Hello,

I'm trying to run selxavg3-sess with FS v4.5 for an FIR analysis where each trial is modeled as a separate condition. The command works smoothly for the majority of subjects, but for 2 different subjects, it exits with the following error:

Saving X matrix to /home/sam/fMRIDATA/samstudies/SoCo/SoCo10/bold/Logit/Xtmp.mat

XCond = 12220.6 (normalized)

ERROR: design is ill-conditioned

>> ------------------------------------------

ERROR: fast_selxavg3() failed\n

details:

I do not definite any contrasts, as I'm only running this command to generate trial specific betas for a separate analysis. Each trial lasts 18 seconds, followed by 4 seconds of fixation. Here is what part of a paradigm file looks like (the paradigm files are identical for all subjects):

4 1 18 1

22 0 4 1

26 2 18 1

44 0 4 1

48 3 18 1

66 0 4 1

70 4 18 1

88 0 4 1

92 5 18 1

110 0 4 1

114 6 18 1

132 0 4 1

136 7 18 1

154 0 4 1

here are the analysis.cfg and .info files:

analysis analysis_name

TR 2

fsd bold

funcstem fmcsm4

tpexclude

parname FIR_logit.par

designtype event-related

nconditions 96

-timewindow 20

-prestim 2

-acfbins 10

-fix-acf

-noautostimdur

-no-fsv3-st2fir

-no-fsv3-whiten

-polyfit 2

-TER 2

-nskip 0

-fwhm 0

-extreg mcextreg

-nextreg 3

-rescale 1000

the command which throws the error:

selxavg3-sess -analysis analysis_name -s subject_name

The strange part for me is that this only seems to be a problem with a couple subjects, despite the fact all the paradigm files are identical. After looking through the mailing archive, I tried playing around with the prestim window, but even it if is = 0, there is 1 subject that exits with an error when I run selxavg3-sess. Do you know what might be going on?

Any guidance would be appreciated.

Thanks,

Sam