A neural network is a connection of neurons. Networks send neural Brain Events to other networks. Self-reinforcing networks also send events to themselves. These events keep the network activated (neurons keep firing).
content
A self-reinforcing network has the ability to store some chunk of changeable information. Your Brain Runs Erlangยป
The mechanism for storing that information is reminiscent of (not identical to) how dynamic memory in computers works (dating back to mercury delay lines 
): the state is represented by electrical charges that will decay back to quiescence if left alone. The self-signaling can act as a memory refresh .
Typically, an activated network will nevertheless gradually become quiescent if it's not often enough stimulated by events from other networks. See a Quick Example of Learning a Smell.
You can think of the self-reinforcing events as being driven by *local clock*, so: `event`, `event`, `event`, `...` rather than a continuous process. There's no global clock to synchronize all networks, though some networks do act synchronously (as with a set of networks receiving periodic pulses from a "timekeeper" network).