Low-Rank SNNs with Gamma Oscillations

This repository contains the simulation and analysis code for the paper:

"Neural oscillation in low-rank SNNs: bridging network dynamics and cognitive function"
by Bin Li, et al.
Accepted in Frontiers in Computational Neuroscience, 2025. https://doi.org/10.3389/fncom.2025.1598138

🔍 Overview

This project investigates how the phase of gamma oscillations in low-rank spiking neural networks (SNNs) modulates cognitive task performance.
We combine biophysically grounded modeling and dynamical systems analysis to bridge network structure and function.

Key components:

• Biophysically realistic SNNs based on the voltage-dependent theta neuron model
• Macroscopic model bifurcation analysis
• Go-Nogo task simulations under oscillatory and stationary states
• Phase-dependent performance evaluation

---

🧠 Main Features

• Predictable network-level dynamics via bifurcation analysis
• Phase-specific modulation of task output
• Low-rank structure in E→E connections
• ING-type gamma oscillation reproduction
• Strict implementation of Dale’s principle

---

📁 Folder Structure

• bifurcation/voltage_dependent_theta.ode: ode file for bifurcation analysis (Software: XPPaut)
• low-rank SNN: code for low-rank spiking neural networks
  • 'configures/': configuration files
  • 'functions.py': functions for SNN
  • 'lowranksnn.py': code of module for low-rank SNN
  • 'main_pytorch.ipynb': main code for low-rank SNN (using PyTorch)
  • 'phase_sensitivity.ipynb': code for phase sensitivity analysis

---

📄 License

This project is licensed under the MIT License.

---

📬 Contact

For questions, please contact:

📧 li [at] neuron.t.u-tokyo.ac.jp

🔗 Academic Homepage: https://libinutokyo.github.io/