Dear Serena, To do this you need to use contrast masking (a button in the Post-Stats tab) with the "Z>0" option, choosing to mask the F-test result by a t-contrast (which you need to setup before) that is on the EV associated with the canonical HRF. This will restrict the F-test results to only those voxels where the main response to the canonical HRF is positive. All the best, Mark On 4 Jul 2012, at 16:53, sere f wrote: > Dear FSL users, > I have generated 3 basis set functions using FLOB in order to obtain canonical HRF, its temporal derivative and its dispersion. > I turn on F-test but in the fzstat map it is not possible to distinguish activation and deactivation. > how can I do? > > Thanks for any advice. > > Serena