External Email - Use Caution

Hello all,

I am currently using the longitudinal implementation of Tracula with a dataset of five subjects each with two time points. All follow-up time points were acquired roughly three weeks after all baseline time points on a different scanner (when our lab upgraded from a Siemens TimTrio to a Prisma) but with an equivalent sequence. In the past, ROI analyses have resulted in similar ROI means between baseline and follow up scans - roughly a < 5% difference between baselines and follow ups.

I was hoping for a similar range of differences between baseline and follow up whole-tract means with Tracula. I've looked at the genu, splenium, and left/right SLF 1; the % differences between baselines and follow ups are all over the place, ranging from ~25% to <1%. After looking at the data, my team believes this wide range of differences to possibly and/or partially be due to some of the tracts sampling CSF. The genu and the splenium in particular yielded very large % differences for us, so our thought is that masking out CSF should cause our % differences to go down.

My question is this: Is there an elegant way to implement CSF masking with Tracula? I don't believe Tracula has the ability to do the CSF masking itself, but is there a best practice of when/how to mask out CSF that works best with Tracula?

Thank you!

Kayti Thorn (she/her) Database Specialist Medical University of South Carolina Department of Neurology Center for Biomedical Imaging 135 Cannon St. CS101 Charleston, SC 29425