Hi SPM Experts
I have a possibly sill question about group level factorial analysis
in SPM and I was wondering if anyone can help me out.
I am trying to run a group level analysis now. The subjects are
divided into 3 groups (CON AND 2 PAT) and for each patient there are
two task conditions. I have run a PPI analysis from seed to whole
brain for each subject for each condition. This works the same as task
based analysis
For my group level analysis I am trying to ask the following questions:
1) which brain regions are significantly more activated in group 1
compared to group 2 and in group 2 compared to group 3 (group 1 >
group 2 > group 3) in condition A minus B. I have set up a full
factorial with three groups & two conditions. I am looking at the
contrasts -1 1 0 0 1 -1 (with the middle group being modelled
implicitly). Is this correct for this purpose?
2) which brain regions show significant group X condition interaction.
For this I am looking at the f contrast group X condition created by
SPM. Is this valid? If so then there are no cluster level stats so how
do I either derive these or determine which clusters are significant
at cluster level?
I would be extremely grateful for any pointers
Sincerely
Prerona
On Wed, Apr 2, 2014 at 10:31 PM, Prerona Mukherjee
<[log in to unmask]> wrote:
> Hi Everyone
>
> I am having an issue with a between group PPI analysis using the gPPPI
> toolbox and I was wondering if anyone could help me.
>
> I have run the analysis from one seed for a group of controls and a group of
> patients. Now I want to run the second level analysis between these two
> groups. However I am getting an error saying no inmask voxels.
>
> I know that this means there are no resulting voxels. But I was wondering if
> anyone could help me figure out why I am getting this error and what I can
> do about it
>
> I have tried looking at all the individual con files (by opening them using
> check reg) and they seem okay (although there are a few very dark spots on
> some of them). I was wondering if this is causing the problem
>
> To set up the second level design I have taken all the con files created by
> the first level PPI and tried to do a 2 sample t test.
>
> I am not sure what is the correct masking option. I have tried both explicit
> (with a mask) and implicit mask. I have also tried changing the spm defaults
> (
> defaults.mask.thresh = 0.8 to 0.2 and also to -Inf).
>
> I have also tried doing a within group analysis of each of the two groups
> and ran into the same problem
>
> PS: The PPPI toolbox creates an SPM.mat file in the output directory and a
> mask.img. This mask.img also looks okay and the SPM.mat looks like it is
> including every subject together in one group
>
> Do you have any thoughts about this? I would really appreciate any help or
> advice
>
>
> Sincerely
>
> Prerona
|