Dear Matthew, dear SPM Marsbar users,
after intense efforts to understand what has been suggested we now are stuck again
and hence describe our problem in more detail:
Our goal was to obtain a statistic on the second level over a ROI, that has been
extracted from an activated area (in another task == Task0). We then wanted to
find out several things about the activation in this area:
1. We would like to know how many voxels are activated inside the ROI.
2. We would like to know about the significance level in this activated area.
What we did:
1. extract from our Task0 the activated areas using MarsBar (no problem)
2. Within MarsBar select basic models in the Design section
- one sample t-test over all contrasts from 1st level analysis
- MarsBar asks here for the ROI.mat which we selected.
- Summary function: mean (default)
3. Within MarsBar: Results-> statistic table-> define new contrast -> t-contrast
(1)
message Stats Done.
This is the point where we are lost: how can we now get the number of activated
voxel etc... ?
Please let us know what we are doing wrong or where we do not understand the
concept of MarsBar.
Thank you again in advance,
F.Castrop P.Erhard
Matthew Brett wrote:
> Hi,
>
> I've a feeling that the buck stops with me with help on MarsBar. So, buck in
> hand...
>
> Second level analyses with MarsBar:
>
> Two approaches:
> 1) Define the ROIs on single subjects; analyze with MarsBar; Run the Statistic
> table command on your chosen t contrast, and extract the contrast value.
> Repeat for each subject; run a single sample t test on the values for the
> single subjects, maybe using SPSS.
>
> 2) Run single subject analyses on normalized data in SPM, and get the relevant
> contrast images for each subject; define ROI, then run MarsBar random effects
> model with ROI, and using the contrast images for each subject for scans -
> this is the standard SPM approach.
>
> I hope that's clear, but please do let me know if not.
>
> Could you give any more details on the problems with the m files?
>
> Thanks a lot,
>
> Matthew
|