Differences between MNINonLinear and MNINonLinear/Native

101 views
Skip to first unread message

lei wei

unread,
Oct 20, 2023, 5:44:09 AM10/20/23
to HCP-Users
Dear HCP expert:

I discovered that the file in the T1 structural image MNINonLinear/Native is the same as the file in MNINonLinear, but with different vertex numbers (~160k and 32k). I believe the term Native indicates that the image has already been aligned to the MNI space but retains the original/Native resolution. Is that correct? It seems slightly different from the manual description, I apologize if I misunderstood it. I projected a metric file onto the MNI_native surface and tried to project the metric file to a standard mesh to compare their differences. Can I use wb_command -surface-resample for the transformation? And which sphere file should I use for the transformation?

Best
leiwei

Glasser, Matt

unread,
Oct 20, 2023, 6:15:59 AM10/20/23
to hcp-...@humanconnectome.org

There are two HCP volume spaces, physical (${StudyFolder}/${Subject}/T1w) and MNI (${StudyFolder}/${Subject}/MNINonLinear).  There are three HCP surface meshes, Native (the original irregular tessellation), 164k_fs_LR (highres standard mesh), 32k_fs_LR (lowres standard mesh).  Volume space and mesh topology are completely independent.  What is it that you are trying to accomplish?

 

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 on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/39d27555-9123-44cf-903e-5b4c636ebacbn%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.

lei wei

unread,
Oct 20, 2023, 8:34:33 AM10/20/23
to hcp-...@humanconnectome.org
Thanks for your reply !
I have followed the code of your paper "Neurite imaging reveals microstructural variations in human cerebral cortical gray matter" published in NeuroImage 2018. I found the NODDI-derived maps were first transform to the MNI space and then projected on the surface in MNINonlinear/Native folder (in code it is "AtlasspaceNativeFolder/"$Subject"."$Hemisphere".midthickness.native.surf.gii ").  Meshes in this folder have different number of vertices between each subject, but all of them have already aligned to the MNI space, I'm confused of this files and tried to transform the metrics file to the 164k and 32k standard space. 

Glasser, Matt <glas...@wustl.edu> 于2023年10月20日周五 18:15写道:

Glasser, Matt

unread,
Oct 20, 2023, 8:39:30 AM10/20/23
to hcp-...@humanconnectome.org, Takuya Hayashi

I have added Takuya who can comment on the exact methods he used.

lei wei

unread,
Oct 20, 2023, 9:28:33 AM10/20/23
to hcp-...@humanconnectome.org
Really thanks for your help !!
I have presented a point cloud picture about the surface in the MNINonLinear/Native folder and MNINonLinear folder. 
It shows that the two surfaces were exactly in the same space but with different numbers of vertices.  It seems not to belong to the three surface spaces ( Native (the original irregular tessellation), 164k_fs_LR (highres standard mesh), 32k_fs_LR (lowres standard mesh).  I think if I project the diffusion map onto the 32K or 164K standard surface, the files would  be transformed into standard space. And that is what I want. 
Screenshot from 2023-10-20 08-53-51.png
But I am still very confused about the surface file in the MNINonLinear/Native folder, the description of these files from manual or the other answers did not match their properties. I'm not sure if I have misunderstood some key points,  could you give me any idea about these files?   Really appreciate your help !
Thanks !

Glasser, Matt <glas...@wustl.edu> 于2023年10月20日周五 20:39写道:
Screenshot from 2023-10-20 08-53-51.png

Takuya Hayashi

unread,
Oct 20, 2023, 1:45:28 PM10/20/23
to Glasser, Matt, hcp-...@humanconnectome.org
Hi leiwei,

Please try using '-metric-resample' to resample metrics in native mesh to those in 164k or 32k.

Takuya

差出人: Glasser, Matt <glas...@wustl.edu>
送信日時: 2023年10月20日 21:39
宛先: hcp-...@humanconnectome.org <hcp-...@humanconnectome.org>
CC: Takuya Hayashi <takuya....@riken.jp>
件名: Re: [hcp-users] Differences between MNINonLinear and MNINonLinear/Native
 

lei wei

unread,
Oct 20, 2023, 1:48:29 PM10/20/23
to hcp-...@humanconnectome.org
Thanks !  I will try to do it.
leiwei

Takuya Hayashi <takuya....@riken.jp> 于2023年10月21日周六 01:45写道:

Tim Coalson

unread,
Oct 20, 2023, 7:42:18 PM10/20/23
to hcp-...@humanconnectome.org
Specifically, you will likely want to use the "MSMAll" native mesh sphere and whichever resolution of fs_LR standard sphere you want to use for cross-subject comparison.  Particularly if you are downsampling to 32k, it is recommended to use the ADAP_BARY_AREA method, in which case you should provide the matching subject midthickness surfaces (native mesh and MSMAll 32k).

If you used ribbon mapping to map your metric to the surface, be aware that particularly on native mesh, some vertices may have a value of zero because they are small, so you should use -metric-dilate on the raw mapping result (distance can be something small, like 1mm) before resampling it to anything else.  Ribbon mapping has a -bad-vertices-out option to tell you which vertices should be dilated into (to keep them separate from any real exact zeros in the volume), which you can pass as the -bad-vertex-roi to -metric-dilate.

Tim


Reply all
Reply to author
Forward
0 new messages