Dear SPM'ers
We have three fMRI runs box-car style on a group of 10 subjects
run1:semantic decision (S1) and low level control (c1)
run2:semantic decision (S2) and same low level control (c2)
run3:phonologic decision (P) and same low level control (c3)
(the actual order was counter-balanced)
these have all been condensed using the random-effects kit to 2 condition
images per run yielding 6 images per subject.
I want to know what is active in S with reference to P.
Can these mean images be treated as conditions in a multi-subject different
conditions analysis, or do I need to use multi-study different conditions?
What is the best way to model the difference I am looking for?
I am worried about the weighting of the two semantic tasks to the single
phonemic task.
If by interaction what weighting?
S1 c S2 c P c
1 -1 1 -1 -1 1
or
1 -1 1 -1 -2 2
Masking:
Is it possible to use a mask to show activity in (S) combined (1 -1 1 -1 0
0) versus P (0 0 0 0 1 -1). If so how do i set this up? The mask feature of
SPM96 seems to show areas in common with the mask rather than outside of the
mask. Is there a way around this?
I have also done a conjunction using the simple effects of each task. Again,
I am worried that the weighting the semantic tasks as separate studies or
conditions is overpowering the phonologic task.
Is it legal to average the 2 mean semantic images to a single condition
image?
thanks for your thoughts.
+++++++++++++++++++
Sterling C. Johnson, Ph.D.
Brain Imaging Lab
Psychiatry
Dartmouth Hitchcock Medical Center
Lebanon, NH 03756
phone: 603-650-5824
fax: 603-650-5842
[log in to unmask]
+++++++++++++++++++
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|