Effect of spatial facilitation on neural network excitability

Miroslav Musila, Mitsuyuki Nakao, Alan Fine, Mitsuaki Yamamoto

Research output: Contribution to journalArticlepeer-review

Abstract

The time evolution of an asymmetrical form of the Hopfield neural network was simulated. The noise assigned to each neuron to represent spontaneous fluctuations of membrane potential had a translated gamma distribution. As a result, in comparison with the generally accepted Gaussian noise with the same mean and standard deviation, for each neuron the instantaneous sum of the noise and the input signals from other neurons could exceed a large threshold more frequently. This imitated the effect of long-term spatial facilitation, elevating the neurons' initial depolarization so that even weak inputs could cause firing. Such gamma-distributed noise increased the overall excitability of the neural network as a modeled subsystem of the central nervous system. As the mean of the translated gamma distribution increased, neuronal activity tended rapidly to the pacemaker bursting type.

Original languageEnglish
Pages (from-to)367-383
Number of pages17
JournalCybernetics and Systems
Volume27
Issue number4
DOIs
Publication statusPublished - 1996 Aug

ASJC Scopus subject areas

  • Software
  • Information Systems
  • Artificial Intelligence

Fingerprint Dive into the research topics of 'Effect of spatial facilitation on neural network excitability'. Together they form a unique fingerprint.

Cite this