''wb_command'' to convert nifti

496 views
Skip to first unread message

Geller Ross

unread,
May 28, 2021, 9:03:19 AM5/28/21
to HCP-Users

Dear HCP experts,

I’m following the instructions for install workbench v 1.5.0 in window
operating system.I‘m using ''wb_command'' to convert nifti asW_LQDO5S7F)}%[DU0TE7CW.png

but an error occurred in the generated file as]]$SAIC2H[](RYSGTU2}TB.png

Coalson, Timothy Scott (S&T-Student)

unread,
May 28, 2021, 3:46:25 PM5/28/21
to HCP-Users
-cifti-convert does not result in a volume file, it puts the matrix of combined vertex and voxel data as-is into a nifti file, with arbitrary spatial dimensions (and thus there will not be any recognizable structure when viewed in a volume viewer).  I don't know why this would cause MRIcron to divide by zero (maybe because the z dimension is length 1?), but you could try -smaller-dims instead of smaller-file, which chooses a near-cube for the spatial dimensions, and then you can see for yourself that the output of -cifti-convert does not organize anatomy the way software expects from a real volume file.  We have called this conversion "fake nifti", and its real purpose is to use surface-based cifti data in volume-based tools that don't require spatial neighborhood structure (I think melodic was one of the early uses for this conversion).

What are you wanting to do with the data?  For cortical data, there is usually a surface-based way to do what you want, which can be more accurate than volume-based approaches.

Tim

From: Geller Ross <7277...@qq.com>
Sent: Friday, May 28, 2021 8:03 AM
To: HCP-Users <hcp-...@humanconnectome.org>
Subject: [hcp-users] ''wb_command'' to convert nifti
 
--
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/8a34da1e-c12d-47fd-909a-c8f6e3d79cefn%40humanconnectome.org.

薄衫少年

unread,
May 28, 2021, 11:43:18 PM5/28/21
to hcp-users
I'm going to do Psychophysiological Interactions,So I need to convert to nifti.


------------------ 原始邮件 ------------------
发件人: "hcp-users" <tsc...@mst.edu>;
发送时间: 2021年5月29日(星期六) 凌晨3:46
收件人: "HCP-Users"<hcp-...@humanconnectome.org>;
主题: Re: [hcp-users] ''wb_command'' to convert nifti

-cifti-convert does not result in a volume file, it puts the matrix of combined vertex and voxel data as-is into a nifti file, with arbitrary spatial dimensions (and thus there will not be any recognizable structure when viewed in a volume viewer).  I don't know why this would cause MRIcron to divide by zero (maybe because the z dimension is length 1?), but you could try -smaller-dims instead of smaller-file, which chooses a near-cube for the spatial dimensions, and then you can see for yourself that the output of -cifti-convert does not organize anatomy the way software expects from a real volume file.  We have called this conversion "fake nifti", and its real purpose is to use surface-based cifti data in volume-based tools that don't require spatial neighborhood structure (I think melodic was one of the early uses for this conversion).

What are you wanting to do with the data?  For cortical data, there is usually a surface-based way to do what you want, which can be more accurate than volume-based approaches.

Tim

From: Geller Ross <7277...@qq.com>
Sent: Friday, May 28, 2021 8:03 AM
To: HCP-Users <hcp-...@humanconnectome.org>
Subject: [hcp-users] ''wb_command'' to convert nifti
 

Dear HCP experts,

I’m following the instructions for install workbench v 1.5.0 in window

operating system.I‘m using ''wb_command'' to convert nifti as

but an error occurred in the generated file as

--
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/8a34da1e-c12d-47fd-909a-c8f6e3d79cefn%40humanconnectome.org.

