The Simulation of Neural Elements by Electrical Networks Based on Multi-Aperture Magnetic Cores

Multilevel storage, gating, controlling threshold-level, and summation are basic functional operations common to several systems for the simulation of neural behavior by electrical networks. It is shown how these may be realized very efficiently in terms of multi-apertured magnetic cores; in a typical example, only two cores of the type described by Crane are required for a circuit with threshold control of three simultaneous gates and non-destructive analog read-out. A logic element equivalent to a self-holding relay exhibits dc current gain and has compatible input-output connections.