⚠ Caution: External sender

My guess is that the variable (G_cingul vol) should be normalized (remove the mean and divide by the stddev).

On 9/6/2022 5:44 PM, Piot, Alex wrote:

External Email - Use Caution

Hello Freesurfer experts,

I am trying to perform a GLM, using the volume of a brain region as a covariate. I am doing a two-group, one covariate design. I originally tried to use QDEC to facilitate the process and received an error message that the mri_glmfit command had failed. I then saw that QDEC was deprecated and followed the instructions set out on the freesurfer wiki for doing it manually (MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysisV6.0).

I received the same error message as I had with QDEC on the command line:

ERROR: MatrixResidualForming(): X is not invertable

RD is not invertable n = 1

This only occurs for certain of the brain regions used as covariates; for others, the QDEC and command-line streams work perfectly. Additionally, I've noticed that if I input different numbers for the covariate values in the fsgd file, the command runs smoothly.

I haven't been able to find a similar error on the mailing list, does anyone have an idea of how to work around this? I would greatly appreciate it.

Please find attached my fsgd file, and the full log of what happens after I run the mri_glmfit command (when it works and when it doesn't)

The hands.fsgd file doesn't work, whereas the hands_works.fsgd file does.

Freesurfer version: 7.2.0, Ubuntu 20.04, Windows WSL

Best wishes,

Alex

--

Alexandre Piot, MSc.

Queen Square Institute of Neurology

President of the Application of Psychedelics Society -- UCL

UNITy Project
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer