A neural network model based on the cortical modularity

Although the individual cortical areas in the brain support very diverse functions, their anatomical organizations are very similar, given by collections of cortical blocks with nearly the same structure. This character suggests that statistical dynamics in the brain may be modeled much efficiently...

Full description

Saved in:
Bibliographic Details
Published in:Journal of the Korean Physical Society 2021, 79(8), , pp.772-784
Main Authors: Cho, Myoung Won, Choi, M. Y.
Format: Article
Language:English
Subjects:
Brain
Free energy
Learning
Mathematical and Computational Physics
Modularity
Neural networks
Original Paper - Cross-Disciplinary Physics and Related Areas of Science and Technology
Particle and Nuclear Physics
Physics
Physics and Astronomy
Theoretical
Visual observation
물리학
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Although the individual cortical areas in the brain support very diverse functions, their anatomical organizations are very similar, given by collections of cortical blocks with nearly the same structure. This character suggests that statistical dynamics in the brain may be modeled much efficiently if interactions between the cortical blocks and learning in them are described through effective reduction. This paper introduces a neural network model based on the cortical modularity. Specifically, in building the model, we postulate on how cortical blocks function and interact with each other and adopt the formalism of path-integral-based interactions and free-energy-based learning. We apply the model to explain the characteristics observed in early visual systems.
ISSN:0374-4884
1976-8524
DOI:10.1007/s40042-021-00301-0