Dear all,
I've run featquery from GUI (FSL version 5.1.0) on first-level analyzed subjects using a mask image defined in the standard space. The registration of the standard space mask to example_func space was bad in some cases and therefore the reported stats values were also bad. The registration options for first-level Feat analyses were BBR (example_func to MPRAGE) and non-linear (MPRAGE to standard). For a subject, where featquery was wrong, we've rerun the registration using the Registration-only in Feat with the same options (BBR & non_linear). After that the featquery was able to correctly register the mask image into example_func space. By inspecting the files in the ..../reg/ directory and comparing them between the cases when featquery worked fine and bad, we realized a difference in the inverse warpfields. The first volumes of the inverse warpfields were the same, but half of the second and full of the third volume of the wrong inverse warpfield file was empty. There was not any error message by featquary.
Does anybody have similar experience with Fearquery? Featquery uses the invfeatreg. Maybe it is an issue related to that rather than featquery..
Thanks,
Gabor
|