Is it possible to use DKE for diffusion kurtosis for scans with 4 b-values (0, 1000, 3000, 5000), and each non-zero b-value having 12 directions? This is from a Siemens scanner. Can DICOMs be in mosaic format?
Thanks!

Thank you for using DKE!

DKE requires 15 directions for each b-value unless you make other assumptions for an underdetermined system. Are your scans taken in vivo? If so, do not use the b=5000 images. The assumptions made by DKE are invalid when using b-values over 2500. You can probably get away with using the b=3000 images. Refer to the following paper for further clarification: https://www.ncbi.nlm.nih.gov/pubmed/20632416.

DICOMS can be in mosaic format.

I hope this is helpful! Let me know if you have any questions.

These are ex vivo in cranio (not fixed) scans. Can you explain a bit more about "...unless you make other assumptions for an underdetermined system"?

Thanks for your clarification.

I have attached an abstract that you may find useful.

Thanks,
Corinne