What does one neural model tell us about another?

I'd love it if a lot of researchers were using fLIF neurons,
but there are a lot of different neural models.
Compartmental models, LIF models, Hopfield nets, Boltzman machines, and many more.
There are a lot of non-neural connectionist systems (MLPs, ART maps, SOMs, RBFs, neural gas, etc.).
Even within LIF models there is variance between time grain, neural types, learning rules (if any), and of course
topology. This includes brain areas, but also laminar architecture, and even local connectivity.
When one researcher implements a model, it is not particularly clear how it informs other models, and it's even less clear about non-neural connectionist systems.