Hi Gabriel

what are you trying to achieve? There's no point in building a gca from a single average subject. The gca is typically built be compiling statistics across many subjects.

Not sure if you need the answers any longers but (1) you can view the gca with freeview as it if were an image file (append #0 and #1 for viewing different "frames", in this case that denotes max probability label and mean intensity of that label). (2) the norm.mgz definitely shouldn't be empty, and (3) yes, mri_ca_label would be the right tool

cheers Bruce

On Wed, 2 Jan 2013, Gabriel Gonzalez Escamilla wrote:

Dear FreeSurfer experts

Happ new year to all of you, and thanx for all your help along the past 2012.

I would like to make some questions regardins a new atlas creation

1. In a first step before mri_ca_train I did use mri_ca_normalize to normalize my FS_MNI152average subject

to the JHU.ICBMtracts.1mm.mgz but this create an empty norm.mgz, is this normal? whean I saw that the norm.mgz was empty I decide to create all the files by: recon-all -s My_MNI152_avg -autorecon1 -autorecon2-inflate1, that create all files as normally does with my every other subjects;

Then I run mri_ca_train, using as -mask the brain.mgz, as -T1 the norm.mgz created during recon-all, directed to my FS_MNI251average subject, and as -parc_dir JHU.ICBMtracts.1mm..mgz; as result got the .gca and a mgh file I use the load_gca.m to check the .gca created, but get the next error message: "version 5 in file $JHU.ICBMtracts_12_2012.gca incorrect - not a known gca file" 2) Is there any way to check at the gca file created? just to make myself sure that everything is placed in the right way

3. And lastly, Which is the next step to parcellate the WM of all my subjects as the new atlas? Can I use

this gca file with mris_ca_label? or some other tool will be better?

Many thanks in advanced, Gabriel.