3dDeconvolve
-input {$subj}_all_runs+orig -num_stimts 4
\
-stim_file 1
../misc_files/all_stims.1D'[0]' -stim_label 1 ToolMov \
-stim_minlag 1 0 -stim_maxlag 1 14 -stim_nptr 1 2
\
-stim_file 2
../misc_files/all_stims.1D'[1]' -stim_label 2 HumanMov\
-stim_minlag 2 0 -stim_maxlag 2 14 -stim_nptr 2 2
\
-stim_file 3
../misc_files/all_stims.1D'[2]' -stim_label 3 ToolPnt \
-stim_minlag 3 0 -stim_maxlag 3 14 -stim_nptr 3 2
\
-stim_file 4
../misc_files/all_stims.1D'[3]' -stim_label 4 HumanPnt\
-stim_minlag 4 0 -stim_maxlag 4 14 -stim_nptr 4 2
\
¥ Input dataset is the catenated thing
created earlier
¥ There are 4 time series models
¥ All
stimuli time series are in one file with 4 columns: ../misc_files/all_stims.1D
¥ The selectors like Ô[2]Õ pick out a particular column
¥ Each stimulus and HRF will be sampled at
TR/2 = 1.0 s, due to the use of -stim_nptr k 2 for each k
¥ Lag from 0 to 14 is about right for
hemodynamic response to a brief stimulus