/usr/local/fsl/bin/flirt -in /TEST_DIR/bold_brain.nii -ref /usr/local/fsl/data/standard/MNI152_T1_3mm_brain.nii -out /TEST_DIR /flirt_output.nii -omat /TEST_DIR /flirt_output.mat -bins 256 -cost corratio -searchrx -90 90 -searchry -90 90 -searchrz -90 90 -dof 12 -interp trilinear
/usr/local/fsl/bin/flirt -in /TEST_DIR /bold_brain.nii -applyxfm -init /TEST_DIR /flirt_output.mat -out /TEST_DIR /post_flirt_AND_applyxfm.nii.gz -paddingsize 0.0 -interp trilinear -ref /TEST_DIR /flirt_output.nii.gz
Perhaps I'll look into FNIRT. I'd feel a lot better nailing this down to a single step. Feels less "noisy" to me.
Damion
Hi Damion,
I can not tell you with clarity. Which command lines did you use for doing this?
I normally use FLIRT and FNIRT.
Kind regards,
Rosalia
El 9 oct. 2016 20:39, "Damion V Demeter" <[log in to unmask]> escribió:Hey Rosalia,Sorry for the late reply. I don't think I have an easy way to send you sample files, but I also think I solved my issue. It feels a little weird, but if I run applyxfm with the .mat file and that reference frame (the one I was talking about) that flirt created....I get my atlas normalized functional image as expected.It feels clunky and I am worried I'm introducing extra steps, but the output passes all visual QA I can think of and looks good. Thoughts on this "workflow"?DamionOn Sat, Oct 8, 2016 at 10:32 PM, Rosalia Dacosta Aguayo <[log in to unmask]> wrote:Hi Damion,
Could you please send me just one subject fmri with its structural as well as the result you got? I will give a look to it but not just now because I am from Spain and here it is 05:30 am esrly in the morning. I will give a look into this in the next few hours.
Rosalia
El 9 oct. 2016 5:28, "Damion V Demeter" <[log in to unmask]> escribió:Hi Rosalia,My apologies for not giving complete info, I was trying to be concise. Yes, I have a T1 image as well and am using it when running the flirt method. The biggest point of confusion is why my output image (that looks great and is definitely aligned to the atlas; verified by overlay in fslview) is just the single frame reference and not an entire 4d image like the input.Any ideas of what I'm overlooking? Or perhaps, if you have time, could you explain the workflow you'd use to accomplish this? From all I've read, I am doing it correctly, but my output says otherwise.DamionHi Damion, do you have structural T1 images for those subjects? A coregistration would be of great help when you have low resolution before MNI registration.
Kind regards,
Rosalia
El 9 oct. 2016 4:55, "Damion Demeter" <[log in to unmask]> escribió:I have searched the archives and tried a few things I read there, but I am not having much luck. This feels like a very basic question, but I can't seem to get it working the way I need. I apologize if this is a repeat post, but I was not able to find a solution that works for me.
What I am trying to do: I am trying to register a functional scan to the MNI atlas. I have successfully done this with flirt (turning off all smoothing, etc because I only want the transform step), but my output file is a single frame (reference file?). I have tried to do this with applyxfm as well, but the image gets moved around and at least one view ends up chopped off (I'm guessing because the voxel size of the acquired image isn't the same as the atlas?).
It's frustrating because this feels like it should be a rather simple function, but I can't seem to find anything about how to do this transform without running it through all the steps of FEAT. I'm hoping someone can point me in the right direction or to the proper tool (or perhaps knows what I'm doing wrong with FLIRT if that is, indeed, the correct tool to use).
Thanks in advance!