Dear Nick, thank you for pointing me out the bbregister problem which is causing FLAIRpial to fail in this case.
I have looked at this particular step and found that the problem is with -init-fsl flag, where initial flirt registration (via fslregister) fails in these data (in contrast to FLAIR_OK subject). The flirt registration .dat file is removed by default, I had to modify expert options of bbregister in order to retrieve it. I cannot see significant difference in quality between FLAIR images of FLAIR_OK and FLAIR_not_OK subject, which I sent to your server, therefore it seems that fslregister is not sufficiently robust with these types of FLAIR images and in some cases it totally fails. When I put -init-header to expert options of bbregister, the registration by bbregister is OK in this subject and FLAIRpial refinement seems to work OK. Some automatic check of registration will be indeed useful in this case. Or would you recommend to modify some option of fslregister to make this step more robust? How the normal user should proceed when one wants to re-run the FLAIRpial step? I expect that running recon-all -s subjid -FLAIRpial -autorecon3 -expert my_expert.opts will do it, but, In my case, I had to manually remove transform files, mri/FLAIR.prenorm.mgz and mri/FLAIR.mgz in order to bbregister run again (with expert options). Otherwise recon-all -FLAIRpial skip bbregister and began on mris_make_surfaces only. I quickly looked at the recon-all code, but I did not grasp how it is decided which steps to run and which not. When -make directive is not specified the make tools are not used for that, are they? Regards, Antonin _____ From: Nick Schmansky, MGH [mailto:ni...@nmr.mgh.harvard.edu] To: Antonin Skoch [mailto:a...@medicon.cz], Bruce Fischl [mailto:fis...@nmr.mgh.harvard.edu] Cc: Douglas N Greve [mailto:gr...@nmr.mgh.harvard.edu] Sent: Tue, 19 May 2015 20:33:05 +0100 Subject: Re: FLAIRpial correction fails in some cases bruce, doug (and antonin), i've put the files here if you want to have a look: /autofs/space/terrier_002/users/nicks/subjects/FLAIR_problems i looked at the FLAIRraw.mgz for the not_OK subject, and in my non-expert opinion, it doesn't look like a high quality image. not sure whats wrong with it, but maybe a lot of motion artifacts. in contrast, the T1 looks very nice. in this instance, i wouldn't include the FLAIR as the T1 surfaces seem ok (load ?h.woFLAIR.pial). also, i noticed that the FLAIR.mgz failed the bbregister procedure, in that FLAIR.mgz is rotated. doug, if you want to have a look into why bbregister failed in this subject (is there any way to detect a major failure like this so that recon-all can exit?), i made a copy of the data here: /autofs/space/terrier_002/users/nicks/subjects/FLAIR_problems/FLAIRpial_not_OK_copy n. On Tue, 2015-05-19 at 09:32 +0200, Antonin Skoch wrote: > Dear experts, > > I tried to process our data with the dev version of freesurfer. > Unfortunately the problem with FLAIR pial surface correction still > persists. I have uploaded the data to your ftp server > ( anonym...@surfer.nmr.mgh.harvard.edu/transfer/incoming ) that you > can inspect the results. > > The files are: > FLAIRpial_OK.tar.gz - subject where it looks that the pial correction > worked > FLAIRpial_not_OK.tar.gz - subject where it seems that the pial > correction failed > > Regards, > > Antonin > > ______________________________________________________________ > From: Z K [mailto:zkauf...@nmr.mgh.harvard.edu] > To: Bruce Fischl [mailto:fis...@nmr.mgh.harvard.edu], Antonin > Skoch [mailto:a...@medicon.cz] > Sent: Wed, 13 May 2015 16:39:26 +0100 > Subject: Re: [Freesurfer] FLAIRpial correction fails in some > cases > > Hello Antonin, > > Please try processing your subject with the dev version of > freesurfer as > it may fix the issue you are encountering. It can be > downloaded and > installed from the following ftp site: > > ftp://surfer.nmr.mgh.harvard.edu/pub/dist/freesurfer/dev > > -Zeke > > > On 05/13/2015 09:43 AM, Bruce Fischl wrote: > > Zeke: can you get Antonin a beta of 6.0 so he can see if it > fixes his > > flairpial problems? > > > > thanks > > Bruce > > On Wed, 13 May 2015, Antonin Skoch wrote: > > > >> Dear FreeSurfer experts, > >> > >> we are using FLAIR images to improve pial surface > detection. In many > >> cases > >> this works OK, however, in some cases this procedure fails > and the > >> corrected > >> pial surfaces are much worse than before. > >> > >> I uploaded 2 example subjects FLAIRpial_OK and > FLAIRpial_not_OK to > >> > >> ftp://surfer.nmr.mgh.harvard.edu/transfer/incoming > >> > >> Could you please look at them and comment on what is the > reason why it > >> did > >> not work in the subject FLAIRpial_not_OK and give > suggestion what to > >> do in > >> such cases? > >> > >> Regards, > >> > >> Antonin Skoch > >> > >> > >> > >> > > > > > > > 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
_______________________________________________ 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.
