 |
 |
Hi Stam i tried qboot on real dataset acquired with 160 graidient directions and a b-value of 2000, the results are really satisfactory, unlike the case of synthetic dataset of 5x6 ODFs with a separating angles spanning from 3° to 90°, that i build using multitensor model (two tensors) with a b-value of 1500 and 60 directions, where the results using qboot with --pf=2. The fiber directions, even close to 90° separating angles, are not recovered. but when i used qboot without --pf=2 on this synthetic dataset the results are inline with the literature (the angular resolution when using csa-ODF is 57° and with tuch-ODF is 68°). I don't know if the problem is in the number of gradients directions (60). i can send you my synthetic dataset, the mask, bvals and bvecs files if you want. Makhlouf