Hello,
I have performed probabilistic tracking using probtrackx2 using left nucleus accumbens (standard space) as seed and left forebrain (standard space) as way point, supplying non-linear transforms to standard space as input.
After visualizing the tacking results (fdt_paths.nii.gz) for 70 subjects in standard space, I see that the fibers connect from the seed region to the forebrain in all the subjects, but there is a huge variation.
i tried to calculate the overlapping regions after thresholding the fdt_paths.nii.gz. I see that the tracts do not overlap except for the region around the seed. I expected a bigger overlap of fibers in standard space across subjects.
I have verified the transformations suppled as input to and from the standard,structural and diffusion space. There seems to be no mistake there.
I am wondering if there is anything that can explain such kind of inter subjection variation of tracked fibers? Is it beter to use FA maps (instead of b0 maps) to get these transformations to standard space for more accuracy? If such inter subject variation exists, it would not be possible to perform voxel-wise statistics on these tracked fibers.
Another question is about --pd option. Would it be better to compare the fibers across subjects with or without --pd option in this case?
Thanks
Jay
|