Dear Manuel,

I would chime in with my current experience with transfering v5.3 reconstructions to v6.0: The wm segmentation results (and therefore white/pial surfaces, apart from other things) are different between versions, especially in ambiguous areas and especially if there is a difference in interpolation during conform step (in v 6.0 the cubic interpolation is off; if it is off or on in v5.3 it depends on whether you used patched version of recon-all or not). It of course also depends on the quality of your data.

In any case, my experience is, if you want to keep high quality of your results, you probably would need to re-inspect your data again and correct newly appeared errors in surfaces.

Antonin Skoch

all the edits will be taken into account regardless of what command lineargs you use

On 6/9/17 3:08 AM, Manuel Delgado wrote:

Dear all,

Regarding this issue:http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg51931.htmlwe have the same situation. We are going to reprocess data previouslyprocessed with FS 5.3.If we proceed with *recon-all -all *will all the edits be taken intoaccount (i.e. skull strip, control points, wm and pial edits) or is itnecessary to proceed with ordered steps (i.e -autorecon-pial, then-autorecon2-cp autorecon3, and so on)?

Thanks in advance,

Manuel