Hello MRtrix experts,
I am currently investigating structural changes in aging (comparing a young cohort with an older cohort) with the one eventual goal of exploring relationships between structural changes and functional/behavioral changes. My question is, is there a single meaningful value to summarize structural connectivity for each subject that can be derived from MRtrix processes? For example, an analog to global FA. Previous literature in aging tends to use global FA to correlate with other values, but I would like to avoid this metric if possible.
I have generated connectomes for all my subjects, using 10 million streamlines and SIFT2, multiplying my output matrices by individual mu values. Additionally, I am currently working on performing FBA with single tissue data. Any single value that could be calculated for each subject from either of these two processes would be great.
I can think of a couple of imperfect analogues to a “global mean FA”:
Calculation of global network metrics such as those provided in the Brain Connectivity Toolbox. A number of these are metrics are calculated at the per-connectome-node level, but are then frequently averaged across nodes to produce a single scalar value per subject.
Staying in the domain of voxel-based measures, you could try calculating metrics such as “complexity” or “sf” within the
fixel2voxel command, and then take the average across the white matter.
If you are able to perform a three-tissue decomposition, you could even just take the mean WM signal intensity (just the l=0 term of the spherical harmonic representation) throughout a mask deemed to correspond to white matter.
Each of these has their own mechanistic pros and cons, and fundamentally measures something different; but maybe it’s enough to get you thinking about what type of experiment you want to perform.