Hi MH
Thanks for your answer. In your experience, do you have any suggestion about which approach or quantitative metric could be used (more suitable) to screen the motion problem of the entire fMRI time series data? I want to get some single metrics based on motion profile to decide whether some dataset have serve motion problem and not to include in further group analysis. Thanks !
Best
Paul
Date: Fri, 5 Jul 2013 18:18:43 +0000
From: [log in to unmask]
Subject: Re: [FSL] retry of mcflirt's mean displacement
To: [log in to unmask]
Hi Paul,
The "fd_rms" option in 'fsl_motion_outliers' just uses the fmri_mcf_rel.rms file generated by MCFLIRT. So, it is not the same as using 'rmsdiff'.
cheers,
-MH
--
Michael Harms, Ph.D.
-----------------------------------------------------------
Conte Center for the Neuroscience of Mental Disorders
Washington University School of Medicine
Department of Psychiatry, Box 8134
660 South Euclid Ave.
Tel: 314-747-6173
St. Louis, MO 63110
Email: [log in to unmask]
From: Chou Paul <[log in to unmask]>
Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
Date: Thursday, July 4, 2013 8:51 PM
To: FSL - FMRIB's Software Library <[log in to unmask]>
Subject: Re: [FSL] retry of mcflirt's mean displacement
Hi Mark
May I ask for one additional question about this thread. Currently I am focusing on some motion related metrics about rsfMRI data and try using some quantitative motion measurement to screen my own dataset. I found there is a command line tool "fsl_motion_outliers"
in FSL. Is the rmsdiff approach which you mentioned here is the same with the "fd_rms" option in "fsl_motion_outliers" ?
Best
Paul
Date: Thu, 4 Jul 2013 21:09:31 +0000
From: [log in to unmask]
Subject: Re: [FSL] retry of mcflirt's mean displacement
To: [log in to unmask]
Hi,
What you are doing looks fine.
The chances are that the differences are just due to the fact that internally within mcflirt it uses the centre of the volume as the centre of a sphere (the region over which the calculations are made), whereas in rmsdiff it will use the centre of gravity
of the image. When the brain is not centred well in the image then this will give different averages. Also note that there is no "correct" solution in this case, because the choice of centre is arbitrary and it affects the balance between the rotational
and translational components. For most applications of this output it will make very little difference which way it is done.
If I had to pick a preferred option then I would actually go for neither of these, but use rmsdiff with the mask derived from the functional image (just called "mask" in a feat directory). This calculates the rms values over a more representative brain
ROI (rather than a sphere) and does not require any origin/centre to be used. I can't objectively defend this as being correct or better, but I think it is more intuitively related to what we think of wrt average movement.
I've added some more info on the wiki relating to rmsdiff. You can find this at:
https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FLIRT/UserGuide#rmsdiff
All the best,
Mark
On 3 Jul 2013, at 12:22, Michael Bietenbeck <[log in to unmask]> wrote:
Hey,
since my bash-scripting skills are even worse than those for matlab, i've tried to run it with the later one
refvol = 005.nii.gz
for i = 1:220
% refvol = 110
mat_in = mat{1,i};
mat_ref = mat{1,110};
% rmsdiff matrixfile1 matrixfile2 refvol
rms_in = ['$FSLDIR/bin/rmsdiff ', mat1, ' ', mat_ref, ' ', refvol];
[~,rms_out] = system(rms_in);
rms_out_num(i) = str2double(rms_out);
end
Thank you for your help
best regards
Michael
The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended
recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone
or return mail. 
