Hello experts,
I'm trying to wrap my head around proj_thresh.
From what I understand, each voxel in each mask generated is the proportion of streamlines from that voxel that reach ANY target mask.
I ran seeds-to-classification targets, seeding the corpus callosum and tracking to left m1 and right m1, seeding 2500 streamlines from every voxel of the seed mask. I want to see the probability of connectivity/an index of "strength" of connectivity from the CC to motor regions in general, not focusing on hemispheric differences in my later analyses.
Secondly, I am confused with the thresholding and interpreting the results. I know there doesn't seem to be a principled way of choosing a threshold value, but in playing around with it I find that using my results vary significantly depending on the value I choose. For example, for looking at the left m1-cc mask, the means I get for:
thresholding at 1 - 0.735
thresholding at 125 - 0.439
thresholding at 625 - 0.585
That doesn't make sense to me. I would expect to see a more linear change in the averages as the threshold increases. But I guess that leads me to my questions:
1) What do I interpret those averages as? They can be thought of as probabilities, correct?
2) What is the utility of providing two masks? i.e. should these be averaged together to take the mean probability overall?
Any guidance is greatly appreciated!
Thanks,
Josh
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|