Hi all
Could anyone help me with this error:
MRISreadAnnotationIntoArray: vertex index out of range: 1499028058 i=5A595959, in_array_size=133467 annot file: /home/virtualuser/apps/freesurfer/subjects/s1/surf/../label/lh.aparc.annot
This appeared after trying to average my subject data. It says ' vertex index out of range '. All acquisitions were made using the same parameters so I don't know what this may mean.
Thanks in advance
Hi Leo, check to see if that subject's analysis is up-to-date. This usually happens when it was only partially re-run. doug
leonardo kay wrote:
Hi all
Could anyone help me with this error:
MRISreadAnnotationIntoArray: vertex index out of range: 1499028058 i=5A595959, in_array_size=133467 annot file: /home/virtualuser/apps/freesurfer/subjects/s1/surf/../label/lh.aparc.annot
This appeared after trying to average my subject data. It says ' vertex index out of range '. All acquisitions were made using the same parameters so I don't know what this may mean.
Thanks in advance
-- Leo Kay.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Dear all, I am combining 3 runs of a subject using fixed effects GLM and I wanted to make sure I am doing the right thing.
For each subject I use: mri_glmfit --y 3runs/lh.cope1.mgh --yffxvar 3runs/lh.varcope1.mgh --ffxdof 126 --osgm --glmdir 3runs/lh.osgm.ffx --surf fsaverage lh --label $SUBJECTS_DIR/fsaverage/label/lh.cortex.label --fwhm 5
, while lh.cope1.mgh contains the concatenated cope1 images of the same subject in fsaverage space (the same for varcope1).
1) Is it the right way? 2) I took DOF from subjectX.feat/stats/dof of one of the runs. Is it correct? 3) The functional data has not been smoothed in the 1st level analysis in FSL (as recommended), and also not smoothed during sampling the copes to a common space. Therefore I want to smooth it here for the first time, but only with 5mm. Does it sound right?
Cheers, Aga
Hi Aga, your commands look right. For the DOF, it should be the sum of the DOFs from all the runs (probably won't make much of a difference). Smoothing is a bit of an issue when you want to look at individual results. Technically, you should smooth before you do the first level analysis (ie, before your compute the varcope), but this would require doing the FEAT analysis directly on surface data. Smoothing after computing the varcope means that the varcope will not be accurate (it will be too large). The penalty is that you will see less activation than you should. At the group level, this is not such a big deal because you're either not using the varcope or you are using it as a weight. doug
Agnieszka Burzynska wrote:
Dear all, I am combining 3 runs of a subject using fixed effects GLM and I wanted to make sure I am doing the right thing.
For each subject I use: mri_glmfit --y 3runs/lh.cope1.mgh --yffxvar 3runs/lh.varcope1.mgh --ffxdof 126 --osgm --glmdir 3runs/lh.osgm.ffx --surf fsaverage lh --label $SUBJECTS_DIR/fsaverage/label/lh.cortex.label --fwhm 5
, while lh.cope1.mgh contains the concatenated cope1 images of the same subject in fsaverage space (the same for varcope1).
- Is it the right way?
- I took DOF from subjectX.feat/stats/dof of one of the runs. Is it
correct? 3) The functional data has not been smoothed in the 1st level analysis in FSL (as recommended), and also not smoothed during sampling the copes to a common space. Therefore I want to smooth it here for the first time, but only with 5mm. Does it sound right?
Cheers, Aga
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Dear Doug, Thank you so much. I completely see you point, but I have re-run the 1st level feat without smoothing just because it has been recommended not to smooth in the volume and then transfer it onto the surface, but rather first smooth on the surface (http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/FslFeatFreeSurfer).
What I plan to do in the end is to include the cortical thickness as the vertex-wise covariate in the functional group analysis.
So the final analysis will be on the group level, where, as you say, the varcopes should not matter that much.
However, I was also thinking of analyzing varcopes in addition to copes (group analysis) to relate the variance of BOLD signal to the thickness.
Would you then recommend going back and taking the initial 1st level analysis with regular smoothing?
Thank you! Aga
On 7/7/11 4:43 PM, "Douglas N Greve" greve@nmr.mgh.harvard.edu wrote:
Hi Aga, your commands look right. For the DOF, it should be the sum of the DOFs from all the runs (probably won't make much of a difference). Smoothing is a bit of an issue when you want to look at individual results. Technically, you should smooth before you do the first level analysis (ie, before your compute the varcope), but this would require doing the FEAT analysis directly on surface data. Smoothing after computing the varcope means that the varcope will not be accurate (it will be too large). The penalty is that you will see less activation than you should. At the group level, this is not such a big deal because you're either not using the varcope or you are using it as a weight. doug
Agnieszka Burzynska wrote:
Dear all, I am combining 3 runs of a subject using fixed effects GLM and I wanted to make sure I am doing the right thing.
