# Lagrange multiplier test in Mixed Level Model

I want to estimate a mixed level model with AR(1) errors and then conduct a Lagrange Multiplier test. The mixed model allows for rich covariance structures but it does not allow for AR(1) errors. Can someone help?