Hi,
Yes you can analyse all these models with "randomise". You can use
the Glm GUI to setup the models (you don't need to use the whole FEAT
GUI).
The only issue is how to deal with the regressors of no interest. The
current version of randomise doesn't have a full proper
implementation of this - the best you can do is put the regressor of
no interest in a separate design matrix file and use the -X option in
randomise - but this will be dealt with much more nicely in the
version in the next release of FSL - if you can hold on a few days :)
Cheers, Steve.
On 3 Aug 2007, at 11:13, Karen Caeyenberghs wrote:
> Dear TBSS users,
>
>
> I have been running the TBSS scripts. I have obtained good (and
> significant ;-))results.
> I was just wondering: Can I perform a two-sample unpaired t test
> with age
> as covariate?
> Also, in the article of Smits et al. (2006) in Neuroimage, they have
> mentioned several examples, where they
> perform correlations within one group, using permutation-based
> inference on
> cluster
> size. How can I perform such kind of analyses? I expect they use
> the FEAT
> toolbox within
> FSL (before running randomise, creating design.con/.mat files).
>
> Specifically: I want to perform the following analyses:
> - a two sample unpaired t test (controls vs patients) with age as
> covariate
> - a correlation between behavioral test results and FA for the two
> groups
> (cons and pats)
> - a correlation between behavioral measures and FA within a group
>
> Best regards,
>
> Karen
------------------------------------------------------------------------
---
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
------------------------------------------------------------------------
---
|