Dimension mismatch between surface datasets

AFNI Message Board
1

Hi,

I’m trying to map the functional data to surface using 3dVol2Surf and then extract the data within surface ROI. Then I met ** FATAL ERROR: Input and mask datasets are not same dimensions!

I checked their number of nodes and found there’s a lack of nodes for the functional data. It’s weird because this doesn’t happen to another run of this subject (scanned just 2 min before the problematic run). Any suggestions to solve this problem?

I also found the number of nodes of functional data varied if I made some changes to the options of @SUMA_AlignToExperiment which may affect the input for -sv of 3dVol2Surf. So something wrong with the alignment?

Thanks in advance!

Dset 0x199c9b0, datum level 0 (0…3 possible)
Number of Links: 0
Dset Name: SPARSE_DATA (-1), isGraph 0, isCIFTI (MultiDomain)0
filename: ./sbj04_rawdata_unblur_rh.niml.dset
label: //media/dongx/0D6…
self_idcode (idcode): AFN_XwpPE5QYr8xv4SQyEjDsTQ
geometry_parent_idcode:
domain_parent_idcode (MeshParent_idcode):
Node Index (NodeDef) Element:
inel->vec_len = 129537
inel->vec_num = 1
inel->vec_filled = 129537
Node Index Column found.
sorted_node_def: Yes
0, 1, …, 129651, 129652, 129653

----- Following are the scripts:

@SUMA_AlignToExperiment -exp_anat a2_brain_shiftrotate_al+orig
-surf_anat …/anat/SUMA/${s}_SurfVol.nii
-align_centers
-echo
-prefix ${s}_SurfVolal

for hemi in “lh” “rh”
do
3dVol2Surf -spec ${s}_${hemi}.spec -surf_A ${hemi}.smoothwm.gii -surf_B ${hemi}.pial.gii
-sv ${datapath}/${s}_SurfVolal+orig
-grid_parent ${datapath}/r08_bandpass.nii.gz
-map_func nzave -f_steps 10 -f_index voxels -out_niml ${datapath}/${s}func${hemi}.niml.dset
done

2

Hello,

Just to be sure, you are not using standard mesh surfaces for this? They allow for comparison across subjects.

3dVol2Surf allows for incomplete output, so that one can make small ROI files. Consequently, it is often good to add options:

   -oob_value    0.0      \
   -oom_value    0.0     \

That should make the output datasets complete in node coverage. Otherwise, if a mask or volume does not fully cover a surface, the result would be missing nodes.

  • rick
3

Thanks for your response! It’s really helpful.