Just wanted to follow up on this a bit. If I am interpreting the results as stated before, aren't I losing the ability to look at a main effect of time?
Is there possibly another solution which would allow me to specify AR1 or another autoregressive structure?