Hi AFNI experts,

I'm trying to run an analysis using TENT functions which are amplitude-modulated by reaction time. I keep getting collinearity errors between TENT regressor pairs, though. The correlation matrix warnings, 3dDeconvolve warnings, and 3dDeconvolve analysis script are below. I appreciate any advice!

Thanks,

Cindy


severity correlation cosine regressor pair
-------- ----------- ------ ----------------------------------------
high: 0.949 0.949 (58 vs. 59) all_gencue#8 vs. all_gencue#9
high: 0.949 0.949 (75 vs. 76) uncueincon#8 vs. uncueincon#9
high: 0.949 0.949 (41 vs. 42) cue_incon#8 vs. cue_incon#9
high: 0.948 0.949 (33 vs. 34) cue_incon#0 vs. cue_incon#1
high: 0.948 0.949 (50 vs. 51) all_gencue#0 vs. all_gencue#1
high: 0.947 0.949 (67 vs. 68) uncueincon#0 vs. uncueincon#1



*+ WARNING: !! in Signal+Baseline matrix:
* Largest singular value=2.13182
* 6 singular values are less than cutoff=2.13182e-07
* Implies strong collinearity in the matrix columns!
*+ WARNING: !! in Signal-only matrix:
* Largest singular value=1.73727
* 6 singular values are less than cutoff=1.73727e-07
* Implies strong collinearity in the matrix columns!
*+ WARNING: +++++ !! Matrix inverse average error = 0.00235849 ** BEWARE **



3dDeconvolve -input pb04.$subj.r*.scale+tlrc.HEAD \
-polort 3 -local_times \
-num_stimts 12 \
-stim_times 1 ${subj}_all_collapsed.1D 'BLOCK(3.5,1)' -stim_label 1 all_collapsed \
-stim_times_AM2 2 ${subj}_all_cueincon_marry.1D 'TENT(0,14,8)' -stim_label 2 cue_incon \
-stim_times 3 ${subj}_all_error.1D 'BLOCK(3.5,1)' -stim_label 3 all_error \
-stim_times_AM2 4 ${subj}_all_gencue_marry.1D 'TENT(0,14,8)' -stim_label 4 all_gencue \
-stim_times 5 ${subj}_all_lies.1D 'BLOCK(3.5,1)' -stim_label 5 lies \
-stim_times_AM2 6 ${subj}_all_uncueincon_marry.1D 'TENT(0,14,8)' -stim_label 6 uncueincon \
-stim_file 7 motion_demean.1D'[0]' -stim_base 7 -stim_label 7 roll \
-stim_file 8 motion_demean.1D'[1]' -stim_base 8 -stim_label 8 pitch \
-stim_file 9 motion_demean.1D'[2]' -stim_base 9 -stim_label 9 yaw \
-stim_file 10 motion_demean.1D'[3]' -stim_base 10 -stim_label 10 dS \
-stim_file 11 motion_demean.1D'[4]' -stim_base 11 -stim_label 11 dL \
-stim_file 12 motion_demean.1D'[5]' -stim_base 12 -stim_label 12 dP \
-fout -tout -x1D X.xmatTENT.1D -xjpeg XTENT.jpg \
-fitts fittsTENT.$subj \
-errts errtsTENT.${subj} \
-bucket statsTENT.$subj