Panel Data analysis for Categorical response variable

I need to estimate a fixed effects panel data model with categorical ordered response variable.

If instead of a panel my data was one-dimensional (one subject over time, or many subjects at one point in time), I would be able to use 'mnrfit' with "'model','ordinal','interactions','off'" options.

If my response variable was not categorical, but continuous, I would be able to use 'nlmefit' by setting up "Mixed Effects" model with fixed effects, but without random effects.

However, I have both - panel AND categorical response. How can I handle it in Matlab?

Thanks!