Many thanks for so fast responce, Bruce!

I will give -make all try again tomorrow on few of the latest "unsuccessful" completes.

Thanks,

Martin

On Sunday 10 May 2009 22:07:58 Bruce Fischl wrote:

Hi Martin,

-make all is what we would suggest. Is there any reason the timestamps would be wrong on your files to make make all rebuild everything?

cheers, Bruce On Sun, 10

May 2009, Martin Kavec wrote:

...

Hi,

I have a bunch of analyzes (about 30), for which recon-all -all terminates prematurely at mris_volmask (and core is created). On one case I ran the mris_volmask and the rest of analysis by executing commands manually, which went fine, but it's laborious. Is there quicker way to complete the recon-all -all without need to go over the whole -autorecon3? ( for some reasons -make all starts the whole analysis over, which is not really necessary).

Thanks in advance,

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

Hi Martin,

there are probably some files it wants to see early in the stream that didn't exist in 2006. I think the only cure is to string together the individual commands you want to run (you can find them in the recon-all help).

cheers, Bruce On Tue, 12 May 2009, Martin Kavec wrote:

Hi Bruce,

-make all is not really helping; recon-all starts from the beginning. However, in these subjects I did not run the whole pipeline. In fact, here is what I did with the subjects: in 2006: recon-all -all, and in 2009: recon-all -autorecon2 -autorecon3, where the autorecon3 failed. It seems that -make all wants to run the whole pipeline, because in 2009 I ran only autorecon2 and autorecon3. I also tried -make autorecon3, which still tried to run the whole pipeline.

Would you have any cure for this situation?

Thanks,

Martin

On Sunday 10 May 2009 22:07:58 Bruce Fischl wrote:

...

Hi Martin,

-make all is what we would suggest. Is there any reason the timestamps would be wrong on your files to make make all rebuild everything?

cheers, Bruce On Sun, 10

May 2009, Martin Kavec wrote:

...

Hi,

I have a bunch of analyzes (about 30), for which recon-all -all terminates prematurely at mris_volmask (and core is created). On one case I ran the mris_volmask and the rest of analysis by executing commands manually, which went fine, but it's laborious. Is there quicker way to complete the recon-all -all without need to go over the whole -autorecon3? ( for some reasons -make all starts the whole analysis over, which is not really necessary).

Thanks in advance,

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

Hi Martin,

I'll leave this for Nick, but we fixed the atlas eTIV calculations in the new version. Not sure which step you need to rerun though. Bruce

On Tue, 12 May 2009, Martin Kavec wrote:

Hi Bruce,

On Tuesday 12 May 2009 13:28:36 Bruce Fischl wrote:

...

Hi Martin,

there are probably some files it wants to see early in the stream that didn't exist in 2006. I think the only cure is to string together the individual commands you want to run (you can find them in the recon-all help).

Your are right. Since it is very close towards the end of the pipeline, I am trying recon-all -cortribbon -aparc2aseg -wmparc, which seems to be running fine so far.

On the other hand I am finding that for some subjects atlas_icv (eTIV) = 2593116 mm^3 (det: 0.75126)

2.5 liters ICV :o ? Is that something I should be worried about? It is pretty high, if it is what I suppose it represents.

Thanks,

Martin

Martin,

That etiv number seems high. Does the -tal-check stage pass? did you rerun the -segstats stage? that calculates etiv.

Nick

On Tue, 2009-05-12 at 15:04 +0200, Martin Kavec wrote:

Bruce,

but I am running 4.3.0

Thanks, Martin

On Tuesday 12 May 2009 14:55:05 Bruce Fischl wrote:

...

Hi Martin,

I'll leave this for Nick, but we fixed the atlas eTIV calculations in the new version. Not sure which step you need to rerun though. Bruce

---

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