Let's assume you would like to write a plasticity model in which induced changes to a synapse require some time to percolate through. Consider that inserting for instance additional AMPA receptors into a postsynaptic density takes time…

Update As of Auryn 0.5 synapse models with multiple states can be implemented more easily using synaptic state vectors – see for sim_bg_lowpass in the example directory to get an idea.

I still owe a tutorial for that here. The old complex synapse tutorial can be found here which describes the procedure before v0.7.