Actually, this is something that has been proposed by Christian Beckmann himself as a potential way of doing it. Theoretically, you could just enter one or two maps, e.g. anatomical maps, or task-fMRI generated maps for that matter...
What you are doing in your case, then, is looking at 10 "seeds" simultaneously, not regarding if those seeds/maps are present in your data, and certainly disregarding the influence of CSF/respiratory/cardiac maps that typically get detected by ICA also. You could add them manually without problems, but you would need to argue how you generated them. There is no "theoretical" problem there - it is just another way of interrogating the data (and hence getting slightly different answers to your questions).
Actually, what I would do is the following: under the assumption of a (more or less) homogeneous sample (which you should corroborate somehow), randomly pick 20-30 subjects and generate a set of 70-90 ICAs from your data and feed that into DR. Alternatively, you could try to reduce the resolution in the Melodic setup (go from 4 to, let's say, 6mm -> you should need 3.375 times less memory) and see if melodic finishes this time. How much RAM has your rig, actually?
With 120 subjects, you should get far more than 25 ICAs, by the way - unless you pre specified that number, that is. And don't expect DR to run much faster than the original Melodic run, then….
Hope that helps,
Cornelius
Am 26.02.2013 um 17:58 schrieb Jacobs H (NP) <[log in to unmask]>:
> Hi all,
>
> Could somebody give me some feedback on whether this approach could be problematic or not advised?
> As mentioned earlier on the list, my ICA is not running well on 120 brains.
> So, as I need to move on with these analyses, I was thinking of running the dual regression with the 10-components-map (from http://www.fmrib.ox.ac.uk/analysis/brainmap+rsns/) as a template (instead of the data-driven ICA file).
> Would that be ok to do? Should I also add white matter and CSF?
>
> Thanks!
> Best
> Heidi
>
>
>
> ================================================
> Dr. Heidi Jacobs
> Postdoc researcher
> Faculty of Health, Medicine and Life Sciences
> School for Mental Health and Neurosciences
> Division Cognitive Neuropsychiatry and Clinical Neurosciences
> Alzheimer Center Limburg
> [log in to unmask]
> www.maastrichtuniversity.nl
> www.heidijacobs.nl
>
> Dr. Tanslaan 12, 6229 ET Maastricht
> P.O. Box 616, 6200 MD Maastricht, The Netherlands
> T +31 43 38 84 090 F +31 43 38 84 092
> ================================================
|