Hi Ben,

Usually when you are calculating a registration between two images, you will get better results by choosing the higher resolution image as the reference.

Task-based fMRI analyses (e.g. using FEAT), and single-subject resting state analysis (e.g. MELODIC) are typically performed in native functional space, so the functional-to-structural registration will have no effect on the results of the analysis. The results of registration are only used to transform the resultant statistic or component images into structural or standard space, for subsequent multi-subject analyses.

Multi-subject ICA is performed on functional data that has been registered to a standard space, but in this case the registration won't defeat the purpose of motion correction - functional-to-structural registration is calculated from a single functional image (either a single reference time point, or the mean across time), and the results of this registration are then applied to every time point. So each time point will have the same transformation applied to it, which means that, for the data for a single subject, there will be no affect across time.

Cheers,

Paul

On 9 August 2018 at 19:50, Ben M <[log in to unmask]> wrote:

Hi,
I'm confused about this flirt command as described in the FNIRT user guide:
flirt -ref my_betted_structural -in my_functional -dof 6 -omat func2struct.mat

My understanding of coregistration is that we take a representative functional image (eg the mean image of motion correction step) and use it as the stationary (reference) image. We then take the anatomical image (the input image) and we move it to align it to the mean functional image.

Assuming that my_functional.nii.gz in the example above is motion corrected (eg the mean image of motion correction), why would I select the structural image as the reference image, and the my_functional as the input image? Won't this jiggle about the motion corrected functional image to match the anatomical image, which would defeat the purpose of motion correction?
Thanks!
Ben

########################################################################

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1