Dear SPMers,
If independent variables A and B are both correlated with dependent variable D, and both independent variables are also mutually correlated, a multiple linear regression analysis (MRA) is mandatory to clarify which independent variable contributes most (or maybe both independent variables make their contribution) to D.
In my experiment, I have 20 subjects. I found that, by simple regression analysis in SPM5 by entering either A or B of each subject as covariates, some brain regions were correlated both with A and B, despite others were specifically activated by A or B.
In order to understand which brain region’s activity is contributed by A or B, or both, I then tried to perform a MRA in SPM5 by entering A and then B of each subject as covariates. After model estimation, I am not sure how to set the contrast to display brain regions as below:
(1) Whose activations were contributed by A solely
(2) Whose activations were contributed by B solely
(3) Whose activations were contributed by both A and B
My way to get that is:
Contrast 1,0 for (1)
Contrast 0,1 for (2)
… Am I right? But I have no idea about the contrast for (3)
Can anyone help me?
Thanks a lot.
|