Might Improve with Changed Fatigue Model

One problem with the model is that after a period of no input, the first spikes come more rapidly than predicted (see Clamped Fig)
As it stands, the model sets the minimum retained fatigue at 0.
There is no reason this could not be negative so that it spikes faster after rest.
Of course, it would then have to recover back to 0 more quickly.
This could also be a perhaps better basis for spontaneous activation.