 |
 |
Dear FSL experts, I am working with some fMRI data from patients with dementia and I am having some trouble in registration with FEAT. The registration of functional images to the structural ones was performed with a linear transformation with 6 DOF. For the registration between the structural images and the standard space a non-linear transformation with 12 DOF was used and a warp resolution of 10mm. For the remaining parameters the standard options were used. I found 2 problems: # For the great majority of subjects, the highres 2standard and the example_func2highres registrations are acceptable, but at least one slice of example_func2standard has a portion of the brain missing in a very irregular way or the red lines do not fit well enough (https://www.dropbox.com/sh/68xlj10i19q57y8/lXhZ5A8kqO, see the problem1 folder, the last 4 brains of example_func2standard.png) # In other few cases, also the highres2standard registration has problems: the red lines don’t adjust to the atrophied brains (https://www.dropbox.com/sh/68xlj10i19q57y8/lXhZ5A8kqO, see the problem2 folder, the highres2standard.png) Could anyone give me an idea of what is causing this unsatisfactory registration and how can I improve it? Sorry to bother you but I am a newbie at fsl. Thank you in advance! == Márcia F. T. Santos MSc Student in Biomedical Engineering Portugal