gonzoBlackMamba
commented
4 months ago Hi @NN3265
The WMTI model requires a DKI dataset, not just DTI. That is you need b = 0, 1000 and 2000 shells. The compartmental diffusivities (extra- and intra-axonal) that are computed require both the kurtosis and diffusivity which DKI provides but cannot be obtained with DTI. The meaning of the parameters is all based on the assumptions involved in the WMTI model (which is linked here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3136876/pdf/nihms-305423.pdf). Basically, the regions where it is most valid are those with highly aligned WM, such as areas that have FA >= 0.4. They are modeling parameters that relate to the exta-axonal space and the intra-axonal space but need to be carefully interpreted based on the regions of interest and the assumptions of the model. Here is a link to a clinical application of the WMTI model: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3962262/#B18.
Nowadays, we recommend getting a 30 direction DKI along with at least 64 directions high value data (b >= 4000 s/mm^2) for an fiber ball imaging (FBI) analysis. This allows us to calculate an fODF related to the fiber geometry of the axons themselves. Together, the DKI and FBI data are combined into a more general model (FBI WM) that does not have the same constraints as WMTI and is valid throughout the WM in the brain. See this publication: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6064190/
Hope this helps, and sorry for the delay in responding.
-Hunter
Dear all,
The WMTI processing spits out diffusivity parameters which are created using DTI. Therefore, I was wondering what the meaning of the DTI parameters is. Are they generated solely using the b0 and b1000, or with all b-values, and hence they are actually the DKI diffusivity parameters?
Many thanks for your help!