Hi - I think the bigger question is whether you want to normalise the
PE/COPE values into % change or use the raw values. If you do decide
to use the % values then the implication of my previous reply is that
if you use COPEs, then statistical interactions with non-orthogonal
regressors (such as motion confound covariates, if used), will be
taken into account.
Cheers.
On 15 Nov 2007, at 15:34, Esther Meerman wrote:
> Dear Steve,
>
> Thank you so much for your answer. Although I am still not sure if I
> should use the PE or COPE values of the participants to calculate
> the correlations between different brain area's. Do you have any
> sugestions?
>
> Thank you so much!
>
> Greetings,
>
> Esther
>
> On Nov 15, 2007 2:27 PM, Steve Smith <[log in to unmask]> wrote:
> Hi,
>
> This is a good question; you might well expect PE1 and COPE1 (if it
> just uses PE1) to give the same answer. In certain simple designs,
> they will.
>
> However, when you have partial correlation between EVs, the effective
> regressor height is changed, and this has a direct impact on the %
> signal change calculations (see Smith, NeuroImage, 2007 for
> mathematical details). Hence the COPE result is different from the PE.
> You can see the estimated effective contrast heights in the design.con
> file.
>
> Note that you only see this difference when you turn on the % change
> option - the raw PE and COPE values are still identical.
>
> Cheers, Steve.
>
>
> On 15 Nov 2007, at 11:21, [log in to unmask] wrote:
>
> > Dear all,
> >
> > I have a question about the results that I got with the featquery
> > option of
> > FSL. In my study I have 6 PE's with their temporal derivatives and I
> > applied motion correction. The COPE's of my study are partly the
> > same as my
> > PE's or they are the substraction of certain PE's. But PE1 is the
> > same as
> > COPE1 and PE3 is the same as COPE2.
> > When I run the featquery over the first-level feats, I get different
> > results for PE1 and COPE1. However, the max data that featquery uses
> > to
> > calculate the min, max, mean etc. is the same for the PE1 and COPE1.
> > Can
> > anyone explain this difference in results? I have used the convert
> > PE/COPE
> > in % option for the featquery. And if I want to calculate the
> > correlation
> > of mean signal change, do I use the mean PE or mean COPE?
> >
> > Thanks!
> >
> > E.E.Meerman
> > email: [log in to unmask]
>
>
> ---------------------------------------------------------------------------
> 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
> ---------------------------------------------------------------------------
>
---------------------------------------------------------------------------
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
---------------------------------------------------------------------------
|