Fatigue
- Turning Fatigue on ruins the experiment
- Fatigue rates are set so that neural fatigue recovers
twice as slowly as it grows
- Consequently, during initial presentation neurons
come on via external stimulation
- After a few steps some neurons go off
- They come back on, but in general two thirds of the neurons in the
CA are off during presentation
- This is bad but the real problem is when presentation ceases
- Here fatigue has built up so that none really persist for more
than 2 steps
- The question can be asked why use neural fatigue?
- The brain may just have picked up this evolutionary detail.
- However, fatigue should enable CAs to shut down.
- Further more, it may be involved in learning (though we do not
currently include it in our learning mechanism).
- Finally, it can reduce the spread of activation and
thus reduce (simulated) epilepsy.
