Dear Supreet,
Our adult pipeline uses a two-stage registration of fmri to structural and then structural to template. The last step (structural to template) is done with non-linear registration but requires that the template and the structural image are of the same modality (e.g. both T1-weighted if using the MNI152). So for your case, if you have a good T2-weighted structural image then I would use this approach. If you don't have a good T2-weighted structural then you will probably have to just try and make do with a linear registration (structural to template if you have a structural, or directly from the fmri to the template).
We also advocate the use of fieldmaps in the fmri to structural registration, to correct for distortions in the fmri (EPI) images. However, to use this you would need the fieldmaps (obviously) as well as a structural image that had sufficient tissue contrast to be able to be anatomically segmented, and also a reasonable amount of tissue contrast in the fmri. If you can't meet these three requirements then you will need to proceed without fieldmap correction, or use an alternative pipeline for it, rather than the current FEAT pipeline.
I hope this helps.
All the best,
Mark
On 18 Jul 2013, at 22:02, Supreet kaur <[log in to unmask]> wrote:
> hi FSL experts,
>
> I'm trying to normalize the pre-term fMRI data to the three-month old Delenhe-lambert T2-weighted template. I need to analyze the resting state fMRI for our pre-term infant cohort.
> Which option should be the optimum for the normalization- FEAT or FLIRT?
> Then which reference image should I choose for the MELODIC ICA analysis?
> I'd appreciate if you could provide your feedback on this.
>
> Thanks,
> Sincerely,
>
> Supreet
>
|