Hi,
If you also have a low b-value (e.g. 1000), you will hardly resolve any crossings. So I would suggest keeping the number of fibres to 1. Changing the burn-in will not help in this aspect.
Cheers,
Stam
On 19 Jan 2011, at 11:20, András Jakab MD wrote:
> Dear FSL Experts,
>
> We wish to perform a study with connectivity-based segmentation of healthy brains. We have a dataset of 40 patients but the problem is that the DTI acquisitions merely have 12 directions. Can you recommend a protocol that might make the segmentation/clustering using ccops more feasible? i.e., if we run bedpostx with only 1 fibers, does that make the probtrackx less sensitive for the negative effects of only 12 directions? Should we increase the "Burnin" value?
>
> Thank you very much.
> András Jakab MD
> University of Debrecen / ETH Zürich Computer Vision Lab
|