Compound Net Model of the Cerebral Cortex
Report Number: AMRL TR 68-189
Author(s): Harth, E. M., Csermely, T. J., Lindsay, R. D., Beek, B.
Corporate Author(s): Syracuse University
Laboratory: Aerospace Medical Research Laboratories
Date of Publication: 1969-06
Pages: 72
Contract: F33615-67-C-1413
DoD Project: 7232
DoD Task: 723205
Identifier: AD0692719
Abstract:
The central nervous system is believed to consist of a network of neurons, combining in its structure genetically determined design with probabilistic features. A model is described in which probabilistic nets are used as building blocks from which systems can be assembled, approximating for example the interaction between sensory projection areas in the cortex with cortical association areas. The operation of such compound nets is based on the dynamics of single probabilistic nets and on a set of assumptions concerning the coding of sensory information into the language of central neuronal activity. Data describing the dynamics of simple nets are presented. It is found that various types of information processing can be explained by applying the hypothesis of synaptic facilitation to compound nets. Sustained neural activity is investigated, particularly the problem of ergodicity. Highly non-ergodic types of reverberations were found both theoretically and by computer simulation for certain types of nets.
Provenance: RAF Centre of Aviation Medicine
Author(s): Harth, E. M., Csermely, T. J., Lindsay, R. D., Beek, B.
Corporate Author(s): Syracuse University
Laboratory: Aerospace Medical Research Laboratories
Date of Publication: 1969-06
Pages: 72
Contract: F33615-67-C-1413
DoD Project: 7232
DoD Task: 723205
Identifier: AD0692719
Abstract:
The central nervous system is believed to consist of a network of neurons, combining in its structure genetically determined design with probabilistic features. A model is described in which probabilistic nets are used as building blocks from which systems can be assembled, approximating for example the interaction between sensory projection areas in the cortex with cortical association areas. The operation of such compound nets is based on the dynamics of single probabilistic nets and on a set of assumptions concerning the coding of sensory information into the language of central neuronal activity. Data describing the dynamics of simple nets are presented. It is found that various types of information processing can be explained by applying the hypothesis of synaptic facilitation to compound nets. Sustained neural activity is investigated, particularly the problem of ergodicity. Highly non-ergodic types of reverberations were found both theoretically and by computer simulation for certain types of nets.
Provenance: RAF Centre of Aviation Medicine