 |
 |
I agree with Tim. Note that IF instead of the above discussion you are happy with NOTB being defined as (B<0) then you can simply define a new contrast which is the negative of B and then use the (>0) option in the FEAT masking GUI. Cheers. On Tue, 29 Jun 2004, Andy Smith wrote: > Thanks, Tim. I'll find avwmaths and try it out. > Yes, I do have a sensible reason for doing it. We're simply trying > to define an ROI within the visual cortex for a subsequent > quantitative analysis and we won't be drawing any statistical > inferences about NOT B! If that were permissible, we'd all be in > Nature every week ;-) > Andy > > > >Hi - you want to be a little bit careful with this. > > > >A feature of null hypothesis testing is that "NOT B" isn't very > >meaningful. That is, we are not able to say that a voxel does not activate > >to stimulus B, we can only say that if this voxel were not activating, > >then we wouldn't be all that surprised to see this data - it is really a > >very weak statement, and you certainly won't be able to perform null > >hypothesis style inference on a mask which is formed as the intersection > >of "A" with "NOT B" . > > > >Hence, there isn't an option for this in Feat > > > > > >Having said that - if you have some sensible reason for performing this > >intersection then you can do it with avwmaths. > > > >something like > > > >avwmaths B -bin -mul -1 -add 1 -mas A A_NOT_B > > > >should give you a mask which defines the intersection of A with NOT B > >which you can use for further analysis. > > > >Cheers > > > >T > > > > > > > >On Tue, 29 Jun 2004, Andy Smith wrote: > > > >> Is any kind of contrast masking possible in FEAT other than a > >> standard exclusive mask? > >> > >> I have defined two contrasts from the same experiment and I want to > >> identify all voxels that are significantly activated in contrast A AND NOT > >> contrast B. It has to be one-way, i.e. I don't want to include voxels that > >> are active in B and not A. An exclusive mask gives voxels active in both > >> but not in either alone, so that's no good. > >> > >> Any ideas on how to do this would be much appreciated. (It shouldn't > >> be too hard to write a script to do it, but there's no point in re-inventing > >> the wheel.) > >> > >> Thanks, > >> Andy Smith > >> > >> PS. I'm turning to you guys because SPM won't do it, so here's a > >> chance for you to gain a convert! > >> > > > >-- > >------------------------------------------------------------------------------- > >Tim Behrens > >Centre for Functional MRI of the Brain > >The John Radcliffe Hospital > >Headley Way Oxford OX3 9DU > >Oxford University > >Work 01865 222782 > >Mobile 07980 884537 > >------------------------------------------------------------------------------- > Stephen M. Smith DPhil Associate Director, FMRIB and Analysis Research Coordinator Oxford University Centre for Functional MRI of the Brain John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK +44 (0) 1865 222726 (fax 222717) [log in to unmask] http://www.fmrib.ox.ac.uk/~steve