A neural maze is like a neural network.

A neural network:
- has a fixed grid like wiring that is created at instantiation.
- calculates every node in the network on every iteration.

A neural maze:
- has no wiring or nodes when created.
- nodes and wiring is added when data is learned which is recorded in the form of the established routings.
- is called a maze because evaluations do not calculate all nodes during every iteration but uses a search algorithm that activates only the nodes that are related to the currently evaluated situation.

A Pulse Source is used in neural mazes that have nodes that route charge but do not generate charge themselves.

Meaning that all signals that travel through the NM are of external origin and not generated by firing nodes.

A Digitron is a far descendent of the 1960 Perceptron which was used in early NN.

Unlike the Perceptron, the Digitron has native support for both spatial and temporal dimensional patterns.

This facilitates building NN and NM that understand order of events and logistic events (meaning the movement of items which is not a static event) and by that can learn both causal and associative relations between such events.

PSNM is based on its bioware equivalent (being the hippocampus and both neocortex and cerebellum).

Meaning it records autonomously life experiences in engrams generating an internal world model.

One should not think of this as artificial intelligence as the constructed agent is not unlike human intelligence and even (by not having a bioware support system) needs some virtual or material construct for embodiment.