Hi Antonin,

almost,

you need to use the existing transforms from the base (when the T1 template get’s created) and apply these to the co-registered T2 of each time point. In recon-all base code you will find a mi_robust_template call that takes existing lta transforms for each time points and simply create an new template (based on the inputs) without actually registering (—inform and some other flag to avoid iteration). I think the orig.mgz in the base was created this way. You would run a similar command on the T2’s of each time point.

Best, Martin

On Oct 20, 2016, at 1:03 AM, Antonin Skoch ansk@ikem.cz wrote:

Dear experts,

I am potentially also interested in correct using of -T2pial feature with longitudinal stream.

The first (standard cross-sectional) and third step (-long ) was already elaborated here, but I am, however, not sure how to process by T2pial in the second step ( -base), which creates the unbiased template and process it by (almost) standard cross-sectional stream. I did not find creation of unbiased template of T2 in the recon-all code.

Therefore, in order to use -T2pial on -base run, I would probably need to create my T2 template by calling mri_robust_template on my T2's first, and then pass the created T2 template as input to recon-all -base run, like:

recon-all -base <templateid> -tp <tp1id> -tp <tp2id> ... -all -T2pial -T2 path_to_T2_template

I am correct?

Regards,

Antonin _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer