Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Parameter space structure of continuous-time recurrent neural networks.

Randall D Beer1

  • 1Department of Electrical Engineering and Computer Science, Case Western Reserve University, Cleveland, OH 44106, USA. rdbeer@indiana.edu

Neural Computation
|October 21, 2006
PubMed
Summary

This study maps the parameter space of continuous-time recurrent neural networks (CTRNNs). It reveals distinct dynamic regions, enabling probability estimates for different circuit behaviors.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

When can neuronal activity-dependent homeostatic plasticity maintain circuit-level properties?

Journal of computational neuroscience·2026
Same author

Matters of life and death in computational cell biology.

NPJ systems biology and applications·2026
Same author

Social contingency in embodied neural networks relies on co-constructed dynamical mechanisms.

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2026
Same author

The role of signal in saltatory pursuit of cryptic stationary targets.

Movement ecology·2025
Same author

MATTERS OF LIFE AND DEATH IN COMPUTATIONAL CELL BIOLOGY.

ArXiv·2025
Same author

(A)Life as It Could Be.

Artificial life·2024

Area of Science:

  • Computational Neuroscience
  • Dynamical Systems Theory

Background:

  • Characterizing the structure of neural circuit parameter space is a fundamental challenge.
  • Continuous-time recurrent neural networks (CTRNNs) are simple yet dynamically universal models.

Purpose of the Study:

  • To systematically study the global parameter space structure of CTRNNs.
  • To understand how parameter variations influence the dynamics of CTRNNs.

Main Methods:

  • Explicit computation of local bifurcation manifolds for CTRNNs.
  • Visualization of manifold structures in net input space for small circuits.
  • Characterization of the combinatorics and geometry of parameter space regions for arbitrary circuit sizes.

Main Results:

Related Experiment Videos

  • Identified extremal saddle node bifurcation manifolds that partition CTRNN parameter space.
  • These manifolds delineate regions with distinct effective dimensionalities of dynamics.
  • Developed an asymptotically exact approximation for these regions.

Conclusions:

  • The identified regions provide a framework for understanding CTRNN dynamics.
  • Enables estimation of the probability of observing different dynamic behaviors within CTRNN parameter space.
  • Offers a foundational step towards a general theory of neural circuit structure.