Some mediation models are hard to figure out, in terms of what they say, or how one would translate their findings in practical terms, one example is this one, which is hard for me to translate into words, and then see what is implying:
Blood, E., & Cheng, D. (2012). Non-linear mixed models in the analysis of mediated longitudinal data with binary outcomes. BMC Medical Research Methodology, 12(1), 5.
Others are more intuitive to some, like: