Hi Patrick
> > Sounds like you are almost there with your monkey data.
> > If you can do your own motion correction and get a reasonable
> > bet result with 0.18 for -f then there isn't much more to fix.
> >
> > Firstly to registration. You will need to define a "standard space"
> > to do your higher level analysis in. Obviously our human brain
> > template is not useful for this. Instead I suggest you pick a single
> > brain image from your monkey data (assumedly an MPRAGE
> > that looks nice) and, for the sake of argument, call this "standard
> > space" for your experiment. Having picked this image you just
> > need to use it in the Standard Space tab of the Feat registration.
> > You also want to put in the individual monkey MPRAGE images
> > in the "Main structural image" tab. If these images look OK then
> > flirt should be able to align them OK.
>
> Is FLIRT still necessary if the images are already realigned and thus
> motion corrected ? (I Am using the Feat GUI). apropos, templates is no
> problem, since we do have EPI and MPRAGE templates of our monkeys, and
> I've created with bet "pure-brain" images out of them.
If you are going to use FEAT at higher level then yes you'll need to use
the registration within FEAT at first level so that FEAT knows where to
find the transforms to "standard space". However, as Mark said (below) you
should then overwrite the resulting .mat files in ??.feat/reg with
identity if you want to ignore the new registrations.
> > One difficulty is that it will still try to register the functional images
> > to the structural ones. You don't want this, but unfortunately, it
> > can't be turned off. However, once registration is run you can
> > overwrite the results by replacing the example_func2highres.mat
> > file with the following contents:
> > 1 0 0 0
> > 0 1 0 0
> > 0 0 1 0
> > 0 0 0 1
>
> So this is not clear to me when you do the overwrite (I'm using the gui):
> is it just before the second level analysis which takes all runs into
> consideration ? or even before ? Because if I'm right the registration
> part comes at the end of the first level analysis stage, is it not ?
yes, so fix the files between first- and second-level FEAT runs.
> > This effectively resets the transformation of the example_func
> > to MPRAGE, as you say they are already aligned. Once you've
> > done this you'll need to regenerate example_func2standard.mat
> > which, in this case is easy, just copy the contents of highres2standard.mat
> > into example_func2standard.mat - that is:
> > cp highres2standard.mat example_func2standard.mat
> >
> > This will then set things up correctly for higher-level feat analyses.
> > It won't, however, fix the images on the report.html page - to do this
> > you need to regenerate the gif images made by slicer - see the
> > report.com file and extract the appropriate commands from there
> > if you need this.
>
> I guess fslview will return the right images ?
no - this is just a viewer - you need the slicer commands as found in
report.com if you want the report web pages to contain the right
registration report images (though this isn't necessary for the actual
running of higher-level FEAT).
> > Now, the BET issue. You shouldn't need to use pre-threshold
> > masking or change the brain/background percentage (although
> > this probably won't be a big deal either way). What you do need
> > to do is to apply you bet result to your functional series. This can
> > be done with the following commands:
> >
> > ${FSLDIR}/bin/bet funcimage funcimage_brain -m -n -f 0.18
> > ${FSLDIR}/bin/avwmaths funcimage_brain_mask -dil funcimage_mask
> > ${FSLDIR}/bin/avwmaths functionalseries -mas funcimage_mask maskedseries
>
> small error I guess (otherwise I'm completely lost), but line 2, shouldn't
> that read: ${FSLDIR}/bin/avwmaths funcimage_brain -dil funcimage_mask ???
> Moreover, what exactly does this function do, I did not see a real
> difference between funcimage_brain and funcimage_mask (btw, our images do
> have a 1mm^3 voxelsize). Is it some kind of smoothing ?
nope - you need to run exactly what Mark wrote...the _mask image is
generated on the first line because of the -m option and is what you want
to feed into the dilation on the second line before using the output of
this in the third line.
Cheers.
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
|