Hi Aditya,
> 1) is it okay to coregister the epi volumes to an
> average T1w image created by averaging all the autistic
> subjects and all controls separately, and then each of them
> to the mni152 template? or is it better that i coregister
> each epi individually to the same mni152 template and then
> do the fibre tracking? or should i use the b=0 images
> acquired instaed?
I would strongly argue against using the b0 if you're using fnirt. The problem with using the T1w is that the distortions in the dw imply that you are not exactly in the same space for the same subject.
I would recommend using fnirt from the native FA to either the FMRIB58 template or an FA study-specific template.
> 2) is it okay to perform the fibre tracking
> in the diffusion space or should i do it in the template/b=0
> space?
It depends what you want to do. If you want to draw some ROI in the standard space and not in the each subject diffusion space, then it is better to perform the fibre tracking in the standard space (but you will need to inverse your warpfields first, see with Saad for the use of probtrackx and fnirt in the standard space).
> 2) after the whole brain diffusion volumes are
> coregistered and normalised, flirted and fnirted, how can i
> get the track directions of the two groups displayed in
> FSLview as i. two separate images, ii. as a difference map
> autism>control and control>autism?
Not sure why you would want to compare the voxel-wise result of the tractography... See archives for discussion on how to compare/use the tractography results between two populations.
To do what you want (presuming you want to compare the mean???), you would need to concatenate all your fdt_paths for all controls/patients (fslmerge) and do a Tmean on it (fslmaths). Then if you want to see the difference, fslmaths -sub.
Cheers,
Gwenaelle
|