--
You received this message because you are subscribed to a topic in the Google Groups "HCP-Users" group.
To unsubscribe from this topic, visit https://groups.google.com/a/humanconnectome.org/d/topic/hcp-users/skQHRKT-pYs/unsubscribe.
To unsubscribe from this group and all its topics, 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/SN6PR01MB499178A9F7C0B38735F90ECB81229%40SN6PR01MB4991.prod.exchangelabs.com.
7A38A5C3@E837F11C.50B8B160.jpg
F56B416F@4FDF8249.50B8B160.jpg

Geller Ross

unread,
May 29, 2021, 1:13:12 AM5/29/21
to HCP-Users, tsc...@mst.edu
l was tried -smaller-dims, This is the result follow)0V2TUD[0]}AY{7UNR%Y~KV.png

Coalson, Timothy Scott (S&T-Student)

unread,
May 29, 2021, 5:10:52 PM5/29/21
to Geller Ross, HCP-Users
As I said, this command will not give you something that looks like a normal scan, it isn't supposed to, it is still vertex-based data, which will always look "wrong" when displayed as voxels.

What are you wanting to do with the data?

Tim

From: Geller Ross <7277...@qq.com>
Sent: Saturday, May 29, 2021 12:13 AM
To: HCP-Users <hcp-...@humanconnectome.org>
Cc: Coalson, Timothy Scott (S&T-Student) <tsc...@mst.edu>
Subject: Re: [hcp-users] ''wb_command'' to convert nifti
 

Geller Ross

unread,
May 30, 2021, 11:22:48 PM5/30/21
to HCP-Users, tsc...@mst.edu, Geller Ross

  • I want to use this data to compute psychophysiological interactions

Coalson, Timothy Scott (S&T-Student)

unread,
Jun 1, 2021, 7:48:34 PM6/1/21
to Geller Ross, HCP-Users
You can do correlation and regression without needing to know neighborhood information, so this should be a valid use for -cifti-convert.

The basic idea is:
1) -cifti-convert -to-nifti to dump the vertexwise data into a format that tools think is a volume file
2) run some tool written for volume files that doesn't need neighborhood information (do not apply smoothing, gradient, or other spatial operations, because the neighbor relations are wrong while in this format)
3) convert the result back to cifti with -cifti-convert -from-nifti

Tim

From: Geller Ross <7277...@qq.com>
Sent: Sunday, May 30, 2021 10:22 PM
To: HCP-Users <hcp-...@humanconnectome.org>
Cc: Coalson, Timothy Scott (S&T-Student) <tsc...@mst.edu>; Geller Ross <7277...@qq.com>

Glasser, Matthew

unread,
Jun 1, 2021, 8:15:55 PM6/1/21
to hcp-...@humanconnectome.org, Geller Ross

There were some posts on the FSL list about this.  I think FSL PPI uses film_gls for the first level, so that part needs to be CIFTI or CIFTI separated into GIFTI and NIFTI.

 

Matt.

l was tried -smaller-dims, This is the result follow

在2021年5月29日星期六 UTC+8 上午3:46:25<tsc...@mst.edu> 写道：

-cifti-convert does not result in a volume file, it puts the matrix of combined vertex and voxel data as-is into a nifti file, with arbitrary spatial dimensions (and thus there will not be any recognizable structure when viewed in a volume viewer).  I don't know why this would cause MRIcron to divide by zero (maybe because the z dimension is length 1?), but you could try -smaller-dims instead of smaller-file, which chooses a near-cube for the spatial dimensions, and then you can see for yourself that the output of -cifti-convert does not organize anatomy the way software expects from a real volume file.  We have called this conversion "fake nifti", and its real purpose is to use surface-based cifti data in volume-based tools that don't require spatial neighborhood structure (I think melodic was one of the early uses for this conversion).

 

What are you wanting to do with the data?  For cortical data, there is usually a surface-based way to do what you want, which can be more accurate than volume-based approaches.

 

Tim

 

From: Geller Ross <7277...@qq.com>
Sent: Friday, May 28, 2021 8:03 AM
To: HCP-Users <hcp-...@humanconnectome.org>
Subject: [hcp-users] ''wb_command'' to convert nifti

 

Dear HCP experts,

I’m following the instructions for install workbench v 1.5.0 in window

operating system.I‘m using ''wb_command'' to convert nifti as

but an error occurred in the generated file as

--
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/8a34da1e-c12d-47fd-909a-c8f6e3d79cefn%40humanconnectome.org.

--
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.

 

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.

Reply all
Reply to author
Forward
0 new messages