Shashwath Meda wrote:
> Dear Will,
>
> Sorry about not being clear about the problem.
>
> I would try to express my concern in a different manner. The issue is I
> first design a matrix for a 2-sample t-test for two groups g1 and g2 by
> selecting scans for group g1 first and then followed by scans for g2,
> further i input 1's for g1 and 0's for g2.
OK. This is the bit I don't understand.
You're entering extra regressors ? Does this mean
you have 4 regressors + constant ?
Are you using SPM5 ?
If so, just use the 'Two-sample T-test design'.
Enter scans for each group.
You should get a design matrix with 2 regressors + constant.
Best,
Will.
I then give a contrast of 1
> -1 (g1>g2) and -1 1 (g2>g1) to yield t-map's of activation pattern say A
> and B respectively. Then i tried to redesign my matrix by inputting 0's
> for g1 and 1's for g2 to see how it behaves, this time though using the
> contrast 1 -1 gives me activation pattern B and a contrast of -1 1
> yields me activation pattern A. I am little confused about how to
> report the results now, do i associate activation pattern A with g1>g2
> or g2>g1. Hope I have made my problem clear. thanks for taking the time
> out and please let me know if any part is not being clear.
>
> Thanks again.
>
> with regards,
> Shashwath.
> On 2/24/06, Will Penny <[log in to unmask]
> <mailto:[log in to unmask]>> wrote:
>
> Dear Shashwath,
>
> Its still not clear to me what the problem is.
>
> Is the reversal you describe to be expected or not ?
>
> Perhaps you could you try explaining the problem in a different way.
>
> Best,
>
> Will.
>
>
>
> Shashwath Meda wrote:
>
> > I actually was referring to the 1's and 0's that we utilise in our
> > design matrix. I remember reading somewhere on the internet where
> they
> > had used the terminology dummy variables to represent these. So
> thought
> > this might be the nomenclature used.
> >
> > On 2/24/06, Will Penny < [log in to unmask]
> <mailto:[log in to unmask]>
> > <mailto:[log in to unmask]
> <mailto:[log in to unmask]>>> wrote:
> >
> >
> >
> > Shashwath Meda wrote:
> >
> > > Dear Dr.Penny,
> > >
> > > I had an simple issue that has been bothering me for a
> couple of
> > days,
> > > thought you would be the best person to seek some advice.
> I am
> > trying to
> > > do an optimised VBM analysis comparing 2 different
> genotypes (g1
> > and g2)
> > > in a bunch of healthy controls.
> > >
> > > I performed a basic 2 sample t-test with a "1-0"
> (g1denoted as
> > 1's-g2
> > > denoted as 0's) design matrix to differentiate my groups
> and my
> > > contrasts for this design were 1 -1 (gi>g2 - activation
> pattern
> > A) and
> > > -1 1 (g2>g1 - activation pattern B).
> > >
> > > Later I tried using a 0-1 design matrix (denoted g1 as
> 0's and
> > g2 as
> > > 1's) with the same contrasts as above and found that my
> patterns
> > A and B
> > > were interchanged.
> >
> >
> > I am confused about the "dummy variables" used here
> >
> > I'm not sure what you mean by dummy variables.
> >
> > > ("1-0 design") what is the exact implication of these
> variables on my
> > > design matrix? why does the changing of "dummy variables"
> produce
> > such a
> > > reversal in my final activation maps? Maybe i am not
> completely sure
> > > about the specifics of the design matrix. Hoping you could
> shed
> > > some light on the issue, to make it possible for me to
> interpret my
> > > results better.
> > >
> > > Thanks a lot in advance.
> > >
> > >
> > > --
> > > with regards,
> > >
> > > Shashwath meda
> > >
> > > ***********************************************
> > > Senior Research Assistant,
> > > Olin Neuropsychiatry Research Center
> > > Hartford Hospital/IOL.
> > > Hartford, Connecticut 06106.
> > >
> > > Cell: (313) 505-6847
> > > Off: (860) 545-7804
> > >
> > > ***********************************************
> >
> > --
> > William D. Penny
> > Wellcome Department of Imaging Neuroscience
> > University College London
> > 12 Queen Square
> > London WC1N 3BG
> >
> > Tel: 020 7833 7475
> > FAX: 020 7813 1420
> > Email: [log in to unmask]
> <mailto:[log in to unmask]> <mailto: [log in to unmask]
> <mailto:[log in to unmask]>>
> > URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
> >
> >
> >
> >
> > --
> > with regards,
> >
> > Shashwath meda
> >
> > ***********************************************
> > Senior Research Assistant,
> > Olin Neuropsychiatry Research Center
> > Hartford Hospital/IOL.
> > Hartford, Connecticut 06106.
> >
> > Cell: (313) 505-6847
> > Off: (860) 545-7804
> >
> > ***********************************************
>
> --
> William D. Penny
> Wellcome Department of Imaging Neuroscience
> University College London
> 12 Queen Square
> London WC1N 3BG
>
> Tel: 020 7833 7475
> FAX: 020 7813 1420
> Email: [log in to unmask] <mailto:[log in to unmask]>
> URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
>
>
>
>
> --
> with regards,
>
> Shashwath meda
>
> ***********************************************
> Senior Research Assistant,
> Olin Neuropsychiatry Research Center
> Hartford Hospital/IOL.
> Hartford, Connecticut 06106.
>
> Cell: (313) 505-6847
> Off: (860) 545-7804
>
> ***********************************************
--
William D. Penny
Wellcome Department of Imaging Neuroscience
University College London
12 Queen Square
London WC1N 3BG
Tel: 020 7833 7475
FAX: 020 7813 1420
Email: [log in to unmask]
URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
|