Dear all,

I think I figured out the contrasts (I'm using [0 0 0 0 0.5 0.5 0 0]...is that correct?) but I've come across an error when trying to do the next step in the tutorial.

I didn't cache the data during recon-all, so I ran the following:

mris_preproc --fsgd MMSE-AD.fsgd --target fsaverage --hemi lh --meas thickness --out lh.MMSE-AD.thickness.00.mgh

It seemed to run smoothly (I've attached the log below) so I ran the next step:

mri_surf2surf --hemi lh --s fsaverage --sval lh.MMSE-AD.thickness.00.mgh --fwhm 10 --cortex --tval lh.MMSE-AD.thickness.10.mgh

I got the following error:

Reading source surface reg /home/virtualuser/freesurfer/subjects/fsaverage/surf/lh.sphere.reg Loading source data mghRead(/home/virtualuser/freesurfer/subjects/lh.MMSE-AD.thickness.00.mgh, -1): could not open file ERROR: could not read lh.MMSE-AD.thickness.00.mgh as type 3

Can anyone tell me where the problem is? Thanks!

Cheers, Linda

The log for mris_preproc (truncated to the last subject and last line):

--------------------------------------------------- #@# 12/12 avo1249 Fri Dec 2 18:20:04 HKT 2011 -------------- ----------------------- mri_surf2surf --srcsubject avo1249 --srchemi lh --srcsurfreg sphere.reg --trgsubject fsaverage --trghemi lh --trgsurfreg sphere.reg --tval ./tmp.mris_preproc.18275/avo1249.12.mgh --sval /home/virtualuser/freesurfer/subjects/avo1249/surf/lh.thickness --sfmt curv --noreshape --no-cortex Source registration surface changed to sphere.reg Target registration surface changed to sphere.reg srcsubject = avo1249 srcval = /home/virtualuser/freesurfer/subjects/avo1249/surf/lh.thickness srctype = curv trgsubject = fsaverage trgval = ./tmp.mris_preproc.18275/avo1249.12.mgh trgtype = srcsurfreg = sphere.reg trgsurfreg = sphere.reg srchemi = lh trghemi = lh frame = 0 fwhm-in = 0 fwhm-out = 0 label-src = (null) label-trg = (null) OKToRevFaceOrder = 1 Reading source surface reg /home/virtualuser/freesurfer/subjects/avo1249/surf/lh.sphere.reg Loading source data Reading curvature file /home/virtualuser/freesurfer/subjects/avo1249/surf/lh.thickness Reading target surface reg /home/virtualuser/freesurfer/subjects/fsaverage/surf/lh.sphere.reg Done Mapping Source Volume onto Source Subject Surface surf2surf_nnfr: building source hash (res=16). Surf2Surf: Forward Loop (163842)

surf2surf_nnfr: building target hash (res=16). Surf2Surf: Reverse Loop (121128) Reverse Loop had 15750 hits Surf2Surf: Dividing by number of hits (163842) INFO: nSrcLost = 0 nTrg121 = 151248, nTrgMulti = 12594, MnTrgMultiHits = 2.2506 nSrc121 = 78171, nSrcLost = 0, nSrcMulti = 42957, MnSrcMultiHits = 2.36099 Saving target data Saving to ./tmp.mris_preproc.18275/avo1249.12.mgh

----------------------- mri_concat ./tmp.mris_preproc.18275/avo0015.1.mgh ./tmp.mris_preproc.18275/avo0260.2.mgh ./tmp.mris_preproc.18275/avo0403.3.mgh ./tmp.mris_preproc.18275/avo0441.4.mgh ./tmp.mris_preproc.18275/avo0493.5.mgh ./tmp.mris_preproc.18275/avo0622.6.mgh ./tmp.mris_preproc.18275/avo0677.7.mgh ./tmp.mris_preproc.18275/avo0920.8.mgh ./tmp.mris_preproc.18275/avo1086.9.mgh ./tmp.mris_preproc.18275/avo1206.10.mgh ./tmp.mris_preproc.18275/avo1241.11.mgh ./tmp.mris_preproc.18275/avo1249.12.mgh --o lh.MMSE-AD.thickness.00.mgh

On 2 December 2011 11:42, Linda Zhang lzhang87@gmail.com wrote:

Dear all,

I have a bit of an amateur question. I've read the slides and tutorial for group analysis, but I'm still confused about the contrasts. I made my fsgd file, which contains two groups: M and F. I also have three continuous variables: Age, MMSE, Education. I would like to test whether thickness is correlated with MMSE scores, after controlling for gender, age, and education. What should my contrast file be in this case?

Thanks for the help!

Cheers, Linda

Hi Linda, you have two class varibles (M and F) and three continuous variables (Age, MMSE, and ED), so you'll have 2*(3+1)=8 regressors, so your contrast will have 6 elements. The order of the elements will be:

1. M-offset 2. F-offset 3. M-AgeSlope 4. F-AgeSlope 5. M-MMSE-Slope 6. F-MMSE-Slope 7. M-Ed-Slope 8. F-Ed-Slope

You are only interested in whether the MMSE changes with thickness, so you're only interested in components 5 and 6 (meaning that all others will be 0), so set your contrast to 0 0 0 0 .5 .5 0 0 The ".5 .5" means that the final value will be (0.5*M-MMSE-Slope + 0.5*F-MMSE-Slope) = average of the two slopes. The p-value will not change if you use 0 0 0 0 1 1 0 0. doug

Linda Zhang wrote:

Dear all,

I have a bit of an amateur question. I've read the slides and tutorial for group analysis, but I'm still confused about the contrasts. I made my fsgd file, which contains two groups: M and F. I also have three continuous variables: Age, MMSE, Education. I would like to test whether thickness is correlated with MMSE scores, after controlling for gender, age, and education. What should my contrast file be in this case?

Thanks for the help!

Cheers, Linda

---

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer