Dear John,
Could you please confirm whether this is correct? The only thing I'm
unsure about is the "Image to base inverse on", is that the image that
had been used for generating the iy* file or the image defining the
functional space? Both give reasonable, but not identical results. I'd
guess it is the former, but I just want to make sure.
Thanks!
Martin
Choose:
Util->Deformations
Composition -> Inverse -> New: Deformation Field
Output -> New: Pushforward
then it looks like this
Composition
. Inverse
. . Composition
. . . Deformation Field: [fill in your iy-file]
. . Image to base inverse on: [Image defining the space of your
iy-file (normally your structural)]
Output
. Pushforward
. . Apply to: [Choose image(s) that need inverse transformation]
. . Weight image
. . Output destination
. . . Output directory: [pick your output directory]
. . Field of view
. . . Image Defined: [image defining your functional space]
. . Preserve: Preserve Concentrations
. . Gaussian FWHM: [0 0 0]
2015-11-06 20:48 GMT+01:00 John Ashburner <[log in to unmask]>:
> If you have an image that defines your predefined image space, then you can
> use the Deformations Utility to achieve these goals. I'd suggest taking a
> look in the "Using Dartel" chapter of the manual (especially section 45.4)
> for hints about how to do this (base it on the example - but perhaps leave
> out the transform that would warp A to the template).
>
> Best regards,
> -John
>
>
> On 6 November 2015 at 18:30, Martin Hebart <[log in to unmask]> wrote:
>>
>> Dear all,
>>
>> In the past, I've done inverse normalization using the now old segment
>> and Dartel, but I haven't found a way to inverse normalize to a
>> predefined image space using SPM12.
>>
>> Using the iy_xxx.nii image and normalise+write, this only allows a
>> specific bounding box and voxel size. However, if I use the voxel size
>> and bounding box from my original image space, the final image space
>> from the inverse normalization is different.
>>
>> I would be happy if anyone could point out to me how to inverse
>> normalize an image to a predefined image space (without first writing
>> to any space and then transforming to the desired space using e.g.
>> resize_img - that interpolation step seems unnecessary).
>>
>> Cheers,
>> Martin
>
>
|