Hi Darren,
if I remember correctly, you should be able to enter multi-line contrast
vectors for F contrasts using MATLAB syntax (i.e. write everything in
one line and use a semicolon ; as a matrix line separator)
Volkmar
On Tue, 2007-10-16 at 07:26 -0500, Darren G wrote:
> Hi Jan
>
> Thank you.
>
> My problem is how to make efficient use of this in the batch or even the
> conman interface
> particularly for T contrasts. I have an analysis in which there are 4
> conditions x 10 FIR basis functions x 4 sessions.
>
> If I use the batch facility for an F-test I will have to set up 4 F
> contrasts to which I have to add 10 vectors each, since batch does not allow
> multiline entries. I can then use the replicate function to get all the
> sessions, however, there is great chance for error in specifying the 40
> vectors.
>
> If I use the batch facility for a T-test I will have to set up 40 individual
> t-vectors. I could use the T (cond/sess) entries, which will cut down on
> errors by allowing me to specify Cond#/ Basis# but it's still a lot of
> entries.
>
> Anyway, being lazy, I was hoping that there was some easy/automatic way to
> have SPM output all the individualt-contrasts across all condtions and bins
> which can then be forwarded to the second level. I'm probably being dense
> about this but I don't see any automatic way to do it.
>
> thanks
> darren
>
>
> ==============Original message text===============
> On Mon, 15 Oct 2007 8:09:51 pm CDT Jan Gläscher wrote:
>
> Hi Darren,
>
> I think you can play around with Kronecker tensor product. This has been
> very useful to me in the past for FIR design matrices and contrasts
>
> For instance, using vectors, you can do the following:
>
> >> kron([1 2 3 4],[1 0])
>
> will give you
>
> [1 0 2 0 3 0 4 0]
>
> But it also work with matrices:
>
> a = eye(2);
> b = [1 2;3 4];
> kron(a,b)
>
> will give you
>
> 1 2 0 0
> 3 4 0 0
> 0 0 1 2
> 0 0 3 4
>
> If you play around with it a bit, you'll probably find a way to efficiently
> specify all the contrasts you need.
>
> Cheers,
> Jan
>
>
> d gitelman wrote:
> > Hi
> >
> > Is there a simple way to automatically create the t-contrasts for analyses
> > using an FIR basis set as opposed to entering them all manually?
> >
> > thanks
> > darren
> >
> >
> > ----------
> > Darren Gitelman, MD
> > Department of Neurology
> > Northwestern University
> > voice: (312) 908-8614
> > fax: (312) 908-5073
> > page: (312) 695-1849
> > email: [log in to unmask]
> > ----------
> >
>
--
Volkmar Glauche
Freiburg Brain Imaging
http://fbi.uniklinik-freiburg.de/
Phone +49(0)761 270-5331
Fax +49(0)761 270-5416
|