Dear fsl user,
I want to run fslvbm on a large number of subjects (n=100) and I have a couple of questions:
1) I tried to run fslvbm_1_bet first, but the results were quite poor since every original image seemed to have a different issue, e.g. one has too much neck, another is not well centred, etc... I decided therefore to write a script that first "flirts" the images on MNI standard space, then "bets" them and finally applies the inverse transformation obtained with flirt to have all the images back in their real space. Is this procedure correct, or applying so many transformations introduces wrong artifacts?
2) On my processed brain extracted images, I ran then fslvbm_2_template and it worked well for all but 36 images... when the job finished I ran it only on those "failed subjects" and this time it worked perfectly...I didn't change anything in the filenames and I can't understand what happened...do you have any idea on how to explain such a mistery?
Thank you in advance for any help,
Cheers,
Chiara
|