 |
 |
Dear Leonardo, I'm not too surprised by your new result but I'm not sure what to suggest next - I'm counting on Torben here but I've heard he is very busy since he signed up for Twitter (@torbenelund) ;-) Personally, I would be quite extreme and mask out this region before analysis (if you think this might affect coregistration/normalisation, globals or smoothness estimates, or spread to other nearby regions due to smoothing). But it might indeed be a bit extreme so let's wait for Torben or anyone else to comment. Best regards, Guillaume. On 12/07/16 08:54, Leonardo Tozzi wrote: > Dear Guillaume, > > Sorry for the late reply. > > Thank you very much for your help. I have now run this procedure leaving > all the default settings and setting “Jacobian deformations: Yes” and I > attach the result on one subject (top: before, bottom: after) . How do > you think the images look like now? The artifact doesn’t seem very > different. Would you suggest changing some of the unwarp options? > My main concern would be using these images in a whole-brain analysis. > Do you think this could still be done while taking the (possible) > results in that small area with extra caution? > Thank you, > > > > > > By the way, when would you suggest running it? As the very first > preprocessing step, like the realignment? > > > Leonardo Tozzi, M.D. > > /PhD Candidate/ > /Discipline of Psychiatry > Trinity College Institute of Neuroscience/ > /College Green > Dublin 2/ > /Ireland/ > / > / > /Office: +35318964234/ > /Mobile: //+353851026166/ > >> On 27 Jun 2016, at 18:42, Guillaume Flandin <[log in to unmask] >> <mailto:[log in to unmask]>> wrote: >> >> Dear Leonardo, >> >> it seems like this option has been rarely used until Torben mentioned >> it... (the help text is rather discouraging). I got it running by >> replacing size(.,1) with size(.) in spm_uw_estimate.m l.630-632: >> >> [nx,mx1] = size(Bx1); [nx,mx2] = size(Bx2); >> [ny,my1] = size(By1); [ny,my2] = size(By2); >> [nz,mz1] = size(Bz1); [nz,mz2] = size(Bz2); >> >> This will be available in the next update of SPM12. >> >> Best regards, >> Guillaume. >> >> >> On 27/06/16 10:17, Torben Lund wrote: >>> That was indeed the one I was thinking about, but I have never tried >>> this option, and I am sorry to se it fail. >>> >>> Best Torben >>> >>> >>> Den 27. jun. 2016 kl. 11.12 skrev Leonardo Tozzi <[log in to unmask] >>> <mailto:[log in to unmask]> >>> <mailto:[log in to unmask]>>: >>> >>>> Dear Torben, >>>> >>>> >>>> Thank you very much for your swift and thorough reply. >>>> >>>> Sadly, the dataset has been already acquired and passed to me for >>>> analysis. Therefore, I would be especially interested in how to >>>> correct for these effects post-acquisition if possible. >>>> >>>> When you talk about Jacobian modulation, do you mean using the >>>> "Realign and Unwarp" dialog and setting the Jacobian deformations >>>> option to Yes? If I do that and leave the other settings as default, >>>> the unwarping fails and I get the following error: >>>> >>>> >>>> Unwarp terminated abnormally. >>>> Failed 'Realign & Unwarp' >>>> Error using size >>>> Too many output arguments. >>>> In file "/Users/tozzil/Documents/MATLAB/Apps/spm12/spm_uw_estimate.m" >>>> (v6456), function "uwAtA2" at line 630. >>>> In file "/Users/tozzil/Documents/MATLAB/Apps/spm12/spm_uw_estimate.m" >>>> (v6456), function "build_AtA" at line 580. >>>> In file "/Users/tozzil/Documents/MATLAB/Apps/spm12/spm_uw_estimate.m" >>>> (v6456), function "spm_uw_estimate" at line 353. >>>> In file >>>> "/Users/tozzil/Documents/MATLAB/Apps/spm12/config/spm_run_realignunwarp.m" >>>> (v6554), function "spm_run_realignunwarp" at line 90. >>>> >>>> The following modules did not run: >>>> Failed: Realign & Unwarp >>>> >>>> >>>> Are there any settings that should be manually changed for this >>>> procedure? >>>> >>>> Thank you very much, >>>> >>>> >>>> >>>> >>>> Leonardo Tozzi, MD >>>> >>>> PhD Student >>>> >>>> Trinity College Institute of Neuroscience >>>> >>>> College Green, Dublin 2 >>>> >>>> Ireland >>>> >>>> ------------------------------------------------------------------------ >>>> *From:* Torben Lund <[log in to unmask] >>>> <mailto:[log in to unmask]> <mailto:[log in to unmask]>> >>>> *Sent:* Monday, June 27, 2016 10:42:50 AM >>>> *To:* Leonardo Tozzi >>>> *Cc:* [log in to unmask] >>>> <mailto:[log in to unmask]> <mailto:[log in to unmask]> >>>> *Subject:* Re: [SPM] Hyperintensity artefact close to edge of field of >>>> view >>>> >>>> Dear Leonardo >>>> >>>> It looks like signal pile-up near air tissue interfaces due to B0 >>>> inhomogenieties. You can reduce this at the acquisition stage by using >>>> in-plane acceleration e.g. GRAPPA, smaller voxels or different >>>> angulation of the slices. Modulation with the jacobian during unwrap >>>> might reduce this. >>>> >>>> Best >>>> Torben >>>> >>>> >>>> Sendt fra min iPhone >>>> >>>> Den 27. jun. 2016 kl. 08.43 skrev Leonardo Tozzi <[log in to unmask] >>>> <mailto:[log in to unmask]> >>>> <mailto:[log in to unmask]>>: >>>> >>>>> Dear experts, >>>>> >>>>> I am currently working on a task fMRI dataset (TR: 2s, TE: 35ms, >>>>> voxel size: 3x3 mm2, 4.8 mm slice thickness, 550 dynamic scans, 23 >>>>> slices per volume). >>>>> While preprocessing the data on SPM, I noticed that it shows a >>>>> hyperintensity in the anterior cingulate cortex and inferior temporal >>>>> lobes, close to the edge of the field of view. This looks like an >>>>> artefact to me: it seems to be common across subjects (see screenshot >>>>> 1: 5 subjects after warping at volume 1) and within the time series >>>>> of each subject (see screenshot 2, one subject volumes 100:108). >>>>> I was wondering if someone had any suggestions on how to correct this >>>>> artefact (if possible) and, otherwise, some insight on how it might >>>>> impact the analysis of task related activity using a GLM. In >>>>> particular, I am concerned it might affect the detection of task >>>>> effects in the subgenual portion of the ACC, especially after >>>>> smoothing. >>>>> >>>>> Thank you for your help. >>>>> Yours faithfully, >>>>> >>>>> >>>>> Leonardo Tozzi, M.D. >>>>> >>>>> /PhD Candidate/ >>>>> /Discipline of Psychiatry >>>>> Trinity College Institute of Neuroscience/ >>>>> /College Green >>>>> Dublin 2/ >>>>> /Ireland/ >>>>> / >>>>> / >>>>> /Office: +35318964234/ >>>>> /Mobile: //+353851026166/ >>>>> <1.png><2.png> >> >> -- >> Guillaume Flandin, PhD >> Wellcome Trust Centre for Neuroimaging >> University College London >> 12 Queen Square >> London WC1N 3BG > -- Guillaume Flandin, PhD Wellcome Trust Centre for Neuroimaging University College London 12 Queen Square London WC1N 3BG