Hi Antonio,
> I would like to know the options available for doing group analysis when there is
> more than one factor involved.
>
> In this case I have 2 groups and, within each group, there are 2 other factors (i.e. the
> eye that was stimulated and the type of stimulus used). Each of the factor level
> combinations was done as a single experiment. I would like to look at, for example,
> interactions between the group factor and the other two.
>
> I was contemplating using the zstat1 image as input for AFNI's 3DANOVA. Is there
> another way to go?
Hi - you can use any simple stats at the second or third level; one
example is the third-level command-line example I sent in to the FSL list
a couple of days ago. Or feeding stats into AFNI's 3DANOVA would be fine
too. If you want the equivalent of the random effects analysis
you should feed in cope images not zstats. the other thing you'll need to
do is get them into standard space as well. If you've already done a
second-level analysis they will already be in standard space, otherwise
get them in using eg
cd <whatever>.feat
flirt -in stats/cope1 -ref standard -applyxfm -init \
example_func2standard -out cope1_2_standard
Hope this sorts it! By the way - the new long-promised version of FEAT
(end of September we hope) will allow general GLM at the second level (ie
will do what you want) as well as having a sophisticated way of estimating
the random effects variance in this model.
ttfn, Steve.
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
|