ok so i have played about abit and now got some pretty blobs, but which
files do i register to the standard brain now? so i can overlay onto
standard in e.g.mricro i thought it would be
flirt -in thresh_zstat1 -ref
/home/andy/EBA/sp_160603/run1/sp_run1.feat/reg/standard -applyxfm -init
/home/andy/EBA/sp_160603/run1/sp_run1.feat/reg/example_func2standard.xfm -ou
tput_file
but this doesnt seem to work, is it the example_func2stadard.xfm bit that
needs replacing with another file?
thanks
andy
----- Original Message -----
From: "Christian F. Beckmann" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Tuesday, June 24, 2003 5:49 PM
Subject: Re: [FSL] Higher Level analysis and Mean's
> Hi
>
> depends on the question you have: If it's the random effects question
> then random effects analysis makes sense (and you potentially have to
> live with an answer you might not like).
> RFX not generating significant results could be for a variety of
> reasons, you should try and interrogate your data further to see if you
> can find a reason for this (e.g. suboptimal first-level model, inflated
> first-level variance due to artefacts etc )
>
> If it's not RFX you're after then yes, use something else. If you're
> happy with just asking about the effect within your sampled population
> of subjects use simple fixed effects analysis.
>
> ta
> christian
>
>
>
>
>
> On Tuesday, Jun 24, 2003, at 16:58 Europe/London, Andy Myers wrote:
>
> > do you think a different type of analysis would be more suitable?
> >
> > andy
> >
> > ----- Original Message -----
> > From: "Christian F. Beckmann" <[log in to unmask]>
> > To: <[log in to unmask]>
> > Sent: Tuesday, June 24, 2003 3:53 PM
> > Subject: Re: [FSL] Higher Level analysis and Mean's
> >
> >
> >> Hi
> >>
> >> well, possible given that you're doing a random effects analysis. Even
> >> if most of these contain 'activation' at the single level that doesn't
> >> imply that the RFX group map will show significant voxels too.
> >> WRT unthresholded maps: look in the BLAH.gfeat/copeX.feat/stats/ for
> >> the unthresholded higher-level zstat maps for each of the copes
> >>
> >> ta
> >> christian
> >>
> >>
> >>
> >> On Tuesday, Jun 24, 2003, at 11:36 Europe/London, Andy Myers wrote:
> >>
> >>> Hello,
> >>>
> >>> I have got my 8 runs of data that the higher level analysis said
> >>> there
> >>> was
> >>> no mean activation on, and looked at the same voxel in standard space
> >>> across
> >>> the runs and there is defintely activation in the same areas a large
> >>> percentage of the time, so i cant work out why the higher level
> >>> anaysis has
> >>> absolutely no activation in it at all. am i feeding in the right
> >>> cope?
> >>> i
> >>> give feat the 8 cope3 images from the feat stats directort, should i
> >>> do
> >>> anything to these first such as standardise them or something, or is
> >>> there a
> >>> problem with my thresholding levels, would that cause a conflict?
> >>>
> >>> also if i wanted to look at the unthresholded zstat image then what
> >>> do
> >>> i
> >>> need to register it to in order to view it? i tried standard space
> >>> like with
> >>> the first level thresh_zstat images but that didnt work (using flirt
> >>> -in
> >>> etc....................................................)
> >>>
> >>> andy
> >>
>
|