converting original mesh to standard mesh

AFNI Message Board
1

Dear AFNI and SUMA experts:

Here is my program which I believe creates the original mesh. After concatenation of runs, removal of TPs and doing tshifting, I do the following set of commands below: I wonder if you see anything obviously wrong??? Since I use +orig space anatomical volume and SurfVol, I am guessing the surface meshes are in original space–Is that correct? I wonder how they can be transformed to a standard mesh. I found these commands (see below) from the MapIcosehedron handout and it said you can use these commandlines to transform an original mesh to a given standard mesh. I am wondering if this is the typical way of doing things or if there are any alternatives? Can you use the result called std.141.SUMAfinal.niml.dset found from below in group data analysis such as 3dANOVA2? Obviously, I need all my subjects to be standard before doing group analysis. Anyway, I would appreciate any comments about this because if I don’t understand the definitions or am not sure of them in this case, I am in big trouble doing the calculations correctly. I wish all of you great Holidays!

3dWarp -deoblique -prefix scat_runtob cat_runt+orig

3dvolreg -dfile motionsuma -base ‘scat_runtob+orig[575]’ -prefix scat_runtob_reg scat_runtob+orig

align_epi_anat.py -anat Volumeob+orig -epi scat_runtob_reg+orig -epi_base 575 -epi_strip 3dAutomask -volreg off

@SUMA_AlignToExperiment -exp_anat Volumeob_al+orig -surf_anat SUMASubject06_SurfVol+orig -strip_skull surf_anat -atlas_followers -overwrite_resp S -prefix SUMASubject06_SurfVol_Alnd_Exp

3dVol2Surf -spec SUMASubject06_rh.spec -sv SUMASubject06_SurfVol_Alnd_Exp+orig -surf_A smoothwm -surf_B pial -f_index nodes -f_steps 10 -map_func ave -oob_value 0 -grid_parent scat_runtob_reg+orig -out_niml SUMASubj06VtoS.niml.dset

SurfSmooth -spec SUMASubject06_rh.spec -surf_A smoothwm -input SUMASubj06VtoS.niml.dset -met HEAT_07 -target_fwhm 6.0 -output SUMASubj06VtoSblurred.niml.dset

1d_tool.py -infile motionsuma -demean -write motionsuma.demean

3dDeconvolve -input SUMASubj06VtoSblurred.niml.dset -polort 2 -num_stimts 12
-stim_file 1 s_PULSE.1D -stim_label 1 “PULSE” -stim_maxlag 1 6 \
-stim_file 2 s_PPI.1D -stim_label 2 “PPI” -stim_maxlag 2 6
-stim_file 3 s_PPI2.1D -stim_label 3 “PPI2” -stim_maxlag 3 6
-stim_file 4 ‘motionsuma[1]’ -stim_label 4 “Roll”
-stim_base 4
-stim_file 5 ‘motionsuma[2]’ -stim_label 5 “Pitch”
-stim_base 5
-stim_file 6 ‘motionsuma[3]’ -stim_label 6 “Yaw”
-stim_base 6
-stim_file 7 ‘motionsuma[4]’ -stim_label 7 “IS”
-stim_base 7
-stim_file 8 ‘motionsuma[5]’ -stim_label 8 “RL”
-stim_base 8
-stim_file 9 ‘motionsuma[6]’ -stim_label 9 “AP”
-stim_base 9
-stim_file 10 ‘e_PulseSubject06.1D’ -stim_label 10 “e_Pulse”
-stim_file 11 ‘e_PPISubject06.1D’ -stim_label 11 “e_PPI”
-stim_file 12 ‘e_PPI2Subject06.1D’ -stim_label 12 “e_PPI2”
-concat concat.1D
-bucket DeconmotionSUMA
-fout -tout -vout -rout -bout

below lines find the chosen freesurfer parcellation in the lookup table and apply its mask

to the final .dset to localize any activity to that brain region.

3dcalc -a ‘rh.aparc.a2009s.annot.niml.dset[0]’ -expr ‘equals(a,9211105)’ -prefix myinsularoi
3dcalc -a myinsularoi.niml.dset -expr “step(a)” -prefix myinsulanewroi
3dcalc -a myinsulanewroi.niml.dset -b DeconmotionSUMA.niml.dset -expr “a*b” -prefix SUMAfinal.niml.dset

These are the applied lines from MapIcosahedron that I believe create the standard mesh for a 141 density using the above #SUMAfinal.niml.dset which was found from taking the deconvolution result (i.e. DeconmotionSUMA) after applying the

#freesurfer parcellation above.

SurfToSurf -i_fs std.141.rh.smoothwm.asc
-i_fs rh.smoothwm.asc
-prefix std.141.
-mapfile std.141.SUMASubject06_rh.niml.M2M
-dset SUMAfinal.niml.dset

2

Hello AFNI folks,
All I need is some reassurance that I am doing things correctly. There is no one here I could ask and I am learning everything by myself. Ultimately, I have a group of subjects that I need to analyze first at the subject level and map their functional to surface space to get the original mesh and then transform to standard mesh space where I can do group data analysis and find the mean across subjects of each BR regressor and cluster to find the most significant node/voxel to use in TMS. I appreciate any time you spend on my posting.
Thanks,
-Linda

3

Hi Linda,

It is not clear why you are not using afni_proc.py for
the first level analysis. To do so, add the surf block
and the -surf_anat and -surf_spec options to a typical
afni_proc.py volumetric analysis command, where the
spec file should be fore the standard mesh surfaces.

Ignoring that, consider using the standard mesh
surfaces directly in the 3dVol2Surf command. There
is probably no reason to use the original surfaces once
the standard mesh ones exist.

Also, 3dWarp -deoblique is is generally not necessary
before running 3dvolreg or align_epi_anat.py. Again,
it might be a very good idea to look at what afni_proc.py
would do.

Consider “Example 8” from the “afni_proc.py -help” output.

Also, consider the sample scripts from the class data,
including the afni_proc.py command script in
AFNI_data6/FT_analysis/s03.ap.surface,
as well as the corresponding analysis script,
AFNI_data6/FT_analysis/s13.proc.FT.surf.

Writing your own scripts is a great way to learn.
But it may be a good idea to try afni_proc.py for
the actual analysis. At least try it and compare.

  • rick
4

Thanks very much Rick!
I am currently looking into afni_proc.py for use for our problem.
I really appreciate your help!