[Abstract]

Activity Transfer Models for Associative Activities in a Brain

Shinji Karasawa (Miyagi National College of Technology, Japan)



This paper describes the paradigm of activity and a network of the programmable and bidirectional digital signal translator that is useful to emulate the brain mechanism. The surroundings ignite activity in a nerve network and the transference accompanies together with forming of circuits. An activity is able to keep in a loop. A neuron laid in upper layer is able to integrate the constituent activities laid in lower layer. The synthetic judgment is carried out by means of pattern matching of plural subsets of activities.
Although each activity is non-linear and it is always on the move, transference of electric charge is able to represent the transference of activity. Since an activity corresponds to a data and a circuit corresponds to a program,the circuit and activity exist concurrently. The concurrently existing subsets of electrons are able to form a MOS decoder. Since transmission of impulse is intermittently, the connecting points of a decoder can be used as the connecting points of the encoder by means of exchanging between inputs and the output.