Hi Everyone,

I realize a lot has been discussed on this issue on the message board but I'm a little unclear on how to create a peri-stimulus time series based on deconvolution analysis. In other words, I want to measure the signal change over the course of say 18 seconds for each event of interest averaged across all trials of that event type. So, if I have happy, fearful, and neutral faces presented in a fast event related design, I want to plot the average waveform over 18s (assuming a TR of 2s). I want to do this initially on a single subject basis. Is the following correct?

3dDeconvolve -input all_Runs+orig \
-polort 3 -num_stimts 9 \
-stim_file 1 fear.txt \
-stim_label 1 fear \
-stim_file 2 happy.txt \
-stim_label 2 happy \
-stim_file 3 neutral.txt \
-stim_label 3 neutral \
-stim_minlag 1 0 -stim_maxlag 1 9 \
-stim_minlag 2 0 -stim_maxlag 2 9 \
-stim_minlag 3 0 -stim_maxlag 3 9 \
-stim_file 4 'moveParam_all.txt[1]' -stim_base 4 -stim_label 4 roll \
-stim_file 5 'moveParam_all.txt[2]' -stim_base 5 -stim_label 5 pitch \
-stim_file 6 'moveParam_all.txt[3]' -stim_base 6 -stim_label 6 yaw \
-stim_file 7 'moveParam_all.txt[4]' -stim_base 7 -stim_label 7 dS \
-stim_file 8 'moveParam_all.txt[5]' -stim_base 8 -stim_label 8 dL \
-stim_file 9 'moveParam_all.txt[6]' -stim_base 9 -stim_label 9 dP \
-iresp 1 fear.irf -iresp 2 happy.irf \
-iresp 3 neutral.irf \
-sresp 1 fear_std -sresp 2 happy_std -sresp 3 neutral_std \
-fitts full_model.fit -errts residual_error.fit \
-fout -tout \
-bucket decon_results