The mismatch is because the medial wall is not excluded from that particular 
file.  The HCP data files (e.g. dense timeseries files) have the medial wall 
excluded.

Peace,

Matt.

From: 
<[email protected]<mailto:[email protected]>>
 on behalf of Timothy Coalson <[email protected]<mailto:[email protected]>>
Date: Friday, May 13, 2016 at 3:09 PM
To: Aaron C <[email protected]<mailto:[email protected]>>
Cc: "[email protected]<mailto:[email protected]>" 
<[email protected]<mailto:[email protected]>>
Subject: Re: [HCP-Users] Some questions about the usage of CIFTI files

Inline replies.

Tim


On Fri, May 13, 2016 at 1:20 PM, Aaron C 
<[email protected]<mailto:[email protected]>> wrote:
Dear HCP experts,

I have some questions about the usage of CIFTI files.

1. I checked the number of grayordinates of the following files using the 
command “wb_command -file-information”, and I noticed some of them have 
different number of grayordinates:

“Parcels_LR.dlabel.nii” from Gordon et al.’s parcellation: 64984 grayordinates 
(32492 left cortex vertices and 32492 right cortex vertices).
“100206.aparc.a2009s.32k_fs_LR.dlabel.nii”: 59412 grayordinates (29696 left 
cortex vertices and 29716 right cortex vertices)
“100206.MyelinMap.32k_fs_LR.dscalar.nii”: 59412 grayordinates (29696 left 
cortex vertices and 29716 right cortex vertices)
“HCP_S900_820_rfMRI_MSMAll_groupPCA_d4500ROW_zcorr.dconn.nii”: 91282 
grayordinates (29696 left cortex vertices, 29716 right cortex vertices and 
31870 subcortical voxels)

Probably the easiest way to fix this is to use 
-cifti-create-dense-from-template to make cifti files that do have matching 
indices.  Modifying a dconn is a bit trickier, and will also need to involve 
-cifti-transpose and -cifti-copy-mapping, and may use a lot of memory, so you 
may simply want to use the 91282 space as the template.

If two CIFTI files are with different number of grayordinates (e.g., 
“Parcels_LR.dlabel.nii” and “100206.MyelinMap.32k_fs_LR.dscalar.nii”), when 
both loaded into the MATLAB using "ciftiopen", then the first file will be just 
a 64984x1 array, and the second file will be just a 59412x1 array, but there is 
no other information about the correspondence of the same grayordinate between 
these two files. My question is how could I locate the same brain grayordinate 
from them?

You can also use -cifti-export-dense-mapping to get the mapping from matrix 
index to vertex/voxel + structure, but this is likely to be more involved.  
However, since you are interested in spatial neighbor information, you'll need 
this to figure that out.

2. I have the time series of only a subset of cortical surface vertices in a 
MATLAB matrix (e.g., a 59412x1200 matrix). How could I save the time series of 
the subset of cortical surface vertices in a CIFTI *.dtseries.nii file? The 
header of existing CIFTI *.dtseries.nii file cannot be used directly since I 
didn’t find such *.dtseries.nii file with a dimensionality of 59412x1200.

Use the ciftisavereset function from 
https://wiki.humanconnectome.org/display/PublicData/HCP+Users+FAQ#HCPUsersFAQ-2.HowdoyougetCIFTIfilesintoMATLAB?

Similarly, I also have the pairwise correlations between only a subset of 
cortical surface vertices in a MATLAB matrix (e.g., just a 59412x59412 matrix, 
not like a 91282x91282 matrix which also includes subcortical voxels). How 
could I save it in a CIFTI *.dconn.nii file? The header of existing CIFTI 
*.dconn.nii file cannot be used directly since I didn’t find such *.dconn.nii 
file with a dimensionality of 59412x59412.

dconn saving from matlab is not particularly well supported at the moment, 
unfortunately.  What you could do is to save it with ciftisavereset to a 
.dtseries.nii name, and then use -cifti-copy-mapping to change the series 
mapping to the same dense mapping.  The fieldtrip-derived cifti functions have 
some support for symmetric dconns, but have only recently started trying to 
preserve the excluded vertices information in the saved file (and not by 
default).

3. How could I obtain surface vertex adjacencies of the fs_LR 32k surface using 
wb_command? I am looking for some commands in Connectome Workbench equivalent 
to the command “caret_command -surface-topology-neighbors” in the Caret 
software.

We don't currently have a command for that in wb_command.  You can load the 
gifti surface in matlab, and use the triangles array to make the information 
directly (all vertices in a single triangle are neighbors).  Most spatial 
things workbench does use geodesic distances, rather than immediate neighbors.  
Immediate neighbors are not the same distance in different parts of the 
surface, or even for a single vertex.  There is a command to produce geodesic 
distances from one vertex to all vertices (with an optional distance limit), 
though we haven't made an all vertices to all vertices command like caret5 has 
(we stopped using matlab to do spatial computations, as it was too slow for our 
purposes).

What kind of spatial things are you wanting to do?  We have commands for some 
of the more common spatial operations.

Thank you very much!


_______________________________________________
HCP-Users mailing list
[email protected]<mailto:[email protected]>
http://lists.humanconnectome.org/mailman/listinfo/hcp-users


_______________________________________________
HCP-Users mailing list
[email protected]<mailto:[email protected]>
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

________________________________
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.

_______________________________________________
HCP-Users mailing list
[email protected]
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Reply via email to