Re: [HCP-Users] Convert subject-specific ICA node maps to volumetric space?

[Nicola Toschi]

Hi Tim,

thanks a lot for your detailed answer and insight!

Best

Nicola


On 11/13/2018 9:06 PM, Timothy Coalson wrote:
-cifti-convert does not do this, it only dumps the matrix into 
different file formats as-is, the spatial relationships are not 
accessible from its output.


If you are only interested in subcortical/cerebellum data, that is 
trivial to extract from cifti as a volumetric nifti using 
-cifti-separate with -volume-all.  In the future, when resolution 
increases sufficiently, we hope to move the cerebellum data to surface 
representation also.


Group volume-registered space has significant drawbacks, in that it 
does not align cortical functional areas to the degree that 
surface/cifti processing can.  This is because much of the cortex has 
variable folding between subjects, which volume registration can't 
unfold to try to match, and because functional areas vary somewhat in 
their position relative to these folds.  See our recent paper:


https://www.ncbi.nlm.nih.gov/pubmed/29925602

This actually means that mapping the cifti data back into the volume 
should have better cross-subject functional correspondence than 
computing group maps from the original volumetric data - the drawback, 
however, is that the folding patterns can't be preserved (since they 
are different across subjects), and you will end up with a much less 
folded cortical ribbon with oddly sharp boundaries (and perfect 
correlation along the normal vector of the surface). Processing the 
data in cifti format will use less memory, be more correct to the 
nature of the data, and make it more obvious that you aren't really 
using a volumetric processing stream.  Matlab and python both have 
support for reading and writing cifti files (though in matlab there 
isn't good support for spatial relationships), as well as a few other 
languages (the cifti code from workbench has been made into a c++ 
library, though you will also need a gifti library to make full use of 
spatial relationships).  wb_command can do several types of spatial 
processing operations on cifti, and -cifti-convert will allow you to 
feed the data through virtually any tool that doesn't use spatial 
neighborhood information.


Tim


On Tue, Nov 13, 2018 at 10:45 AM, Nicola Toschi 
mailto:tos...@med.uniroma2.it>> wrote:


Thank you Steve,

I was hoping to get them out of the CIFTI versions (also so I
don't have to reprocess the 4D FIX data for all subjects). Maybe
with cifti-convert? The brain would not be fully covered here i
realise but much better than nothing.

Alternatively, do you think there is a chance of obtaining them
somehow (since they have been calculated by you guys anyway)? Can
i put in a request somewhere maybe?

Thank you in advance!

Nicola


On 11/13/2018 5:13 PM, Steve Smith wrote:

Hi - we calculated them in order to estimate the volumetric group
maps. However AFAIK we didn't bundle them in the PTN release.
You can easily get them though by regressing node timeseries into
4D FIX-cleaned data.

Cheers.



On 13 Nov 2018, at 16:09, Nicola Toschi mailto:tos...@med.uniroma2.it>> wrote:

Dear HCP list and experts,

I was wondering if the subject-specific node maps from the
latest PTN
release (which are available in CIFTI format) are also available in
volumetric space (which would aid the specific analysis we are
running).

Or if not, maybe a pointer on how to convert them?

Thank you very much in advance!

nicola

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus 

___
HCP-Users mailing list
HCP-Users@humanconnectome.org 
http://lists.humanconnectome.org/mailman/listinfo/hcp-users




---
Stephen M. Smith, Professor of Biomedical Engineering
Head of Analysis,  WIN (FMRIB) Oxford

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 610470
st...@fmrib.ox.ac.uk 
http://www.fmrib.ox.ac.uk/~steve 
---

Stop the cultural destruction of Tibet 













Virus-free. www.avast.com




<#m_-1125881958447918270_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

___
HCP-Users mailing list
HCP-Users@humanconnectome.org 

Re: [HCP-Users] Convert subject-specific ICA node maps to volumetric space?

[Timothy Coalson]

-cifti-convert does not do this, it only dumps the matrix into different
file formats as-is, the spatial relationships are not accessible from its
output.

If you are only interested in subcortical/cerebellum data, that is trivial
to extract from cifti as a volumetric nifti using -cifti-separate with
-volume-all.  In the future, when resolution increases sufficiently, we
hope to move the cerebellum data to surface representation also.

Group volume-registered space has significant drawbacks, in that it does
not align cortical functional areas to the degree that surface/cifti
processing can.  This is because much of the cortex has variable folding
between subjects, which volume registration can't unfold to try to match,
and because functional areas vary somewhat in their position relative to
these folds.  See our recent paper:

https://www.ncbi.nlm.nih.gov/pubmed/29925602

This actually means that mapping the cifti data back into the volume should
have better cross-subject functional correspondence than computing group
maps from the original volumetric data - the drawback, however, is that the
folding patterns can't be preserved (since they are different across
subjects), and you will end up with a much less folded cortical ribbon with
oddly sharp boundaries (and perfect correlation along the normal vector of
the surface).  Processing the data in cifti format will use less memory, be
more correct to the nature of the data, and make it more obvious that you
aren't really using a volumetric processing stream.  Matlab and python both
have support for reading and writing cifti files (though in matlab there
isn't good support for spatial relationships), as well as a few other
languages (the cifti code from workbench has been made into a c++ library,
though you will also need a gifti library to make full use of spatial
relationships).  wb_command can do several types of spatial processing
operations on cifti, and -cifti-convert will allow you to feed the data
through virtually any tool that doesn't use spatial neighborhood
information.

