Dear Christian,
I am aware that the lGI may occasionnally exit with the error you mentionned in some rare cases of pial surfaces with normal euler neumbers. I am working on debugging it.
In the mean time, if this is only for one subject, is it possible for you to recreate the surfaces just for the hemisphere where you have this problem (i.e. relaunching recon-all -hemi ?h -autorecon2 and -autorecon3)? This would probably be the easiest way to overcome the aberrantly high lGI issue, as it will change very slightly the position and the total number of vertices of the white and pial surfaces, which will normally solve the problem. However, that implies that you would have to redo any statistical analyses that you may have already done on cortical thickness or cortical volumes (but there is no reason that the results would change just with reprocessing one hemisphere with the same freesurfer version).
Let me know how it goes,
Marie
Quoting Christian Scheel demi_bee@gmx.net:
Dear all,
I have successfully run 'recon-all -lgi' for all my subjects. However the computation process stopped in one of the brains claiming an lGI vertex value was aberrantly too high and giving me the following error message:
#####
... remeasuring lGI value for vertex iV = 32501. It may take a few minutes. WARNING -- Problem for vertex iV = 32501, lGI value is aberrantly high (lGI=40.5282)... ...lGI computation will be stopped. This may be caused by topological defects, check mris_euler_number on the pial surface.
ERROR: compute_lgi did not create output file '/home/freesurfer/christian/34/surf/tmp-mris_compute_lgi-rh.pial/rh.pial_lgi.asc'!
#####
I have already restarted the computation twice without success. The script always crashes at the same vertex. The error message and the wiki page mention such an error in connection with possible topological defects. However when I check the mris_euler_number for the subject's rh.pial it interestingly does not show me any topological defect. Here is the output:
#####
[freesurfer@localhost ~/christian]$ mris_euler_number /home/freesurfer/christian/34/surf/rh.pial euler # = v-e+f = 2g-2: 133775 - 401319 + 267546 = 2 --> 0 holes F =2V-4: 267546 = 267550-4 (0) 2E=3F: 802638 = 802638 (0)
total defect index = 0
#####
In all other brains the lGI calculation was successful. Does anyone of you have an idea? Any help is greatly appreciated.
Best wishes, Christian Scheel
University of Cologne Department of Psychiatry _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer