Hi David,
yes the number of rows should equal the number of
subjects you have in your 3 groups. For the F test
using randomise, please look at Tom emails from the
22nd of Februrary on this exact subject.
Hope this helps,
Gwenaëlle
--- David Kideckel <[log in to unmask]> a écrit
:
> HI Tim,
>
> One more thing. Will the number of inputs be equal
> to the number of subjects
> I have?
> David
>
> On 2/28/08, David Kideckel
> <[log in to unmask]> wrote:
> >
> > Hi Tim. I am trying to set up the design matrix
> and am getting errors. I
> > am trying to compare 3 groups (as described
> above), and this is the info I
> > have included in the design matrix:
> > 3 EV'S with 6
> > inputs: the first 2 inputs are 1 0 0, the second
> two are 0 1 0, and the final 2 are 0 0 1 (as per
> your email above).
> >
> > The contrasts include 3 contrsats and 1 F-test.
> The first contrast is 1 -1 0, the second contrast is
> 1 0 -1, and the third is 0 1 -1 (as per your email
> above). All 3 C's are checked with yellow in the
> Glm_gui, but the F's are not.
> > Can you please tell me why it isn't working. The
> error message when I
> > click View Design
> > is: "problem with processing the model: F-test 1
> isn't valid - each included contrast cannot be a
> linear combination of the others. child process
> exited abnormally"
> >
> > I look forward to hearing from you.
> >
> > David
> >
> > Right. Can I set up each TBSS script (ie TBSS1...
> >
> > On Jan 30, 2008 12:39 PM, Tim Behrens
> <[log in to unmask]> wrote:
> >
> > > Hi
> > > On 30 Jan 2008, at 17:32, David Kideckel wrote:
> > >
> > > Thank you for the response. How exactly do I do
> this. Do I have to run
> > > all TBSS preprocessing steps before this though,
> and place all subjects' FA
> > > maps into the original file folder?
> > >
> > >
> > > Yes
> > >
> > >
> > > Once I get to the design matrix, can I set up a
> script that will allow
> > > me to do this 3 group comparison with a cluster
> threshold of 3?
> > > Cheers.
> > >
> > >
> > > What question do you want to ask of your three
> groups?
> > > The simplest one you can ask is where are there
> any differences between
> > > any of the groups.
> > > For this, you design matrix looks like
> > > 1 0 0
> > > 1 0 0
> > > ...
> > > 0 1 0
> > > 0 1 0
> > > ...
> > > 0 0 1
> > > 0 0 1
> > >
> > > your contrast looks like
> > >
> > > 1 -1 0
> > > 1 0 -1
> > > 0 1 -1
> > >
> > > and then you compute an F-statistic across all
> three contrasts.
> > >
> > > You can then use randomise with whatever
> F-threshold you want using -F
> > >
> > > Cheers
> > >
> > > T
> > >
> > >
> > >
> > > David
> > >
> > > On Jan 30, 2008 12:26 PM, Tim Behrens
> <[log in to unmask]> wrote:
> > >
> > > > You can set up a GLM with an EV for each
> group, and test any
> > > > comparison you like with the contrast matrix -
> all in one TBSS run.
> > > >
> > > >
> > > > T
> > > > On 30 Jan 2008, at 17:21, David Kideckel
> wrote:
> > > >
> > > > > Dear FSLers,
> > > > >
> > > > > Can one compare 3 groups in TBSS all at
> once, or is it a
> > > > > requirement of TBSS to make three separate
> comparisons, essentially
> > > > > running three separate TBSS analyses to
> obtain all possible
> > > > > comparisons, i.e., CON vs EXP1, CON vs EXP2,
> and EXP1 vs EXP2.
> > > > >
> > > > > Thank you.
> > > > >
> > > > > David Kideckel
> > > >
> > >
> > >
> > >
> >
>
_____________________________________________________________________________
Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail http://mail.yahoo.fr
