Hi everyone!
I am reposting this query again. I have been looking for some inputs for the longitudinal VBM study consisting of 3 groups (1-patients, 2-high risk population and 3-healthy controls) and 2 time points using a flexible factorial design. Some materials are available for 2 groups and 2 to 3 time points dataset in SPM. However, no documentation/support is available for the 3 groups and 2 time points (3x2) statistical design. Although constructing a design matrix for 3x2 looks pretty much straightforward, creating a contrast matrix (F-contrast) for testing the different study hypothesis looks complex and more prone to error. Any valuable inputs/suggestions will be greatly appreciated!
Flexible factorial design
Subjects: [1 1;1 2] [2 1;2 2] [3 1;3 2]
Main effects and interaction:
Main effects -subject(1)
Main effects -group(2)
Main effects -time(3)
Interaction effects: group x time [2 3]
The first three main research questions in my longitudinal VBM study (3x2) are the following. Please see the screenshot of the design matrix attached.
1.) Main effects of group
Eg. patients<high risk population<healthy controls (reduced grey matter volume in patients and high risk population)
-1 -1 1 zeros(1,2) -ones(1,2)/2 -ones(1,2)/2 -ones(1,25)/25 -ones(1/21)/21 ones(1/20/20)
2.) Main effects of time
Eg. Time 1 > Time 2
zeros(1,2,3) [1 -1] 25/66*[1 -1] 21/66*[1 -1] 20/66*[1 -1] zeros(1,66)
3.) Interaction effects of group by time
zeros (1,3) zeros (1,2) [1 -1] [-1 1] zeros (1,66)
sincere thanks,
Varun
|