3dDeconvolve Error

Hi,
I was running preprocessing on set of resting-state data and run into issue with 3dDeconvolve on one of the subjects while it seems to be fine for others.
Below is the command and the error, do you have any suggestions what could be causing the problem?

Best regards and thanks,
Kasia

3dDeconvolve -input pb04.fmri_name.r01.blur+orig.HEAD -censor censor_fmri_name_combined_2.1D -ortvec bandpass_rall.1D bandpass -ortvec motion_demean.1D mot_demean -ortvec motion_deriv.1D mot_deriv -polort 3 -num_stimts 0 -fout -tout -x1D X.xmat.1D -xjpeg X.jpg -x1D_uncensored X.nocensor.xmat.1D -fitts fitts.fmri_name -errts errts.fmri_name -x1D_stop -bucket stats.fmri_name

++ 3dDeconvolve extending num_stimts from 0 to 110 due to -ortvec
++ 3dDeconvolve: AFNI version=AFNI_19.0.01 (Jan 22 2019) [64-bit]
++ Authored by: B. Douglas Ward, et al.
++ STAT automask has 428786 voxels (out of 3655125 = 11.7%)
++ Skipping check for initial transients
++ Input polort=3; Longest run=314.0 s; Recommended minimum polort=3 ++ OK ++
++ Number of time points: 157 (before censor) ; 25 (after)

  • Number of parameters: 114 [114 baseline ; 0 signal]
    ** ERROR: *** Censoring has made regression impossible :frowning: ***
    ** FATAL ERROR: 3dDeconvolve dies: Insufficient data (25) for estimating 114 parameters
    ** Program compile date = Jan 22 2019

Hi, Kasia-

Too many time points in your dset have been censored to do your modeling afterward.

This line tells you that you start with 157 time points, and end up with only 25 after:


++ Number of time points: 157 (before censor) ; 25 (after)

That is a quite large loss-- is this subject very motiony?

This line then tells you how many parameters you are trying to use (I guess this is resting state, because of the 0 signal-- and perhaps you are bandpassing the data, too, and hence there are so many baseline params?):


+ Number of parameters: 114 [114 baseline ; 0 signal]

And these lines tell you that you have too few time points (25) to fit a model with so many parameters (114):


** ERROR: *** Censoring has made regression impossible :( ***
** FATAL ERROR: 3dDeconvolve dies: Insufficient data (25) for estimating 114 parameters

You have to have more time points than parameters, as a requirement for modeling the data.

–pt