SBI-trained posterior models for cortical circuit parameter inference from EEG features

This repository contains the trained Simulation-Based Inference (SBI) models used in the study:

“A Framework Integrating Spiking Cortical Circuit Modeling and Simulation-Based Inference to Probe Biomarkers of Cortical Dysfunction in Alzheimer’s Disease”.

The uploaded files include trained posterior estimators (model.pkl), inference objects (inference.pkl), and feature scalers (scaler.pkl) corresponding to multiple SBI models trained during the development and evaluation of the study.

These models were trained using simulation-based inference methods implemented with the sbi Python library and were designed to infer latent neurophysiological parameters of cortical circuit models from EEG/CDM-derived feature representations.

The repository includes models trained with different feature configurations, mechanistic models, and experimental settings explored throughout the study.

The associated publication is available at:

https://link.springer.com/article/10.1007/s12539-026-00817-8