[Camino-users] Differences in number of streamlines computed with FSL and Camino

[Cook, Philip]

Hi,

Yes, the tractography is likely to be different in a number of ways, and the way in which connectivity between ROIs is counted also differs. 

If you run probtrackx with no args you will see there are many options, only some of which have a reasonably similar equivalent in Camino (eg, step size). And there are almost certainly other differences that are not exposed to the user, for example how the bedpost data is sampled, or how to treat seed points in voxels containing multiple fibers.

Here's my best guess of some options to start with, if you are looking to make the results of simple tracking (ie, the heat maps one gets from probtrackx / procstreamlines -outputacm) more similar

probtrack               track
--nsamples=5000   -iterations 5000
--nsteps=2000        [hard coded to 1000mm, so same as probtrack if default step size of 0.5mm is used]
--cthr=0.2               -ipthresh 0.2 and -curveinterval 0.5 is my best guess
--fibthresh=0.01      -bedpostxminf 0.01
--steplength=0.5     -stepsize 0.5 
 
I'd use "-tracker euler -interpolator linear" unless using --modeuler in which case you'd want "-interpolator prob_nn".

I would expect the results to be even more different if you do any postprocessing of the streamlines, eg with waypoints or target masks. 

> On Jan 23, 2018, at 3:47 PM, Suraj Thyagaraj <sxt492 at case.edu> wrote:
> 
> Hi all,
> 
> I am using both FSL and Camino based tractography to compute the connectivity between different brain structures as a function of the total number of streamlines between any pair of ROIs. With Camino, I import the data from FSL's bedpostx to run both probabilistic as well as deterministic tractography. 
> 
> Here's my question: Why do the number of streamlines between any two ROIs (even the total number of streamlines from an ROI) computed by Camino based probabilistic tractography differ from those computed by FSL based probabilistic tractography, even when Camino is using the data from FSL? Is it because of different models being used to perform tractography? 
> 
> Cheers!
> _______________________________________________
> Camino-users mailing list
> Camino-users at www.nitrc.org
> https://www.nitrc.org/mailman/listinfo/camino-users