Dear Yu-Shin,
I am no expert in VBM, but I assume here that it is in general
comparable to fMRI 2nd level analysis.
I think what you are looking at is the whitening of your design,
correcting it for dependencies or iid violations, which indeed depends
on the degree of depencies in your data (and hence your design matrix
will be slightly changed after a first pass of GLM estimation when
estimating this dependency).
Usually this is what you want.
But correct me if I'm wrong, I never ran any VBM analysis.
Cheers,
Bas
Yi-Shin Sheu schreef:
> Dear SPM experts,
>
> This question might be a bit strange, but it has been bothering me for a while.
>
> I am doing VBM analysis; basically, I compare the differences of gray matter
> volume between groups.
>
> My design is the "two-sample t-test", in which the input scans are normalized and
> smoothed T2 maps, and I also model some non-interest covariates.
>
> What bothers me is that, most of the time when I conduct the analysis, the first column
> and second column which represent the group are normal as below.
> 1 0
> 1 0
> 1 0
> 0 1
> 0 1
> 0 1
>
> However, sometimes when I process other subjects in another study with exactly the
> same procedure and parameter, the first two columns will become uneven
> number, an example is shown below. Furthermore, it does not happen when I specify the
> model, only happens after I hit "estimate".
>
> 1.05 0
> 1.05 0
> 1.05 0
> 0 0.95
> 0 0.95
> 0 0.95
>
> I just can't figure out what caused the number difference. In my knowledge, the first two
> column should be always composed of 1 and 0.
>
> Can anyone tell me what is wrong and will that effect the accuracy of my analysis?
>
> A screenshot of my design matrix is attached.
>
> Thank you
>
> Yi-Shin
>
>
> ------------------------------------------------------------------------
>
>
> ------------------------------------------------------------------------
>
--
--------------------------------------------------
Dr. S.F.W. Neggers
Division of Brain Research
Rudolf Magnus Institute for Neuroscience
Utrecht University Medical Center
Visiting : Heidelberglaan 100, 3584 CX Utrecht
Room B.01.1.03
Mail : Huispost B.01.206, P.O. Box 85500
3508 GA Utrecht, the Netherlands
Tel : +31 (0)88 7559609
Fax : +31 (0)88 7555443
E-mail : [log in to unmask]
Web : http://www.fmri.nl/people/bas.html
--------------------------------------------------
|