 |
 |
Dear FSL users, we are currently using the FSL tools for VBM analysis. Unfortunately, we are somehow stuck at the first step after brain extraction. FAST classifies in almost all our brains parts of the meninges, especially in the tentorium, but also elsewhere as GM with a probility of 1. The scans were taken with a 3 T Siemens Trio Scanner, MPRAGE Sequenze, voxel size 1x1x1mm. We used the standard options as in the FLSVBM_2 script (fast -R 0.3 -H 0.1) and also tried increasing main-loop an bias-field iterations. Our last attempt was to repeat segmentation with own priors (registering the templates to each brain), which didn't really improve our results. I fear that this will substantially influence our results. Is there anything, that can be done about this? Thanks for your help. Regards Richard