[Eeglablist] ICA - low density, spatially restricted montages

[Makoto Miyakoshi]

Dear Tarik,

> 8 channels is quite sparse, ICA and eye blink removal algorithms rely on
> relatively higher scalp coverage.

When I tested, 3 ch ICA showed as good EOG as 33 ch ICA.

http://sccn.ucsd.edu/eeglab/japanese/Clipboard01.jpg
(This is just for a temporary upload and I will delete it soon).

Makoto

2012/8/29 Tarik S Bel-Bahar <tarikbelbahar at gmail.com>:
> 8 channels is quite sparse, ICA and eye blink removal algorithms rely on
> relatively higher scalp coverage.
> However because all your channels are frontal, ICA might be resolving
> correctly. Some examination of the ERP, EEG signal before/after blink IC
> removal, and comparison of no-blink vs. after blink-removal is warranted,
> but you just might be able to proceed, with caution...
>
> some possible resources from goggle scholar below,
> also to give you some ideas about how test the validity of what you're doing
>
>
> 16 electrodes...
>
> Automatic removal of eye-blink artifacts based on ICA and peak detection
> algorithm
>
> J Gao, P Lin, Y Yang, P Wang - Informatics in Control, …, 2010 -
> ieeexplore.ieee.org
>
> 8 electrodes
>
> Detection and removal of ocular artifacts using Independent Component
> Analysis and wavelets
>
> H Ghandeharion… - Neural Engineering, 2009. …, 2009 - ieeexplore.ieee.org
>
> Analysis of Blind Source Separation Techniques for Eye Artifact Removal
>
> TH Aspiras, VK Asari - Wireless Networks and Computational Intelligence,
> 2012 - Springer
>
> Validation of ICA as a tool to remove eye movement artifacts from EEG/ERP
>
> M Mennes, H Wouters, B Vanrumste, L Lagae… - …, 2010 - Wiley Online Library
>
> [HTML] The correction of eye blink artefacts in the EEG: a comparison of two
> prominent methods
>
> S Hoffmann, M Falkenstein - PLoS One, 2008 - dx.plos.org
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> On Wed, Aug 29, 2012 at 12:09 AM, Yaseen Gerhold
> <sunshineafterdusk at gmail.com> wrote:
>>
>> ICA-gurus
>>
>> I have a low density  montage, 8 channels over the frontal and prefrontal
>> regions. I want to know if it is feasible to use ICA in this context for
>> artifact correction, or is ICAs application limited to higher density
>> montages?
>>
>>
>> RUNICA appears to resolve all the eye-blink data into one neat component
>> on a trial run: One can see this in spatial (scalp maps), frequency (fractal
>> dimension) and time domain. My concern is that with only 8 components to
>> resolve to, and limited spatial sampling (frontal region only), important
>> the cortical data may also be resolved to the artifactual component.
>>
>>
>> Can someone please share any experience they have in this context, or
>> provide some theoretical guidance in terms of ICA performance under such
>> conditions?
>>
>>
>> Yaseen
>>
>>
>> _______________________________________________
>> 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
>
>
>
> _______________________________________________
> 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