9 Apr
2010
9 Apr
'10
9:15 a.m.
Hi,
I get this error message when running glm_fit (v.3.02):
CEPHES ERROR: fdtrc domain error (I see you've handled this error message before, but there was no solution there: http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg11602.html)
I'm using a model with two groups, and age and ICV as continuous regressors. The error message comes when estimating the 0 0 0 1 contrast, i.e. ICV. Matrix condition = 494.18.
The beta for this covariate shows very small numbers, i.e. 0-6*10^-1. And the sig.mgh file for it shows all vertices = 1e+10.
yours,
Dr. Lars M. Rimol