longitudinal seed based tractography and sampling from images

Hello everyone,

I have 12 subjects with 5 follow ups. For each subject, we have 4 functionally defined ROIs. We also have quantitative MRI data, which we want to sample along the tracts between ROIs to see the effect of learning. What is the best way of doing this sort of analysis? I see two ways of doing it:

Option 1:

  1. Run seed-based tractography on Day 1 only (then we would end up not using the other available diffusion data)
  2. estimate the transformation matrix from qMRI data of the other days to the Day1 diffusion data (I guess we do not have to reslice them, or do we?)
  3. use tcksample to sample along the tracts of Day 1

Option 2:

  1. Run seed-based tractography on each day separately (but then we won’t be able to sample from the same tract across days…)
  2. estimate the transformation matrix from qMRI data to diffusion data of the same day (I guess we do not have to reslice them, or do we?)
  3. use tcksample to sample along the tracts for each day separately, but then just take the mean of the tracts

Option 1 makes more sense to me because we make sure that we sample from the same region (tract) across days, but we won’t be using the diffusion data from all the days. I have been looking better ways to approach longitudinal diffusion data and sampling, but couldn’t find a reasonable way. I appreciate any type of input, or ideas!

All the best,
Ayberk