Hello Hauke,
Sorry, we didn't answer your previous question about group-level F-test
stuff either yet....there isn't currently a simple way within the GLM
framework to carry F-test results up to second level. There are various
things that could be done:
a) You could to a second-level F-test across the first-level PEs of
interest (for example the ones you described) but this isn't always very
useful as it insists that the balance between the basis functions (etc) is
the same in all first-level sessions, ie doesn't test for the _power_
across subjects (etc) but for any variation, including the balance between
EVs.
b) You could take the first-level zfstats and do simplified tests on them
yourself, after resampling into standard-space. For example,
sum(Z)/sqrt(n) gives simplified fixed-effects, whilst doing simple OLS on
Z stats probably gives some kind of random-effects-related analysis
(though maybe don't quote me on that!)
c) If you are not TOO concerned about phase variations in response to your
stimulation it might be best to simply carry the primary sinusoid's PE up
to second level...
This problem is similar to the one of comparing submodels of basis
functions at first-level (eg two original EVs with 3 basis functions each,
causing 6 real EVs). They way people do this at the moment is to pair-wise
contrast the basis functions and do an F across these 3 contrasts, but
this doesn't ask the simple question "how does the power in one submodel
compare with the other".
Thanks, Steve.
On Thu, 16 Jan 2003, Hauke R. Heekeren wrote:
> Hi there,
> Is there a way to bring f-test results to a higher level?
> As a practical example, our input function is sinusoidal (32s) and I
> analysed a couple of data sets with feat (included two harmonics in the
> model).
> I'd now like to compute a 2nd level analysis (across subjects) based on
> the zfstat maps.
>
> Any comments on whether and how this works are greatly appreciated,
>
> Thanks,
>
> Hauke
>
Stephen M. Smith
Head of Image Analysis, FMRIB
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
|