probably something wrong with your stimulus schedule (paradigm). Can you send the X.mat (or Xtmp.mat)?

On 2/28/2023 12:12 PM, Caspar M. Schwiedrzik wrote:

        External Email - Use Caution        

Hi!
I am running into an issue with fast_selxavg3b.
I am attempting a first level PPI analysis.

The command is
mkanalysis-sess -fsd bold -native -delay -2/2 -notask -TR 2 -polyfit 6 -analysis [xxx] -funcstem [xxx] -runlistfile mooneyppi.list -nuisreg motion.dat 6 -tpef tpef_dvars.dat -acfbins 5 -acffwhm 0 -mask [xxx] -force -nuisreg juice.dat 1 -nuisreg fixation.dat 1 -taskreg ppi.dat 4

I get the following error:

GLS Residual Pass
  run 1    t= 0.0
Error using chol
Matrix must be positive definite.

Error in fast_selxavg3b (line 942)
Wseg = inv(chol(toeplitz(acfrun))');

Before that, I am getting several warnings of this kind: 
> In fast_selxavg3b (line 895)
Warning: Matrix is singular, close to singular or badly scaled. Results may be
inaccurate.

A few of them, but not run 1, have RCOND = NaN.

This is with versions 7.2 and 5.3.
I found one previous conversation on the mailing list regarding this error in the context of a resting state analysis, but no conclusion. Any advice would be appreciated.
Thanks!
Caspar

_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer