(I resend the mail due exceeding max message size.)

Hello, Niels.

Thanks for the advice!

I did perform fslrerorient2std on my images before running run_first_all.

I share my input and output images of a subject from each step-wise process like the following:

(1) original T1 file

[log in to unmask]">

(2) brainmask.mgz file converted to nifti file for FSL (mri_convert)

[log in to unmask]">

(3) nifti file reoriented to MNI152 using fslreorient2std

[log in to unmask]">

(4) resulting image of "all_fast_firstseg.nii.gz" after run_first_all was performed on the reoriented nifti image

[log in to unmask]">

This subject resulted in a total jumble of different subcortical regions disoriented through out the axis.

Other subjects have problems like (1) run_first_all simply seems to stall, (2) doesn't result in "all_fast_firstseg.nii.gz", (3) random subcortical regions are missing or distorted.

What could've happened to all these images?

Thanks so much.

Sincerely, Sooyun Cho.

--------- 원본 메일 ---------

보낸사람 : Niels Bergsland <[log in to unmask]>
받는사람 : <[log in to unmask]>
받은날짜 : 2016-11-28 (월) 18:22:45

제목 : Re: [FSL] "run_first_all" not working for children (ages 11~13)

Hi,

It would be helpful to see your input image as well. Regarding your question, they don't need to be manually aligned beforehand. However, you may want to try running fslreorient2std on your input first (e.g. fslreorient2std input input_reoriented) 

On Mon, Nov 28, 2016 at 9:48 AM, "조수연" <[log in to unmask]> wrote:

Hi, Gabor.

Thanks for the comment.

As I've said in my previous mail, I'd already tried using the non brain extracted T1 images for run_first_all, but a same error regarding non-brain tissue removal (assuming either for either BET or FLIRT) seemed to occur for a couple of my subjects' images.

So I reran run_first_all using brainmask images obtained from Freesurfer using the '-b' option.

Yet there still seems to be issues since some subjects are stalled during run_first_all, some do not produce the 'all_fast_firstseg.nii.gz' file, and those files that did produce 'all_fast_firstseg.nii.gz' file have subcortical areas like the R_Caud missing.

I doubt there's anything wrong with the protocol used for the acquisition for the image data since our protocol was based on ADNI.

Do you have any assumptions as of what the underlying problem might be?

To add, is it a must for run-first_all to assign each subjects' T1 images' AC-PC line and (0, 0, 0) centering manually? I don't recall doing so for my previous study with adult subjects but wanted to ask anyways just in case.

Thank you.

Sincerely, Sooyun Cho.

--------- 원본 메일 ---------

보낸사람 : Gabor Perlaki <[log in to unmask]>
받는사람 : <[log in to unmask]>
받은날짜 : 2016-11-28 (월) 16:29:13
제목 : Re: [FSL] "run_first_all" not working for children (ages 11~13)

For run_first_all you should use the original non brain extracted T1 images.

https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FIRST/UserGuide "The argument -i specifies the original T1-weighted structural image (only T1-weighted images can be used). " 

If you use a brain extracted image as input, you should use the -b option. However, I think you should try to run FIRST on the non brain extracted data first. 

Best Regards,
Gabor

--

Niels Bergsland
Integration Director
Buffalo Neuroimaging Analysis Center
100 High St. Buffalo NY 14203
[log in to unmask]