Tim


On Tue, Nov 13, 2018 at 10:45 AM, Nicola Toschi 
wrote:

> Thank you Steve,
>
> I was hoping to get them out of the CIFTI versions (also so I don't have
> to reprocess the 4D FIX data for all subjects). Maybe with cifti-convert?
> The brain would not be fully covered here i realise but much better than
> nothing.
>
> Alternatively, do you think there is a chance of obtaining them somehow
> (since they have been calculated by you guys anyway)? Can i put in a
> request somewhere maybe?
>
> Thank you in advance!
>
> Nicola
>
>
> On 11/13/2018 5:13 PM, Steve Smith wrote:
>
> Hi - we calculated them in order to estimate the volumetric group maps.
> However AFAIK we didn't bundle them in the PTN release.
> You can easily get them though by regressing node timeseries into 4D
> FIX-cleaned data.
>
> Cheers.
>
>
> On 13 Nov 2018, at 16:09, Nicola Toschi  wrote:
>
> Dear HCP list and experts,
>
> I was wondering if the subject-specific node maps from the latest PTN
> release (which are available in CIFTI format) are also available in
> volumetric space (which would aid the specific analysis we are running).
>
> Or if not, maybe a pointer on how to convert them?
>
> Thank you very much in advance!
>
> nicola
>
> ---
> This email has been checked for viruses by Avast antivirus software.
> https://www.avast.com/antivirus
>
> ___
> HCP-Users mailing list
> HCP-Users@humanconnectome.org
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users
>
>
>
> 
> ---
> Stephen M. Smith, Professor of Biomedical Engineering
> Head of Analysis,  WIN (FMRIB) Oxford
>
> FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
> +44 (0) 1865 610470
> st...@fmrib.ox.ac.ukhttp://www.fmrib.ox.ac.uk/~steve
> 
> ---
>
> Stop the cultural destruction of Tibet 
>
>
>
>
>
>
>
>
>
>
>
> 
>  Virus-free.
> www.avast.com
> 
> <#m_-1125881958447918270_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>
> ___
> HCP-Users mailing list
> HCP-Users@humanconnectome.org
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users
>

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] Convert subject-specific ICA node maps to volumetric space?

[Nicola Toschi]

Thank you Steve,

I was hoping to get them out of the CIFTI versions (also so I don't have 
to reprocess the 4D FIX data for all subjects). Maybe with 
cifti-convert? The brain would not be fully covered here i realise but 
much better than nothing.


Alternatively, do you think there is a chance of obtaining them somehow 
(since they have been calculated by you guys anyway)? Can i put in a 
request somewhere maybe?


Thank you in advance!

Nicola


On 11/13/2018 5:13 PM, Steve Smith wrote:
Hi - we calculated them in order to estimate the volumetric group 
maps.  However AFAIK we didn't bundle them in the PTN release.
You can easily get them though by regressing node timeseries into 4D 
FIX-cleaned data.


Cheers.


On 13 Nov 2018, at 16:09, Nicola Toschi > wrote:


Dear HCP list and experts,

I was wondering if the subject-specific node maps from the latest PTN
release (which are available in CIFTI format) are also available in
volumetric space (which would aid the specific analysis we are running).

Or if not, maybe a pointer on how to convert them?

Thank you very much in advance!

nicola

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users



---
Stephen M. Smith, Professor of Biomedical Engineering
Head of Analysis,  WIN (FMRIB) Oxford

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 610470
st...@fmrib.ox.ac.uk  
http://www.fmrib.ox.ac.uk/~steve 

---

Stop the cultural destruction of Tibet 











___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] Convert subject-specific ICA node maps to volumetric space?

[Steve Smith]

Hi - we calculated them in order to estimate the volumetric group maps.  
However AFAIK we didn't bundle them in the PTN release.  
You can easily get them though by regressing node timeseries into 4D 
FIX-cleaned data.

Cheers.


> On 13 Nov 2018, at 16:09, Nicola Toschi  wrote:
> 
> Dear HCP list and experts,
> 
> I was wondering if the subject-specific node maps from the latest PTN 
> release (which are available in CIFTI format) are also available in 
> volumetric space (which would aid the specific analysis we are running).
> 
> Or if not, maybe a pointer on how to convert them?
> 
> Thank you very much in advance!
> 
> nicola
> 
> ---
> This email has been checked for viruses by Avast antivirus software.
> https://www.avast.com/antivirus
> 
> ___
> HCP-Users mailing list
> HCP-Users@humanconnectome.org
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users


---
Stephen M. Smith, Professor of Biomedical Engineering
Head of Analysis,  WIN (FMRIB) Oxford

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 610470
st...@fmrib.ox.ac.ukhttp://www.fmrib.ox.ac.uk/~steve 

---

Stop the cultural destruction of Tibet 









___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

[HCP-Users] Convert subject-specific ICA node maps to volumetric space?

[Nicola Toschi]

Dear HCP list and experts,

I was wondering if the subject-specific node maps from the latest PTN 
release (which are available in CIFTI format) are also available in 
volumetric space (which would aid the specific analysis we are running).

Or if not, maybe a pointer on how to convert them?

Thank you very much in advance!

nicola

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users