sure are artifacts (eye blinks, lateral eye movements, temporal muscle). So 20 might be too many. But again, there is no rule.

Best,

Arno

On Nov 4, 2014, at 3:36 PM, Bin Wang <binwang87828 at gmail.com> wrote:

> Hi, Makoto,
> 
> Thank you very much. You are right. 
> 
> I am using a dataset with 64 channels. Normally I can get 64 ICs, so if I removed 20 artefact-ICs, I still have about 40 ICs that are confirmed to be brain activity, which is totally reasonable. As long as I can make sure that these 20 IC are related with artefacts, I can remove them and won’t remove much of my data variance. Am I right?
> 
> Bin 
> 
> 
> ++++++++++++++++++++++++++++++
> Bin Wang
> Cognitive Neuroscience
> Institute of Neuroscience and Medicine (INM-3)
> Forschungszentrum Juelich, 
> Leo-Brandt-Str. 5 
> 52425 Juelich
> Email: b.wang at fz-juelich.de
>      &  binwang87828 at gmail.com  
> Telefon: 02461-61-9542
> ++++++++++++++++++++++++++++++
> 
> 
> 
> 
>> On 04 Nov 2014, at 21:11, Makoto Miyakoshi <mmiyakoshi at ucsd.edu> wrote:
>> 
>> Dear Bin,
>> 
>> We only use up to around 20 ICs that is confirmed to be brain EEGs regardless of the channel number.
>> 
>> By the way you can't say how many IC removal is appropriate. If you remove IC1-10 then you'll remove more than 50% of your data variance, while if you remove IC90-100 it could be less than 1%.
>> 
>> Makoto
>> 
>> On Wed, Oct 29, 2014 at 1:22 PM, Bin Wang <binwang87828 at gmail.com> wrote:
>> Hi, everyone,
>> 
>> One simple question: I am using ICA to remove the artefact right now. How many removed components is reasonable? I am using a completely automatic algorithm (ADJUST) now. It seems work for most of subjects. But, it is wired for one subject, I need remove 20 components of 60. Is it reasonable? 
>> 
>> Thank you in advance,
>> Bin
>> 
>> 
>> 
>> 
>> 
>> _______________________________________________
>> 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
> 
> _______________________________________________
> 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

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20141106/cc53738d/attachment.html>