Questions about the 1D file for 3dLSS

AFNI Message Board
1

Hi,

When I use 3dLSS, in a single run, I have three different stimulus conditions: left-eye condition, right-eye condition, and binocular condition. Each condition consists of eight blocks, each block lasts for 8 seconds, accompanied by a blank screen for 12 seconds. When I generate a 1D file with 3dDeconvolve, should I put these three stimulus conditions in one 1D file? If so, should I arrange them according to method (1) or method (2)? And is the method in (3) acceptable?

(1)

3dDeconvolve -nodata 240 2.0 -polort 1 -num_stimts 3 \
    -stim_times 1 RE.txt 'BLOCK(8,1)' \
    -stim_times 2 BE.txt 'BLOCK(8,1)' \
    -stim_times_IM 3 **LE**.txt 'BLOCK(8,1)' \
    -x1D Xmat_LSS_LE.1D -x1D_stop -overwrite

(2)

3dDeconvolve -nodata 240 2.0 -polort 1 -num_stimts 1 \
    -stim_times_IM 1 **LE_RE_BE**.txt 'BLOCK(8,1)' \
    -x1D Xmat_LSS_LEREBE.1D -x1D_stop -overwrite

(3)

3dDeconvolve -nodata 240 2.0 -polort 1 -num_stimts 1 \
    -stim_times_IM 1 **LE**.txt 'BLOCK(8,1)' \
    -x1D Xmat_LSS_LE.1D -x1D_stop -overwrite
3dDeconvolve -nodata 240 2.0 -polort 1 -num_stimts 1 \
    -stim_times_IM 1 **RE**.txt 'BLOCK(8,1)' \
    -x1D Xmat_LSS_RE.1D -x1D_stop -overwrite
3dDeconvolve -nodata 240 2.0 -polort 1 -num_stimts 1 \
    -stim_times_IM 1 **BE**.txt 'BLOCK(8,1)' \
    -x1D Xmat_LSS_BE.1D -x1D_stop -overwrite
2

(Modified per Rick's suggestion)

From the modeling perspective, the first two methods would yield the same result. However, there is some practical difference: method (2) is more concise with 3dDeconvolve, but method (1) is better for bookkeeping of result labeling with 3dLSS.

With method (3), you would run three separate models in 3dLSS, which is not preferable.