That command is meant for that purpose, yes, but you will first need to use
one of the -cifti-create-* commands to make a template cifti in the
downsampled space (you can use things like -surface-vertex-areas,
-volume-create, and -volume-affine-resample to get the needed inputs for a
-cifti-create-* command).  I would also suggest considering parcellation
instead of downsampling, so that the entirety of an area contributes to a
single output value.  Choosing a good parcellation for this is important,
though.

Tim


On Thu, Oct 29, 2015 at 2:36 PM, Jennifer Elam <[email protected]> wrote:

> Hi,
>
> Tim C. can confirm, but if you are trying to resample a CIFTI fMRI data
> file, such as a dense connectivity *.dconn, I would try using Connectome
> Workbench <http://humanconnectome.org/software/connectome-workbench.html>:
>  wb_command -cifti-resample
>
> Here’s the help information
> <http://humanconnectome.org/software/workbench-command.php> for the
> -cifti-resample command:
>
>
>
> RESAMPLE A CIFTI FILE TO A NEW CIFTI SPACE
>
>    wb_command -cifti-resample
>
>       <cifti-in> - the cifti file to resample
>
>       <direction> - the direction of the input that should be resampled
>
>       <cifti-template> - a cifti file containing the cifti space to
> resample to
>
>       <template-direction> - the direction of the template to use as the
>
>          resampling space
>
>       <surface-method> - specify a surface resampling method
>
>       <volume-method> - specify a volume interpolation method
>
>       <cifti-out> - output - the output cifti file
>
>
>
>       [-surface-largest] - use largest weight instead of weighted average
> when
>
>          doing surface resampling
>
>
>
>       [-volume-predilate] - dilate the volume components before resampling
>
>          <dilate-mm> - distance, in mm, to dilate
>
>
>
>       [-surface-postdilate] - dilate the surface components after
> resampling
>
>          <dilate-mm> - distance, in mm, to dilate
>
>
>
>       [-affine] - use an affine transformation on the volume components
>
>          <affine-file> - the affine file to use
>
>
>
>          [-flirt] - MUST be used if affine is a flirt affine
>
>             <source-volume> - the source volume used when generating the
> affine
>
>             <target-volume> - the target volume used when generating the
> affine
>
>
>
>       [-warpfield] - use a warpfield on the volume components
>
>          <warpfield> - the warpfield to use
>
>
>
>          [-fnirt] - MUST be used if using a fnirt warpfield
>
>             <source-volume> - the source volume used when generating the
>
>                warpfield
>
>
>
>       [-left-spheres] - specify spheres for left surface resampling
>
>          <current-sphere> - a sphere with the same mesh as the current left
>
>             surface
>
>          <new-sphere> - a sphere with the new left mesh that is in register
>
>             with the current sphere
>
>
>
>          [-left-area-surfs] - specify left surfaces to do vertex area
>
>             correction based on
>
>             <current-area> - a relevant left anatomical surface with
> current
>
>                mesh
>
>             <new-area> - a relevant left anatomical surface with new mesh
>
>
>
>          [-left-area-metrics] - specify left vertex area metrics to do area
>
>             correction based on
>
>             <current-area> - a metric file with vertex areas for the
> current
>
>                mesh
>
>             <new-area> - a metric file with vertex areas for the new mesh
>
>
>
>       [-right-spheres] - specify spheres for right surface resampling
>
>          <current-sphere> - a sphere with the same mesh as the current
> right
>
>             surface
>
>          <new-sphere> - a sphere with the new right mesh that is in
> register
>
>             with the current sphere
>
>
>
>          [-right-area-surfs] - specify right surfaces to do vertex area
>
>             correction based on
>
>             <current-area> - a relevant right anatomical surface with
> current
>
>                mesh
>
>             <new-area> - a relevant right anatomical surface with new mesh
>
>
>
>          [-right-area-metrics] - specify right vertex area metrics to do
> area
>
>             correction based on
>
>             <current-area> - a metric file with vertex areas for the
> current
>
>                mesh
>
>             <new-area> - a metric file with vertex areas for the new mesh
>
>
>
>       [-cerebellum-spheres] - specify spheres for cerebellum surface
> resampling
>
>          <current-sphere> - a sphere with the same mesh as the current
>
>             cerebellum surface
>
>          <new-sphere> - a sphere with the new cerebellum mesh that is in
>
>             register with the current sphere
>
>
>
>          [-cerebellum-area-surfs] - specify cerebellum surfaces to do
> vertex
>
>             area correction based on
>
>             <current-area> - a relevant cerebellum anatomical surface with
>
>                current mesh
>
>             <new-area> - a relevant cerebellum anatomical surface with new
> mesh
>
>
>
>          [-cerebellum-area-metrics] - specify cerebellum vertex area
> metrics to
>
>             do area correction based on
>
>             <current-area> - a metric file with vertex areas for the
> current
>
>                mesh
>
>             <new-area> - a metric file with vertex areas for the new mesh
>
>
>
>       Resample cifti data to a different brainordinate space.  Use COLUMN
> for
>
>       the direction to resample dscalar, dlabel, or dtseries.  Resampling
> both
>
>       dimensions of a dconn requires running this command twice, once with
>
>       COLUMN and once with ROW.  If you are resampling a dconn and your
> machine
>
>       has a large amount of memory, you might consider using
>
>       -cifti-resample-dconn-memory to avoid writing and rereading an
>
>       intermediate file.  If spheres are not specified for a surface
> structure
>
>       which exists in the cifti files, its data is copied without
> resampling or
>
>       dilation.  Dilation is done with the 'nearest' method, and is done on
>
>       <new-sphere> for surface data.  Volume components are padded before
>
>       dilation so that dilation doesn't run into the edge of the component
>
>       bounding box.
>
>
>
>       The recommended resampling methods are ADAP_BARY_AREA and CUBIC
> (cubic
>
>       spline), except for label data which should use ADAP_BARY_AREA and
>
>       ENCLOSING_VOXEL.
>
>
>
>       The <volume-method> argument must be one of the following:
>
>
>
>       CUBIC
>
>       ENCLOSING_VOXEL
>
>       TRILINEAR
>
>
>
>       The <surface-method> argument must be one of the following:
>
>
>
>       ADAP_BARY_AREA
>
>       BARYCENTRIC
>
>
>
> Best,
>
> Jenn
>
>
>
> Jennifer Elam, Ph.D.
> Outreach Coordinator, Human Connectome Project
> Washington University School of Medicine
> Department of Anatomy and Neurobiology, Box 8108
> 660 South Euclid Avenue
> St. Louis, MO 63110
> 314-362-9387
> [email protected]
> www.humanconnectome.org
>
>
>
> *From:* [email protected] [mailto:
> [email protected]] *On Behalf Of *?
> *Sent:* Thursday, October 29, 2015 3:15 AM
> *To:* hcp-users
> *Subject:* [HCP-Users] downsample
>
>
>
> hi, I have a question about downsample the fmri grayordinate data.
>
>  30k vertices means too many for me, I wander if I can find the
> downsampled data from the database, if not, what should I do to downsample
> using my own computer.
>
> Thank you.
>
> _______________________________________________
> HCP-Users mailing list
> [email protected]
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users
>
> _______________________________________________
> HCP-Users mailing list
> [email protected]
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users
>

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

Reply via email to