Hi,
I am using 3dDeconvolve to analyze a 1-run ER design with TR=1.2, 80 trials and tot run duration of 6 min and 40 sec. I have some questions on how to best set up the TENT response model parameters given the time aspects of my design:


Each trial lasts 1.65 sec (which includes stimulus presentation and response time), and is followed by a jittered inter-trial interval ranging from 2 to 4.7 sec (avg jittering time 3.35 sec). Does it make sense to have ‘TENT (0,5,3), that is, modeling the response occurring from 0 to 5 sec and having 3, (or maybe 4?), tent functions?

And these tent functions, in which, if I understand correctly, I input the stimulus starting times in sec through the run, are already taking into account the usual 4-6 sec shift for reaching the peak bold activation?

In addition, I'd like to have four regressors, since I have two conditions (ex A and B), but I also want to differentiate between correct and incorrect responses in each of them (ex. correctA, correctB, incorrectA, incorrectB, hence I picked numstimts of 4).

I’d appreciate it if you could also take a look at this script, and tell me if something is wrong or missing?

3dDeconvolve \
-input 16588+orig \
-TR_1D 1.2 \
-automask \
-numstimts 4 \
-stim_times 1 regressor_CorrectStop.1D ‘TENT (0,5,3)’ \
-stim_label 1 ‘CorrectStop’ \
-stim_times 2 regressor_CorrectGo.1D ‘TENT (0,5,3)’ \
-stim_label 2 ‘CorrectGo’ \
-stim_times 3 regressor_IncorrectStop.1D ‘TENT (0,5,3)’ \
-stim_label 3 ‘IncorrectStop’ \
-stim_times 4 regressor_IncorrectGo.1D ‘TENT (0,5,3)’ \
-stim_label 4 ‘IncorrectGo’ \
-polort 1 \
-xout \
-fout \
-tout \
-fitts output_model_fitts \
-xjpeg Xmat_jpg.jpg \
-x1D Xmat_1D.x1D \
-bucket 16588.buck \


Thank you very much,
Alessandra