Hi Gwenaelle,
thank you for your reply.
>This sounds ok to me. But this is roughly what fslvbm_1_bet -N actually does... Just checking: you're inverting the bet >mask back to native space and then apply it to the native, original, struc image, right?
Actually I didn't create a brain mask with bet: I first flirted the original images (=brain+everything else) to the std space in order to have a common cohordinates system where to apply common bet options. Then I just inverted, using "convert_xfm", the flirt matrixes (one for each subject) and applied, one by one, each inverted transformation to the corresponding betted image...I don't know if this explanation is clear...If it is, does it still sound ok? I continue to doubt that such an approach doesn't introduce distortions...
About the 36 subjects forgotten by fslvbm_2 I don't know what happened: actually, since then everything is working fine, so probably I did something wrong that I fixed at some point without knowing it...sorry about that.
Thank you again for your help,
Cheers,
Chiara
|