Hi,
I'm using probtrackx to look at group differences between patients and controls in terms of subcortical-cortical connectivity using classification targets.
I've used the defaults in probtrackx (5000 streamlines etc), but I'm not sure what threshold to use for proj_thresh so that I can make appropriate group comparisons. The literature and the list cite a range of thresholds.
The group sizes are modest (16 vs 16), so if I want to avoid too much noise would a threshold of 25% seem appropriate? (The same seems to be used for determining CDRs).
Cheers,
Ed
|