Deep Learning to Analyze Chaotic Dynamics in Computational Neuroscience

Authors: Carmen Mayora-Cebollero (Universidad de Zaragoza), Roberto Barrio (Universidad de Zaragoza), Álvaro Lozano (Universidad de Zaragoza), Ana Mayora-Cebollero (Universidad de Zaragoza)

Presentation type: Poster at SNUFA 2026 online workshop (4-5 Nov 2026)

Abstract

The detection of different dynamical regimes in computational neuroscience models can provide valuable insights into the behavior of neural populations. Although such dynamical analyses are usually performed using classical techniques, in recent years, Deep Learning has emerged as a promising alternative. In this presentation, we use Deep Learning to detect chaotic dynamics and to approximate the full Lyapunov exponents spectrum of a mean-field model of neural populations. Our results show that Deep Learning can accurately characterize the system dynamics while offering several advantages, such as a substantial reduction of computational time (in chaos detection analyses, Deep Learning requires less than 0.5% of the computational time of a classical algorithm) and the ability to work with partial information (only single-variable time series are needed to accurately approximate the full Lyapunov exponents spectrum of the system).

Are you the author and do you want to add some related links to this page? If so, check out our guide to adding related links and videos.