specparam-sbi

Tutorials on chaining together LFP simulation (neurodsp), spectral parameterization (fooof), and simulation-based inference (sbi).

Colab Link	Tutorial
	Demo 1: minimal example of how to use fooof and simulating time series.
	Demo 2: reproducing the Poisson EI-1/f exponent model using neurodsp.
	Demo 3: minimal example of how to use sbi.
	Demo 4: combining everything to do inference and probabilistic spectral parameter estimation.

Packages required (all installable via pip install):

• neurodsp (repo)
• fooof (repo)
• sbi (repo)

To set up and run in a new conda environment:

conda create -n sbsp python=3.8
conda activate sbsp
pip install uv
uv pip install jupyter fooof neurodsp sbi==0.22

For more complex mechanistic model simulations with spiking neural networks, go to AutoMIND.