Hi Tudor, Good to know that it went fine! Normally the lGI takes about 2-3h per subject to complete, so if you launched them sequentially and not in parallel, I'm not surprised that it took about 3 days. (you can shorten the computational time by launching as many parallel sessions as you have available processors, e.g. 4 batches of 9-10 subjects)
As for the ones that had aberrant values, the best way to solve this problem is to rerun the surface reconstruction for the hemisphere that is concerned, including the randomness flag (recon-all -autorecon2-wm -randomness -autorecon3 -hemi ?h -s #SUBJECT) and then rerun the lGI for this hemisphere. That should allow the lGI to finish properly. Let me know if not. Best, Marie On Jun 17, 2013, at 12:27 PM, Tudor Popescu <tud...@gmail.com<mailto:tud...@gmail.com>> wrote: Hi Marie After checking that Matlab on FSL indeed has the image processing toolbox, I ran the batch command (-localGI for every subject) through FSL, and it all worked fine - although it took something like 3 days to complete for all 38 subjects :) Unfortunately, for two of my subjects, ?h.pial_lgi files were not created at all, with no error logs that I could find, and for two additional subjects, only the lh surface was created, with the following error found in the log files corresponding to the rh: Problem for vertex iV = 40101, lGI value is aberrantly high (lGI=48.5662)... (and similar for the other subject whose rh was missing). If I open the /mri/brain.mgz and /surf/?h.white of those 4 problem subjects in freeview, they all seem fine. What other volumes/surfaces should I check, in order to identify what might have led to that aberrantly high lGI value? Or, otherwise, do you know why this error might come about? Thank you very much for your help Marie! Tudor On 14 June 2013 19:53, Marie Schaer <marie.sch...@unige.ch<mailto:marie.sch...@unige.ch>> wrote: Hi Tudor, Then the problem is certainly related to the fact that you don't have matlab (the folder /usr/local/freesurfer/matlab is not the matlab distribution, rather the matlab functions enclosed in the FreeSurfer distribution), and not the fact that you have several copies of make_outer_surface.m. So, I am not a specialist of FreeSurfer running on a virtual machine, so I don't know if I can help you much. I guess the easiest work around (if you can) is to install matlab on your virtual ubuntu. I am not aware of any way to point to windows distribution of software from the virtual machine, I would guess that it's not possible given the different architecture of the OS, but maybe others from the mailing list could help here? Given that lGI uses both FreeSurfer and matlab functions, it's technically feasible to run part of the processes on your virtual machine for the FreeSurfer parts, and part of the processes in your windows matlab session. It's not impossible, but it would be a huge gymnastic and I strongly discourage you to do it, unless you really know what you do. Then, as for the FSL possibility to run FreeSurfer processes, I have to say that I am very unaware of how it works. I would say that as long as you can run a recon-all process that way, then the recon-all -lgi will work as well… but same thing, it will need the getmatlab command to point to the correct location. Hope you will find a way, sorry to be not more helpful, Marie On Jun 14, 2013, at 11:22 AM, Tudor Popescu <tud...@gmail.com<mailto:tud...@gmail.com>> wrote: Many thanks for your reply Marie. "getmatlab" returns "ERROR: Cannot find matlab. Does not exist or is not in path." There is a matlab folder in my Virtualbox installation of FS, at the path /usr/local/freesurfer/matlab, however it only contains .m files rather than the software itself. Among those files, the make_outer_surface function has a date stamp of 2007/11/14, with no other version information. I did a search in /usr/local, and that one seems to be the only copy of make_outer_surface, so there probably isn't any duplicate elsewhere. On my Windows PC I have Matlab 2010a. Can the localGI functions in any way be pointed to it? If not, then could this be done via FSL (on FMRIB cluster) maybe? So far I ran my recons through FSL, then did further analyses locally. When I type matlab in my FSL terminal, it says "MATLAB not allowed on this host. Run inteactive sessions on your local computer or processing jobs on the cluster". On the other hand, if I double click on /usr/local/bin/matlab2006a (in my FSL connection), the 2006a version of Matlab starts up, and seems to be working fine! What do you make of this? Thanks again! Tudor On 14 June 2013 17:31, Marie Schaer <marie.sch...@unige.ch<mailto:marie.sch...@unige.ch>> wrote: Hi Tudor, Yes you will need matlab with the image processing toolbox installed. The issue may be that you are running FreeSurfer in a virtual machine, with matlab not installed there. But, as Doug mentioned, I would assume the algorithms to answer you that it didn't find matlab. So, a few ideas: First, try to run getmatlab to see whether you get any answer, and to which matlab it points to. If you have a matlab installed, please let me know which version, as I have had issue with the most recent versions (but not the same issue that you are reporting) Second, the specific issue of too many or not enough arguments could be found if you have different version of a function with the same name, and the other ones need less arguments. You can use the which function in matlab to see whether this is the case for make_outer_surface.m. Let me know, Marie NB: as for your last question, there is no need to delete anything before rerunning when you will have solved the problem, it will simply overwrite. On Jun 14, 2013, at 7:20 AM, Douglas Greve <gr...@nmr.mgh.harvard.edu<mailto:gr...@nmr.mgh.harvard.edu>> wrote: Hi Tudor, the lGI stream definitely uses matlab, so that will be a problem at some point if not now. I would have thought that it would have reported that it could not find matalb. Marie would know for sure. doug On 6/14/13 3:25 AM, Tudor Popescu wrote: Dear Freesurfer & localGI experts, I want to compare local GI (gyrification index) between my two groups. Following the help<https://surfer.nmr.mgh.harvard.edu/fswiki/LGI>, I ran this command for one of my subjects: recon-all -s M01 -localGI but it resulted in the error: make_outer_surface('/media/math/all38subj/M01/surf/tmp-mris_compute_lgi-lh.pial/lh.pial.filled.mgz',15,'/media/math/all38subj/M01/surf/tmp-mris_compute_lgi-lh.pial/lh.pial-outer'); exit ================= ERROR:: Too many arguments. Someone previously<http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg23414.html> received this error due to a space in their path, but I don't specify any path in the above command, and my $SUBJECTS_DIR path doesn't have any spaces or any other dodgy characters. I'm running FreeSurfer v5.2 from inside Virtualbux/Ubuntu, so Matlab is not included in there, although it exists separately on my Windows partition. I could run recon through FSL, but I don't see any matlab folder there either. Could Matlab be the problem? Also, which files/folders will I need to delete after I'll have fixed the problem, before re-running the recon for all subjects? Or will it just overwrite any left-overs? Many thanks in advance for any help! Tudor _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu<mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
