Dear Rene
> 1st problem
> The cluster is in gray matter. I think if it is not soluted I can make algorithm which will calculate the nearest white matter area where I will insert my ROIs of white matter (seed, target or part of white matter where the pathway goes through).
Depending on the location/size of the fMRI clusters you may be able to use them as they are. If you feel there is not enough anisotropy within them to drive tractography, you could dilate the ROIs to reach adjacent WM (e.g. using fslmaths).
> 2nd problem
> I think this one is worse. Can I simply use this areas of white matter as seed and target? There is a problem that the tractography accurancy falls near by gray matter. Or is it more precise calculate global connectivity and looking only for pathways which are going through both areas of white matter?
> If second one is better which seed should I use for global connectivity estimation? Only corpus calosum or some another part of white matter?
Seeding from the whole brain will definitely take much longer than simply trying to track between two defined regions. Depending on the tract you want to estimate and the quality of your data, the latter approach may be good enough. You can either use each of the ROIs as seed and the other as waypoint/termination mask or have a look at the network option in probtrackx.
Good luck
Stam
|