Persistent State

• Using a LIF model, it's easy to build a state that is persistent.
• Take several neurons as the state. I used 5, but now use 10.
• Give them large internal connections.
• Activate them all simultaneously.
• They'll then reactivate themselves.
• That's what I call a binary Cell Assembly.
• With 10, I make 2 inhibitory. When the CA starts to fire too fast, the inhibitory neurons fire, and slow it down.
• If you use a fatiguing model, you can have two sets of neurons that activate each other.
• This is not the way the brain does it, but it works for programming.