Hi everyone,

I’m conducting a meta-analysis on the effects of transcranial stimulation on functional connectivity using fMRI data.
My main covariate of interest is the stimulation site, which is typically reported as categorical labels corresponding to the 10–20 EEG system (e.g., F3, C4, P3, etc.).
When I try to include stimulation site as a covariate in the meta-regression, I get the following error:

gsl: ../../gsl-2.6/linalg/lu.c:449: ERROR: matrix is singular
Default GSL error handler invoked.

How should I correctly handle this categorical covariate in SDM-PSI? Should I manually create dummy variables for each site, or is there a recommended way to encode and include categorical predictors in the software?

Any advice is greatly appreciated.
Merci,
Shiva