Dear Marco,
> 1): In SPM99 when I specific and estimate a model I always get the message: "gray -> beta
> not uniquely specified".
...
> I can't find why I have the message "beta not uniquely specified" that I have with
> different model too. Is this
> message a error message or only a warning message?
Your model is perfectly ok. The text 'gray -> beta not uniquely
specified' just means that if you see some grey boxes to the left of
this text, then the parameters to be estimated are not uniquely
specified. In your case, the boxes are white and this message just
indicates that everything is fine...
>
> 2): What does mean "remove global effects" and when I must use these option?
> (I see that, sometime, activations map is strong depending on the use of these option)
In SPM99 for fMRI you can apply proportional scaling to the images, i.e.
each image is divided by an estimate of its global mean intensity prior
to the statistical analysis. I think it's usually a good thing to do,
especially if you analyse several sessions. For a single session it's
maybe not that important, because much of the (low-frequency) effect,
which the prop. scaling would take care of, is also removed by the
high-pass filter.
>
> 3): Does high-pass filter and low pass-filter modify sample data or only the model?
> (I think the last one)
Both filters (in combination called a bandpass filter) are applied to
the data (and the model) before estimating the parameters.
>
> 4): In "realign" what does mean "adjust sampling errors" and when I must use these option?
I leave this question to John!
Stefan
--
Stefan Kiebel
Functional Imaging Laboratory
Wellcome Dept. of Cognitive Neurology
12 Queen Square
WC1N 3BG London, UK
Tel.: +44-(0)20-7833-7478
FAX : -7813-1420
email: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|