Direct and retrograde signal propagation in unidirectionally coupled Wilson-Cowan oscillators

Guy Elisha, Richard Gast, Sourav Halder, Sara A. Solla, Peter J. Kahrilas, John E. Pandolfino, Neelesh A. PatankararXiv 2024 · DOI · arXiv

Abstract

Certain biological systems exhibit both direct and retrograde propagating wave signals, despite unidirectional neural coupling. However, there is no model to explain this. Therefore, the underlying physics of reversing the signal's direction for one-way coupling remains unclear. Here, we resolve this issue using a Wilson-Cowan oscillators network. By analyzing the limit cycle period of various coupling configurations, we determine that intrinsic frequency differences among oscillators control wave directionality.

Introduction: The Puzzle of Reversing One-Way Signals

In biological systems such as the esophagus or uterus, coordinated waves of muscle contraction are vital for physiological function. While these waves typically move in a single direction (direct propagation), certain pathological or physiological states can cause them to reverse (retrograde propagation).

The mechanical puzzle lies in the neural architecture, which is often "unidirectionally coupled"—the physical wiring only permits signals to flow from one point to the next. Previous models explained wave reversal by assuming bidirectional wiring or shifting the stimulus location. This study proposes a more biologically plausible mechanism: how a fixed, one-way network can flip its signal direction without structural rewiring or changing the initiation point.

Methods and Network Architecture

The researchers utilized a network of Wilson-Cowan (WC) oscillators, a standard model for interactions between excitatory ($E$) and inhibitory ($I$) neuronal populations.

The study compared two primary configurations:

  1. Chain Configuration: A linear sequence of 70 oscillators where each influences only its immediate neighbor.
  2. Ring Configuration: A benchmark where the last oscillator connects back to the first to represent the "natural" frequency of the coupled system.
Network topology (a, b) and bifurcation analysis (c) showing the sensitivity of the system near critical points.

Stability and sensitivity were analyzed using bifurcation theory. Specifically, the team looked for Bogdanov-Takens (BT) bifurcations, regions where the oscillation period becomes extremely sensitive to external input changes.

Key Insights: Frequency Mismatch and Sensitivity

The direction of the traveling wave is governed by a "leader-follower" dynamic. In a unidirectional chain, the first oscillator ($i=1$) acts as the driver, and the rest of the chain synchronizes to its frequency.

The wave direction depends on the difference between the driver's intrinsic period ($T_s$) and the period the rest of the network "prefers" when coupled ($T_R$):

  • Direct Wave ($T_s < T_R$): Occurs when the driver is faster than the group.
  • Retrograde Wave ($T_s > T_R$): Occurs when the driver is slower than the group.

By operating near a homoclinic bifurcation associated with the BT point, the period of the leading oscillator changes more drastically than the rest of the network in response to global inputs. This disproportionate sensitivity allows a simple global signal to flip the local wave direction.

Results: Global Control of Wave Direction

The simulations demonstrated that adjusting only the global excitatory input ($S_E$) can successfully reverse wave propagation. For instance, a high $S_E$ might produce a direct wave, while a slight decrease flips it to retrograde.

Space-time plots (a, b) demonstrating wave reversal and the parameter space map (c) defining the direct and retrograde regimes.

The researchers mapped the $S_E - S_I$ parameter space, identifying a clear boundary between direct and retrograde regimes. Crucially, this switching occurs on the timescale of the input change itself, suggesting that neuromodulators like dopamine or serotonin could trigger these reversals in real-time without needing slow synaptic plasticity.

Conclusion and Future Directions

While the current 1D model simplifies the 3D complexity of biological tissue, it provides a fundamental "blueprint" for understanding signal reversal in one-way systems. Beyond biology, these findings could inform the design of Central Pattern Generators (CPGs) for soft robotics, enabling machines to switch between forward and backward locomotion using a single global control parameter rather than complex internal reprogramming.

Summary updated

Abstract

Certain biological systems exhibit both direct and retrograde propagating wave signals, despite unidirectional neural coupling. However, there is no model to explain this. Therefore, the underlying physics of reversing the signal's direction for one-way coupling remains unclear. Here, we resolve this issue using a Wilson-Cowan oscillators network. By analyzing the limit cycle period of various coupling configurations, we determine that intrinsic frequency differences among oscillators control wave directionality.
BibTeX
@misc{elisha2024direct,
  author = {Guy Elisha and Richard Gast and Sourav Halder and Sara A. Solla and Peter J. Kahrilas and John E. Pandolfino and Neelesh A. Patankar},
  title = {{Direct and retrograde signal propagation in unidirectionally coupled Wilson-Cowan oscillators}},
  year = {2024},
  eprint = {2402.18100},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2402.18100},
}
Figure