---------- Forwarded message ----------
From: Wu Ming-ting <[log in to unmask]>
Date: 2007/5/17 ¤U¤È 11:38
Subject: Re: JISCmail Data Protection policy
To: [log in to unmask]
Dear SPM list:
We are performing a VBM analysis on gray matter tissue from two
conditions (C1 : no disease and C2: with disease).
For the two groups, we have a categorical covariant name "S".( 1
=male, 2=female)
Our goal is to see whether S interacts with condition to make the
results different.
We used a "single subject conditions and covariates" model.
We selected 2 conditions (scans from C1 and scans from G2), and one
covariate of interest "S".
We then select the condition by covariate interaction option.
The contrast matrix is shown in the attached jpg file.
Now, it's the interpretation of contrast makes us very confused.
We think that the following contrast designs can be read this way:
1,-1,0,0 = difference between C1 and C2 (C1-C2)
0,0,1,0 = the effect of being female on C1? or what?
0,0,0,1 = the effect of being female on C2? or what?
0.0.-1,1= the effect of being female with C2 than C1?
Are we right in reading the contrast, or we make the whole process wrong?
Sincerely
§d»Ê®x
Ming-Ting
=======================
Ming-Ting Wu, MD
Chief,
Section of Thoracic and Circulation Imaging
Department of Radiology
Kaohsiung Veterans General Hospital
Associated Professor of Radiology
National Yang Ming University
[log in to unmask]
FAX 886-7-3468301
--
Sincerely
§d»Ê®x
Ming-Ting
=======================
Ming-Ting Wu, MD
Chief,
Section of Thoracic and Circulation Imaging
Department of Radiology
Kaohsiung Veterans General Hospital
Associated Professor of Radiology
National Yang Ming University
[log in to unmask]
FAX 886-7-3468301
|