ICA and artifact rejection

[Nicolas Robitaille]

Dear Floris

I used ICA to remove noise from MEG signal. I can share my experience. If calculation time is an issue, I suggest to use FastICA instead of infomaxICA, which is the default. You need to download the toolbox yourself (http://www.cis.hut.fi/projects/ica/fastica/), but Fieldtrip take charge of it afterward, it's a new option in componentanalysis. For both infomaxICA and FastICA you can specify to the routine a specific numbers of components (like 100 instead of 275) to output, which should also reduce computing time. You can also select a subset of your data that contains all the artefacts you are interested to remove, run ICA on that, and apply the signal reconstruction without the artefactual-component to the entire dataset.

For artefactual-component identifications, I suggest to look simultaneously at the topography and the time-course of the components. My position is that we do not absolutely need robust objective criteria for artefactual-component identification, unless we want a fully automatize process (that I would not trust for now). Furthermore, it's always a good idea to look at the raw data you have, prior to any signal processing. The EOG and ECG components are obvious, you will pick them easily. Just find a time-window of raw data that shows these artefacts, and look at the time-course of the components. I suppose this would be the same for muscle-related artefact. 

Others artefacts are more complex for ICA. I would remove segments with jumps and head movements. Breathing artefacts tended, in my MEG data, to spread across severals components, so it was tedious to identify them. You may also consider if the time requiered to spare a noizy subject is worth, comparing to the time (and cost) to test another one. 

Hope this help,

Nicolas

----------------------------------------
> Date: Tue, 27 May 2008 16:27:41 +0200
> From: florisdelange at GMAIL.COM
> Subject: [FIELDTRIP] ICA and artifact rejection
> To: FIELDTRIP at NIC.SURFNET.NL
> 
> Dear Fieldtrippers,
> 
> I've started to play with using ICA to remove artifacts from the MEG 
> signal. Now, there's a couple of questions that came to my mind. I hope 
> anyone of you can help me with them.
> 
> 1) ICA can be pretty slow, and it has been suggested in the Fieldtrip 
> mailing list to use PCA and estimate only the 50 components that explain 
> most variance. However, when I do this to e.g. identify ECG-related 
> artifacts, I run into trouble when trying to decompose the ECG-locked 
> data into the components, using the previously estimated 50 components 
> (in order to calculate the coherence between the components and the ECG 
> signal). The ECG-locked data is 275 channels X number of time points, so 
> it cannot be simply decomposed into 50 components. Does anyone know how 
> to go about this?
> 2) For EOG I suppose I could do like for ECG, and calculate the 
> coherence with VEOG in order to select components that are artifacts. 
> But this is not so easy for head muscle-related artifacts (jaw and neck 
> movements, etc.), since we typically don't measure neck and jaw muscle 
> EMG. Apart from looking at spatial topography, is there a robust 
> criterion that one could use for detection of EMG artifact components?  
> Or would it be better to use 'classical' artifact rejection routines for 
> these kinds of artifacts? And would ICA also be suitable to remove 
> things like jumps and head movements in the MEG, or is it better/safer 
> to just throw away those kind of data segments?
> 
> Thanks for your help,
> Best wishes,
> Floris
> 
> ----------------------------------
> The aim of this list is to facilitate the discussion between users of the FieldTrip  toolbox, to share experiences and to discuss new ideas for MEG and EEG analysis. See also http://listserv.surfnet.nl/archives/fieldtrip.html and http://www.ru.nl/fcdonders/fieldtrip.

_________________________________________________________________

----------------------------------
The aim of this list is to facilitate the discussion between users of the FieldTrip  toolbox, to share experiences and to discuss new ideas for MEG and EEG analysis. See also http://listserv.surfnet.nl/archives/fieldtrip.html and http://www.ru.nl/fcdonders/fieldtrip.