Hello-
I have a study where I'm using the GLM gui and flameo to run stats on
several groups although, for discussion, I'm only interested in a
comparison of methods for 5 subjects. I've checked the cope 4D file
that I feed flameo to verify that the ordering of the inputs is the same
(the first 5 inputs in the 4D file). zthresh and pthresh are the same
and "dlh" and "volume" are properly calculated (from smoothest) to feed
into cluster.
For one group, mean only, I have the following setup:
EVs Tab:
/ContrastName1 group mean
/NumWaves 1
/NumContrasts 1
/PPheights 1.000000e+00
/RequiredEffect 25.033
/Matrix
1.000000e+00
---------------------------------------------------------------
Contrasts & F-tests Tab:
/NumWaves 1
/NumPoints 5
/PPheights 1.000000e+00
/Matrix
1.000000e+00
1.000000e+00
1.000000e+00
1.000000e+00
1.000000e+00
For another setup, two groups where one group is the same as above (and,
in this case, I'm interested in the group mean only):
EVs tab:
/NumWaves 2
/NumPoints 11
/PPheights 1.000000e+00 1.000000e+00
/Matrix
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
---------------------------------------------------------------
Contrasts & F-tests Tab:
/ContrastName1 Amean
/ContrastName2 Bmean
/ContrastName3 A>B
/ContrastName4 B>A
/NumWaves 2
/NumContrasts 4
/PPheights 1.000000e+00 1.000000e+00 1.000000e+00
1.000000e+00
/RequiredEffect 4.414 4.029 5.976 5.976
All values in the ".con" files are what the gui populated for me. For
the second test, they are set as group1 and group2, with group1 being
the same 5 subjects that were used in the first example.
If I look at the outputs, the mean image is different. In the first
case my output is Lhemi only, but in the second, there is a significant
portion of the Rhemi that is introduced. Clearly I'm misinterpreting
the meaning of the contrast in the latter case, or my assumption that I
should be getting the same group mean is incorrect. Could someone
please shed some light on this situation?
Thanks,
Joel
|