Checking linearity assumption in a mixed linear model ?

I'm trying to fit a mixed linear model to explain a variable with four fixed regressors and three random factors.
I'm not sure that the relationship between my dependant variable and each of my regressor is linear. According to some literature, it might be, but it might not be also. Maybe I should be fitting a GAMM or something but before building something too complex i would like to fit a LMM and check the linearity relationships between my variables.

So, I'd like to do a sort of partial regression for each regressor in order to look at these relationships. So I was wondering : is it okay if I just make a new model by taking off a regressor, and then check graphically the relationship between that "excluded" regressor and the fitted values of that new model ? And so, I'd do this for each regressor ? Or maybe there are some parameters I'm not taking into account ?

If there's a simpler/more robust way, do you know how to do it in R ? I've been searching a little. I know the existence of "variable added plots" for linear models. But I couldn't find an equivalent for mixed models.

Thank you a lot
Last edited: