This is a repo to make inference on a mean-filed model of spiking QIF neurons, using Optimization (DE/PSO), Simulation-based Inference (SBI), Hamiltonian Monte Carlo (HMC), and Neural ODEs.
For fast simulation (JIT/Jax), see the folder Simulation.
For Linear stability analysis, see the folder LSA.
Based on our benchmark, we recommend using SBI for efficient inference on generative parameters. Please refer to the folder SBI for a demonstration.
For exact (but computionally expensive) inference, see the folder HMC.
For fast point estimation, see the folder Optimization.
To infer the dynamics from simulations, particularly to learn vector fields and nullclines (travesering across scales), see the folder Neural ODEs.
citation:
@article{Baldy2024InfMFM,
title={Inference on the Macroscopic Dynamics of Spiking Neurons},
author={Baldy, Nina and Breyton, Martin and Woodman, Marmaduke M and Jirsa, Viktor K and Hashemi, Meysam},
journal={Neural Computation},
volume = {36},
number = {10},
pages = {2030-2072},
year = {2024},
issn = {0899-7667},
doi = {10.1162/neco_a_01701},
url = {https://doi.org/10.1162/neco\_a\_01701},
publisher={MIT Press 255 Main Street, 9th Floor, Cambridge, Massachusetts 02142, USA}
eprint = {https://direct.mit.edu/neco/article-pdf/doi/10.1162/neco\_a\_01701/2469857/neco\_a\_01701.pdf},
}
}
This research has received funding from EU’s Horizon 2020 Framework Programme for Research and Innovation under the Specific Grant Agreements No. 101147319 (EBRAINS 2.0 Project), No. 101137289 (Virtual Brain Twin Project), and government grant managed by the Agence Nationale de la Recherch reference ANR-22-PESN-0012 (France 2030 program).