How to map DWI measures from native volume space to standard surface space?

[Yang Hu]

Dear experts,

I would like to map DWI-derived measures (e.g., NDI from NODDI modeling) from the native T1 volume space to the fsLR 32k surface space, and then extract ROI-wise mean values based on the HCP-MMP1 atlas. I followed the approach described in the following paper and its associated code (https://github.com/RIKEN-BCIL/NoddiSurfaceMapping):

Fukutomi, H., Glasser, M. F., Zhang, H., Autio, J. A., Coalson, T. S., Okada, T., Togashi, K., Van Essen, D. C., & Hayashi, T. (2018). Neurite imaging reveals microstructural variations in human cerebral cortical gray matter. NeuroImage, 182, 488–499. https://doi.org/10.1016/j.neuroimage.2018.02.017

I have a few questions and would greatly appreciate your advice:

1. Is the myelin-style resampling method still recommended for mapping DWI measures (such as NDI, ODI, etc.) to the cortical surface?
2. How should the FWHM smoothing parameter be chosen when the voxel size of my data differs from that of the HCP-YA dataset? Currently, I use 2.5 times the voxel size, as in the original code.

For clarity, here is the code I am currently using (using NDI as an example):

## Map from native to MNI space
wb_command -volume-resample NDI.nii.gz \
            ./MNINonLinear/T1w_restore.nii.gz \
            CUBIC NDI_MNI.nii.gz \
            -warp ./MNINonLinear/xfms/acpc_dc2standard.nii.gz \
            -fnirt ./T1w/T1w_acpc_dc_restore.nii.gz
## Extract the cortical ribbon
fslmaths ./MNINonLinear/ribbon.nii.gz -thr 3 -uthr 3 -bin ribbon_L.nii.gz
fslmaths ./MNINonLinear/ribbon.nii.gz -thr 42 -uthr 42 -bin ribbon_R.nii.gz
## Map from MNI volume to native surface
DiffRes="`fslval ./T1w/Diffusion/data.nii.gz pixdim1 | awk '{printf "%0.2f",$1}'`"
NODDIMappingFWHM="`echo "$DiffRes * 2.5" | bc -l`"
NODDIMappingSigma=`echo "$NODDIMappingFWHM / ( 2 * ( sqrt ( 2 * l ( 2 ) ) ) )" | bc -l`
for curr_hemi in L R
do
  wb_command -volume-to-surface-mapping NDI_MNI.nii.gz \
                ./MNINonLinear/Native/${curr_sub}.${curr_hemi}.midthickness.native.surf.gii  \
                ${curr_hemi}.NDI.native.func.gii \
                -myelin-style ribbon_${curr_hemi}.nii.gz \
                ./MNINonLinear/Native/${curr_sub}.${curr_hemi}.thickness.native.shape.gii "$NODDIMappingSigma"
done
## Map from native surface to fsLR 32k space
for curr_hemi in L R
do
  wb_command -metric-resample ${curr_hemi}.NDI.native.func.gii \
                ./MNINonLinear/Native/${curr_sub}.${curr_hemi}.sphere.MSMAll.native.surf.gii \
                ./MNINonLinear/fsaverage_LR32k/${curr_sub}.${curr_hemi}.sphere.32k_fs_LR.surf.gii \
                ADAP_BARY_AREA ${curr_hemi}.NDI.MSMAll.fsLR_32k.func.gii \
                -area-surfs ./MNINonLinear/Native/${curr_sub}.${curr_hemi}.midthickness.native.surf.gii \
                ./MNINonLinear/fsaverage_LR32k/${curr_sub}.${curr_hemi}.midthickness_MSMAll.32k_fs_LR.surf.gii \
                -current-roi ./MNINonLinear/Native/${curr_sub}.${curr_hemi}.roi.native.shape.gii
  wb_command -metric-mask ${curr_hemi}.NDI.MSMAll.fsLR_32k.func.gii \
                ./MNINonLinear/fsaverage_LR32k/${curr_sub}.${curr_hemi}.atlasroi.32k_fs_LR.shape.gii \
                ${curr_hemi}.NDI.MSMAll.fsLR_32k.func.gii
done

Thank you in advance for your time and expertise!

Best regards,

Yang Hu

[Glasser, Matthew]

1. Ideally, we would have some kind of partial volume corrected volume to surface mapping; however, we have not yet developed that.
2. If you just are going to average within parcels, no smoothing is needed.

Matt.

--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/6c400b78-6c1b-4dd3-a22c-6d591a4aae94n%40humanconnectome.org.

 

---

The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.

[Yang Hu]

Thank you for the quick reply.

Regarding point 2, I realize I may have used the wrong term. The FWHM parameter is not for spatial smoothing, but for weighting voxels within the cylinder in myelin-style mapping.

If my final goal is to average within parcels, should this parameter still be set to a specific value (e.g., 2.5 × voxel size), or can it be adjusted? Or does it matter less when the final step is parcel-wise averaging?

Thanks again for your help.

[Glasser, Matthew]

I see.  I would probably keep that as is.

[Takuya Hayashi]

Thank you for raising this issue. We chose myelin-style mapping because midthickness-weighted averaging better reduces partial volume effects than ribbon-constrained mapping. However, its effectiveness depends on 'resolution relative to cortical thickness': a sigma smaller than the diffusion voxel size is not meaningful (note that diffusion data commonly has a larger voxel size than structural T1w/T2w images). Additionally, NODDI also accounts for partial volume effects during model fitting. So, we parameterized the FWHM as 2.5× the voxel size rather than using a fixed cortical thickness-based value. We've evaluated the partial volume effect and found it likely negligible (see supplements),but further validation would be valuable. A principled partial volume corrected mapping, as Matt suggests, would therefore be a welcome development. Regarding smoothing, I agree it is not strictly necessary for parcel-wise analyses; the smoothing in our code was intended for vertex-wise analyses. 

Takuya

---

差出人: Glasser, Matthew <glas...@wustl.edu>
送信日時: 2026年2月19日 23:26
宛先: Yang Hu <huyan...@gmail.com>; HCP-Users <hcp-...@humanconnectome.org>
CC: Takuya Hayashi <takuya....@riken.jp>
件名: Re: [hcp-users] How to map DWI measures from native volume space to standard surface space?