Hello,
I am encountering an error in FSL 5.0.10 that I can't seem to resolve. When I try to run film_cifti, I get an error during film_gls as it tries to compute parameter estimates for the surfaces. The volume runs through just fine, but the surface gifti gives a "process fails to converge" error.
I have only seen one post related to this issue in the past, but there was not enough information posted to help me resolve the issue on my own. The post and its update were from March 21-22, 2017, entitled "film_gls surface mode error "Newmat: incompatible dimensions."" The hint to solving the issue is that "the command line works fine when the values in the medial wall are non-zero and the time series are not constant."
I found that the values in my cifti are non-zero, but when film_cifti separates the cifti into giftis (surfaces and volume), each surface gifti suddenly has "zero" values. I do not know why this is happening or whether it is the main cause of my "process fails to converge" error.
Here is an example of my film_cifti command:
film_cifti -i REST1_Atlas.dtseries.nii -o BOLD_299-07_124-1 -l 30299.L.midthickness.32k_fs_LR.surf.gii -r 30299.R.midthickness.32k_fs_LR.surf.gii --filmOptions "--con=BOLD_299-07_124-1.con --pd=BOLD_299-07_124-1.mat"
I also tried calling film_gls independently:
/mnt/max/software/fsl-5.0.10/bin/film_gls --mode=surface --in=outLeft.func.gii --rn=statsLeft --epith=0 --ms=4 --con=BOLD_299-07_124-1.con --pd=BOLD_299-07_124-1.mat --in2=30299.L.midthickness.32k_fs_LR.surf.gii
In both cases, I get the following error:
film_cifti error:
paradigm.getDesignMatrix().Nrows()=226\nparadigm.getDesignMatrix().Ncols()=2\nsizeTS=226\nnumTS=32492\nCalculating residuals...\nCompleted\nEstimating residual autocorrelation...\nCalculating raw AutoCorrs... Completed\nTukey M = 15\nTukey estimates... Completed\nCompleted\nPrewhitening and Computing PEs...\nPercentage done:\n\n\n
An exception has been thrown\nRuntime error:- detected by Newmat: process fails to converge\n\n
MatrixType = Rect # Rows = 2; # Cols = 2\nTrace: SVD; pinv.\n\n'
film_gls error:
Prewhitening and Computing PEs...
Percentage done:
An exception has been thrown
Runtime error:- detected by Newmat: process fails to converge
MatrixType = Rect # Rows = 2; # Cols = 2
Trace: SVD; pinv.
The design matrix has 226 rows (time points) and 2 columns (predictors). It was created using the "GLM Setup" function from the drop-down menu of the main FSL GUI.
Please let me know if you have any ideas as to why this program is throwing an error. I can provide files if necessary.
Thank you,
Sam
|