Hi, me again...(sorry)
now that we've settled the first question, I am stuck with figuring out how to extract the new maps (I did split the melodic_IC file, as a first step -.-)
Question 1) I assume that the intensities within the components in the ICA output are important for the dual_regression stage_1, is that correct?
Therefore, I thought that a handdrawn binary mask of the voxels pertaining to a component (e.g. frontoparietal) and a region of interest (e.g. SPL) won't do.
Question 2) I stumbled across this command in the forum:
"You could create a structural mask from one of the atlases provided with FSL (they're in $FSLDIR/data/atlases/). Use fslmaths to create the structural ROI and binarize it, then multiply that by your activation map - it will give you areas that are significantly active AND inside the structure of interest." (posted by Kx on Feb 23rd, 2012 [Re: extract roi from activation map])
- Would this work for me and would this leave me with a mask that contains the intensities as in the original ICA map or with a binary mask?
Question 3) If there is no (better) way to extract the subregions of the ICA maps, including intensities, could I simply take the relevant ICA-map (e.g., frontoparietal, as extracted with fslsplit), overlay the relevant structure from one of the atlases (e.g., SPL), and erase all voxels of the ICA-map that are not within the SPL? ...it appears very crude, but maybe it's valid...
Thanks again for your help!
Helen
________________________________________
From: FSL - FMRIB's Software Library [[log in to unmask]] On Behalf Of Lückmann, Helen (Stud. FPN) [[log in to unmask]]
Sent: 24 June 2013 13:41
To: [log in to unmask]
Subject: Re: [FSL] Splitting components from ICA output into relevant subregions for further analysis
Great, thanks a lot!
________________________________________
From: FSL - FMRIB's Software Library [[log in to unmask]] On Behalf Of Stephen Smith [[log in to unmask]]
Sent: 24 June 2013 13:34
To: [log in to unmask]
Subject: Re: [FSL] Splitting components from ICA output into relevant subregions for further analysis
Hi
On 24 Jun 2013, at 02:55, Lückmann, Helen (Stud. FPN) <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Hi, thanks for the quick reply!
I am still a bit confused, though, because I do not understand how I can get separate time-courses for the "new" subcomponents.
You will generate a new 4D NIFTI file that you will use to feed into the dualreg instead of the original group-ICA spatial maps - then re-run dualreg and you'll get a new set of stage1 timeseries files.
To split group-level components, use fslsplit to create a whole set of 3D NIFTI files from the 4D group maps file, split the individual masks in whatever way is convenient (e.g. hand masking in fslview and saving out different masks), and then remerge with fslmerge (being careful to list the input files in the right order).
Cheers.
For FSLNets, I would need files as provided by dr_stage1, where each column contains a time-course. For my purpose, I would need those not for the original ICA output, but for the subregions I created from the components.
Is that even possible, considering the way the components were created?
Thanks again for your help.
Helen
________________________________________
From: FSL - FMRIB's Software Library [[log in to unmask]<mailto:[log in to unmask]>] On Behalf Of Stephen Smith [[log in to unmask]<mailto:[log in to unmask]>]
Sent: 23 June 2013 22:24
To: [log in to unmask]<mailto:[log in to unmask]>
Subject: Re: [FSL] Splitting components from ICA output into relevant subregions for further analysis
Hi
On 23 Jun 2013, at 15:54, Lückmann, Helen (Stud. FPN) <[log in to unmask]<mailto:[log in to unmask]><mailto:[log in to unmask]>> wrote:
Dear experts,
I ran MELODIC and dual regression on my resting state dataset and am now interested in the connectivity between three components that MELODIC produced. I'd like to investigate this using FSLNets.
If I understood correctly, FSLNets provides information about (partial) correlations between the components entered into the analysis. While this always refers to the relation between the complete components, I am rather interested in the relation between subparts (e.g. posterior part of the DMN or the PCC) of the two networks (within and between network connectivity) as well. I then want to compare correlations between groups.
I think this would mean I would have to split my component in two or more parts. What would be the best way to do this? Should I do this before Dual regression, on the stage2 component file (concatenated for the whole group),
Yes - that would be best. Split them and recombine into a new 4D file that will have an increased number of "time" points - that solves your question below too.
Cheers.
or for the individual stage 2 files? Most importantly: How do I get different time-courses for these components (as the ICA and DR seem to only provide component-wise, not voxel-wise, time-series)? Could I just use parcellations from the Harvard-Oxford atlas or is there a more precise way to do this?
Thanks for your help.
Helen
---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask]<mailto:[log in to unmask]><mailto:[log in to unmask]> http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------
Stop the cultural destruction of Tibet<http://smithinks.net>
---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask]<mailto:[log in to unmask]> http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------
Stop the cultural destruction of Tibet<http://smithinks.net>
|