Hi Gabriel,

I would suggest you take a look at the Gibbs tracking approach, since especially for quantification it seems a promising approach to me. It is implemented either in the DTI&fibertools for SPM or also in the MITK diffusion toolkit (where also other algorithms are implemented).

cheers, Andy

Hi,

In my opinion, the use of which DTI technique is only one thing, and yes there are several modern possibilities. But the real high-level comes when you are able to imply other techniques in order to cross validate the DTI results whatsoever, e.g. post-mortem, primate brain, or ther imaging modalities.

Best, Markus 

 High level tractography is probably the wrong way to indicate what you want to do.

Many experts of the field are convinced that CSD is the most cost effective approach to obtain fairly better results than DTI.

At the moment one of the most popular software for this approach is MR-trix that can give multiple kind of outputs (i prefer .trk to use it on Trackvis) and that allows to compare many different algorithms including deterministic and probabilistic CSD. Its main drawback is that it is not very user friendly (it depends on your background).

Other popular software are Explore-DTI (from Alexander Leemans, Utrecht) and Star-Track (from Flavio Dell'Acqua, Kings College London). I mostly use the latter.

For quantitative matters Explore-DTI may be the best choice because it provides multiple metrics and tools.

In my personal opinion streamlines number and length are not reliable enough in most of the conditions. However the best metrics to use depend on which phenomenon or pathology you are studying.

Thank you  Andy, Markus  and Guiseppe,

I've started to work with ExploreDTI as appart from being so user friendly it affectively as lots of tool for extracting metrics for streamline analysis. Also as it runns under matlab, has the same format that can help to program myself more metrics and allows me to a programmatic-pipieline.

The draw back I've found is that only a deterministic apprach has been set or at least in the version I do have there's only the deterministic CSD algorithm is available. and something it could be important to have the probabilistic one.

I also will check at the MITK diffusion toolkit, so I'll be able to compare different algorithms (any suggestions about how to proceed in here?).

Bests,

Gabriel

In my opinion the best software for tractography now is a combination ExploreDTI/MRtrix, using CSD and probabilistic algorithm. As you mentioned yourself, ExploreDTI does not offer great probabilistic algorithms, yet is the software I use to manage the preprocessing, MRtrix cannot do preprocessing at all. Using them in combination with HARDI, I have been able to obtain difficult tracts such as the optic radiations, also getting similar results to dissection studies and other tractography papers. So ExploreDTI could be the best preprocessing software, MRtrix the best CSD tractography software.

StarTrack did not yield good  tractography for me, streamlines were more noisy with more erroneous directions. This is despite using HARDI. But I think Flavio was working on integrating a probabilistic algorithm as this option was greyed out when I used it.

Good luck.

I agree with previous comments for MRtrix that is a wonderful software for post-processing, especially when using Track-weighted imaging. TWI maps bring new insights in brain anatomical pathways as well as for studying peripheral nerves. Yet, quantitative microstructural analysis remains difficult except for basic DTI derived parameters such as FA or MD.