Hi Mike,

I've run into the same problem you see in number 1 - the SPLINE_CUBIC_SET - Fatal Error!, with a fine talaraich.  I was able to fix this by using the -no1d flag in mri_normalize.  You can add this as an xopt to that subject so that it is always run:
http://surfer.nmr.mgh.harvard.edu/fswiki/recon_2dall#head-25c704d7d03b1c22e5f712383f6045d2b4156957

I think this occurs when there is a bigger bias field in left-right direction than in the anterior-posterior direction (although, maybe it's I/S, I forget now).

I'll leave the other two problems for someone more qualified!
Jenni
--
Learning and Memory Lab
The University of Texas at Austin
Department of Psychology
1 University Station A8000
Austin, TX 78712
phone: (512) 471-2727
fax: (512) 471-5935
email: jpacheco@mail.utexas.edu


On Tue, Feb 26, 2008 at 3:29 PM, Michael Harms <mharms@conte.wustl.edu> wrote:

Hello,
We recently just ran 80 elderly brains through -autorecon1 using FS
v4.0.2, and encountered 3 different types of errors that we are not
entirely sure how to handle:

(1) During mri_normalize of 1 subject we encountered:
SPLINE_CUBIC_SET - Fatal error!
 The number of data points N must be at least 2.
 The input value is 1.

The FS archive indicates to check the talairach transform given this
error, but the talairach.xfm appears fine for this subject (as assessed
using tkregister2).

(2) During mri_watershed of 1 subject:
Sorting...
mri_watershed Error:
 Problem in the COG calculation

This subject had a slightly warped talaraich.xfm, but I think more
relevantly, its talairach_with_skull.lta was way off.

(3) During mri_watershed of 5 subjects:
mri_watershed Error:
 GLOBAL region of the brain empty !

For all 5 of these subjects, the talairch_with_skull.lta also had
obvious problems.

RE (1): Any insight would be appreciated.

RE (2) and (3): I suspect that we should try running with the -no-
wsgcaatlas flag, so that talairach_with_skull.lta is longer used by
mri_watershed.  Should we then substitute the -wsatlas flag instead?
(Its not clear to me which atlas is used when that particular flag is
included).

Last, I thought you might be interested in these errors, because they
directly relate to my recent question regarding the robustness of
talairach_with_skull.lta.  At least during the -autorecon1 (i.e.,
initial) creation of this lta file, it appears that there is some
possibility of a bad talairach_with_skull registration (6 of 80 in this
particular subject set).

thanks,
Mike H.
--
Michael Harms, Ph.D.
--------------------------------------------------------------------
Conte Center for the Neuroscience of Mental Disorders
Washington University School of Medicine
Department of Psychiatry, Box 8134
Renard Hospital, Room 6615           Tel: 314-747-6173
660 South Euclid Ave.                Fax: 314-747-2182
St. Louis, MO 63110                  Email: mharms@wustl.edu
--------------------------------------------------------------------


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