A team of researchers from the University of Pavia has introduced a tool that automatically converts detailed spiking neural network models into simpler “mean field model” (MFM) equations – mathematical descriptions that capture the dynamics of populations of neurons.
The approach, called Auto-MFM, translates microscopic properties – such as single-neuron firing behaviour and synaptic connectivity into mesoscale models while preserving key functional features.
The framework connects several tools in the EBRAINS ecosystem: detailed neural circuits can be reconstructed with Brain Scaffold Builder and simulated with NEST, after which Auto-MFM automatically derives a simplified population-level model that can be integrated into The Virtual Brain.
Tested on a cerebellar circuit, the generated mean field model closely matched the activity of the spiking network’s populations, reproducing both average and time-varying firing patterns across different simulation conditions. The framework also allowed researchers to explore disease-related changes, including connectivity alterations linked to ataxia and variation in excitatory synaptic strength.
The authors suggest that Auto-MFM could provide a flexible way to build efficient, brain-region-specific models for studying both healthy and pathological brain dynamics. By linking cellular and synaptic changes to activity patterns observed at the scale of entire brain regions, the approach could help researchers investigate how local circuit abnormalities contribute to the large-scale signals associated with neurological disorders.
Read more
Automated derivation of mean field models from spiking neural networks for the simulation of brain dynamics
Roberta M. Lorenzi, Marialaura De Grazia, Claudia A.M. Gandini Wheeler Kingshott, Fulvia Palesi, Egidio D’Angelo, Claudia Casellato
bioRxiv 2026.03.18.712631; doi: https://doi.org/10.64898/2026.03.18.712631
Author: Helen Mendes
Contact: press@ebrains.eu
Create an account
EBRAINS is open and free. Sign up now for complete access to our tools and services.