For each subject I use: mri_glmfit --y 3runs/lh.cope1.mgh --yffxvar 3runs/lh.varcope1.mgh --ffxdof 126 --osgm --glmdir 3runs/lh.osgm.ffx --surf fsaverage lh --label $SUBJECTS_DIR/fsaverage/label/lh.cortex.label --fwhm 5
, while lh.cope1.mgh contains the concatenated cope1 images of the same subject in fsaverage space (the same for varcope1).
- Is it the right way?
- I took DOF from subjectX.feat/stats/dof of one of the runs. Is it
correct? 3) The functional data has not been smoothed in the 1st level analysis in FSL (as recommended), and also not smoothed during sampling the copes to a common space. Therefore I want to smooth it here for the first time, but only with 5mm. Does it sound right?
Cheers, Aga
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The problem with doing the smoothing at the first level in FEAT is that it will be volume-based, not surface-based. I'm not sure what to tell you about using the varcopes. They will certainly be very noisy without smoothing, so probably smoothing is a good idea, even if it's volume-based.
doug
Agnieszka Burzynska wrote:
Dear Doug, Thank you so much. I completely see you point, but I have re-run the 1st level feat without smoothing just because it has been recommended not to smooth in the volume and then transfer it onto the surface, but rather first smooth on the surface (http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/FslFeatFreeSurfer).
What I plan to do in the end is to include the cortical thickness as the vertex-wise covariate in the functional group analysis.
So the final analysis will be on the group level, where, as you say, the varcopes should not matter that much.
However, I was also thinking of analyzing varcopes in addition to copes (group analysis) to relate the variance of BOLD signal to the thickness.
Would you then recommend going back and taking the initial 1st level analysis with regular smoothing?
Thank you! Aga
On 7/7/11 4:43 PM, "Douglas N Greve" greve@nmr.mgh.harvard.edu wrote:
Hi Aga, your commands look right. For the DOF, it should be the sum of the DOFs from all the runs (probably won't make much of a difference). Smoothing is a bit of an issue when you want to look at individual results. Technically, you should smooth before you do the first level analysis (ie, before your compute the varcope), but this would require doing the FEAT analysis directly on surface data. Smoothing after computing the varcope means that the varcope will not be accurate (it will be too large). The penalty is that you will see less activation than you should. At the group level, this is not such a big deal because you're either not using the varcope or you are using it as a weight. doug
Agnieszka Burzynska wrote:
Dear all, I am combining 3 runs of a subject using fixed effects GLM and I wanted to make sure I am doing the right thing.
For each subject I use: mri_glmfit --y 3runs/lh.cope1.mgh --yffxvar 3runs/lh.varcope1.mgh --ffxdof 126 --osgm --glmdir 3runs/lh.osgm.ffx --surf fsaverage lh --label $SUBJECTS_DIR/fsaverage/label/lh.cortex.label --fwhm 5
, while lh.cope1.mgh contains the concatenated cope1 images of the same subject in fsaverage space (the same for varcope1).
- Is it the right way?
- I took DOF from subjectX.feat/stats/dof of one of the runs. Is it
correct? 3) The functional data has not been smoothed in the 1st level analysis in FSL (as recommended), and also not smoothed during sampling the copes to a common space. Therefore I want to smooth it here for the first time, but only with 5mm. Does it sound right?
Cheers, Aga
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Hi, During running fixed effects glm to combine runs, the analysis of all but one subject went fine.
The log file contains this error message: mri_surf2surf --srcsubject 1250 --srchemi lh --srcsurfreg sphere.reg --trgsubject fsaverage --trghemi lh --trgsurfreg sphere.reg --tval 3runs/tmp.mris_preproc.63877/1250.1.mgh --sval 3runs/tmp.mris_preproc.63877/subjsurfvals.mgh --noreshape --no-cortex ERROR: dimension inconsistency in source data Number of surface vertices = 132315 Number of value vertices = 132299
I see that mri_vol2surf --src run1.feat/stats/varcope1.nii.gz --srcreg run1.feat/reg/freesurfer/anat2exf.register.dat --hemi lh --out 3runs/tmp.mris_preproc.63877/subjsurfvals.mgh --noreshape
Results in creating a file with dimensions 132299 1 1.
Which data source data has then 132315 vertices? How can I correct this?
Thank you for your help!
Best, Aga
freesurfer@nmr.mgh.harvard.edu
-
Agnieszka Burzynska -
Douglas N Greve -
leonardo kay