Posted by Phoebe on March 14, 2001 at 18:02:48:
Hello all,
I have some questions in using 3dDeconvolve to calculate t-test.
I got a 3d time dataset (bold+orig) which is collected with the following paradigm:
Epoch No. No. of time points On/off
1 20 99999
2 20 0
3 20 10
4 40 0
5 20 10
6 20 0
7 10 99999
What I would like to do is to compare the 'on' and 'off' states by t-test for this particular dataset, instead of using 3dfim+. Then I type in the following command:
3dDeconvolve -input bold+orig -nfirst 20 -nlast 139 -polort 1 -num_stimts 1 -stim_file 1 para.1D -stim_label 1 "para" -stim_maxlag 1 20 -fout -rout -tout -bucket bold.buck
and in Define Function, switch the Func to 'base t^1 coef' and the Thr to 'base t^1 t-st'. However, I found that the signals on the images is very different from those when I used 3dfim+. I would like to ask whether there's something wrong in the above procedures in doing the t-test?
Thank you.
Phoebe.