Hi, Gang.
I have a question regarding the output for glt tests. Consider the following model and the associated output report for sub-brick 3:
Model:
3dLME -prefix /home/thomaskm_02/huntrh/ESMF_3dLME_TEST/TEST_5/ESMF_3dLME_TEST_5_OUT_1 -jobs 24 \
-model "cond" \
-ranEff '~1' \
-SS_type 3 \
-num_glt 3 \
-gltLabel 1 'emot-shape' -gltCode 1 'cond : 1*emot -1*shape' \
-gltLabel 2 'emot' -gltCode 2 'cond : 1*emot ' \
-gltLabel 3 'shape' -gltCode 3 'cond : 1*shape' \
-dataTable @/home/thomaskm_02/huntrh/ESMF_3dLME_TEST/TEST_5/DataTable.dat \
Output:
-- At sub-brick #3 'emot-shape Z' datum type is short: -32767 to 0 [internal]
[* 8.76054e-05] -2.87057 to 0 [scaled]
statcode = fizt
The output is labelled with a "Z", but the values in the output image exactly match my hand computed t-test values. Are you sure the output is supposed to be in z? Or am I mis-understanding something?
If the output image is in t-values, then one can always (assuming we know the dof) convert to z-values using:
3dcalc -a t_image.nii.gz -expr 'fitt_t2z (a,dof)' -prefix z_image.nii.gz
On that note, I notice that the output table has the dof for the F-test, but not for the t(z)-test. Is there a reason for the difference?
Thanks,
Ruskin
