Infant structural T1 data segmentation

[rosella1234]

Hi,
I would like to run developing HCP pipeline to perform segmentation of neonatal T1 scans. I know it surely works on T2 data , I would like to know if it is suitable to segment T1, too. Indeed, it is the only structural kind of scan I have,
Best
Rosella

[ecr05]

Hi, no, I'm afraid it won't work with T1. It may be that the baby connectome or freesurfer have some options you can use. Emma

[romainVala]

Hello

I would be very interested to have a working solution for T1w volume of the dHCP dataset. because both acquisition have different artifact, I think it is important to have an independent segmentation of both modality (if not a joint solution)

I did try with the current version to give a T1w image as argument of the -T2 option (without a -T1 option) :
It did not work for the brain mask, but if I give a cropped T1w volume then I get a reasonable result ... but the quality of the segmentation is worse than with the T2w as input. so I guess there is still something wrong ...

Why is it a problem ? I though the the drawEM is contrast agnostic, and it should work as good for T1 or T2 images no ?

Is there any hope to have a working solution ?

Many thanks
Romain