Dear folk,
I'm new to second level analyses using fMRI.
I have 11 subjects who were all entered into a single design matrix for the
first level analysis. I created the appropriate contrast image for each subject
and then entered these into a simple models, one sample t test.
The peak voxel of my main activation has a T value of 8.7 and a Z score of 4.54,
but a corrected p value of only 0.368 (presumably due to only having 10 degrees
of freedom).
Is the convention for reporting results in RFX models the same as 'old-style'
fixed effects models ie: report Z scores at corrected thresholds (assuming no
prior regional hypothesis).
Thanks for your thoughts.
Cheers.
Alex.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|