help
help > RE: Connectome MVPA and mixed contrast
Dec 11, 2013 01:12 AM | Alfonso Nieto-Castanon - Boston University
RE: Connectome MVPA and mixed contrast
Hi Emanuele,
some thoughts on your questions below
Best
Alfonso
Originally posted by Emanuele Pravata':
Yes, exactly.
Each MVPA map represents an individual component of a principal component decomposition of the (voxel-specific) voxel-to-voxel covariance. Rather than analyze them separately you should treat them jointly as a multivariate representation of the connectivity between each voxel and the rest of the brain (in practice this means that you should select all of them in the "measures" list and then enter a 'between-measures contrast' of eye(n) -in your case [1 0 0 0;0 1 0 0;0 0 1 0;0 0 0 1]-, with whatever combination of 'between-subjects' and 'between-conditions' contrasts that you wish).
3) Is it already possibile to reference this method?
Not yet, sorry, but I am hoping to send this manuscript out before the end of the year!
some thoughts on your questions below
Best
Alfonso
Originally posted by Emanuele Pravata':
Hi again!
I have some questions regarding Connectome MVPA:
1) Is it possible to apply the following contrasts, that you suggested in my previous Mixed ANOVA post, to investigate the interaction of group x session in a 3 x 3 design?
[1 -1 0; 0 1 -1] for subjects effects (3 groups) and [1 -1 0; 0 1 -1] for conditions (3 sessions)
I have some questions regarding Connectome MVPA:
1) Is it possible to apply the following contrasts, that you suggested in my previous Mixed ANOVA post, to investigate the interaction of group x session in a 3 x 3 design?
[1 -1 0; 0 1 -1] for subjects effects (3 groups) and [1 -1 0; 0 1 -1] for conditions (3 sessions)
Yes, exactly.
2) What the different MVPAs 1,2,3,4 refer to?
What's their meaning?
Each MVPA map represents an individual component of a principal component decomposition of the (voxel-specific) voxel-to-voxel covariance. Rather than analyze them separately you should treat them jointly as a multivariate representation of the connectivity between each voxel and the rest of the brain (in practice this means that you should select all of them in the "measures" list and then enter a 'between-measures contrast' of eye(n) -in your case [1 0 0 0;0 1 0 0;0 0 1 0;0 0 0 1]-, with whatever combination of 'between-subjects' and 'between-conditions' contrasts that you wish).
3) Is it already possibile to reference this method?
Not yet, sorry, but I am hoping to send this manuscript out before the end of the year!
Threaded View
| Title | Author | Date |
|---|---|---|
| Emanuele Pravata' | Dec 4, 2013 | |
| Alfonso Nieto-Castanon | Dec 11, 2013 | |
| Emanuele Pravata' | Mar 26, 2014 | |