Single Neurons Detect Spatiotemporal Activity Transitions through STP and EI imbalance

Code for simulations, data analysis, and figure generation.

Copyright (C) 2026 Aditya Asopa and Upinder S. Bhalla, National Centre for Biological Sciences, Tata Institute of Fundamental Research, Bangalore, India.

All code in this project is licensed under GPL 3.0

This paper is submitted to eLife 2026. A preliminary version is on bioRxiv:

"Short-term plasticity of EI balance at single neurons can detect pattern transitions"

Aditya Asopa and Upinder S. Bhalla

bioRxiv 2024.10.30.621034; doi: https://doi.org/10.1101/2024.10.30.621034

Experimental data are available on Zenodo: https://doi.org/10.5281/zenodo.20179470

About

This is the code and data repository for this study, notably including the entire model specification, figure generation, and data analysis code.

Large experimental data files (>5 MB) are available on Zenodo rather than in this repository. Scripts that require them expect the data files in relative subdirectories to the plotting script, and these relative paths may need editing to line up with the data files downloaded from Zenodo. The data filenames are the same.

Dependencies

All the code here uses Python version 3.10.

All the simulation code uses MOOSE version 4.1.0 Jhangri.

The model optimization code also depends on HOSS and FindSim.

Some of the MOOSE graphics depends on the 3-D rendering package vpython.

MOOSE is the Multiscale Object Oriented Simulation environment. It is good for running ODE-based signaling models defined in the standard SBML, as well as multiscale models combining electrical and chemical signaling. The current project depends on its multiscale capabilities.

FindSim is the Framework for Integrating Neuronal Data and Signaling Models.

HOSS provides a set of methods for performing hierarchical optimization of signaling and other models.

Jardesigner is three things: A JSON format for storing multiscale neuronal model definitions; a Python library for building models using this format, and a Javascript app for building and editing jardesigner models in a GUI. The Python library is included in the Zenodo repository with this paper. The working version of jardesigner on GITHUB is a8c8373. All the simulation scripts here use a model defined in the Jardesigner JSON format.

CONTENTS

This repository contains the following subdirectories:

• FIG1_basic_data/ - Loads and plots basic experimental EI dynamics data for Fig 1. Also contains datapaths.py (used by all figure scripts) and the eidynamics Python package.
• FIG2_E_I_component_heatmaps/ - Plots E and I synaptic component heatmaps for Fig 2.
• FIG3_EI_balance_expts/ - Plots EI balance experimental data for Fig 3 and the supplementary panel for Fig 3.
• FIG4_Model_layout/ - Generates dynamic 3-D images of the multiscale model in your browser.
• FIG5_burst_EPSP/ - Simulates burst EPSP responses (jgenFig5.py) and plots Fig 5 (plotFig5.py).
• FIG6_Poisson_stim/ - Simulates Poisson stimulus responses and plots Fig 6, including no-STP and no-NMDA control variants.
• FIG7_Mismatch_Heatmaps/ - Simulates mismatch responses for panels AB, C, and E (jitter), and plots Fig 7 heatmaps.
• FIG8_spiking/ - Simulates spiking mismatch responses including frequency sweep and theta rhythm variants, and plots Fig 8.
• FIG9_param_sensitivity/ - Simulates and plots parameter sensitivity analysis for Fig 9.
• Fig10_summary_schematic/ - Source file (SVG/ODG) for the Fig 10 summary schematic.
• ANIMATION/ - Uses MOOSE 3D graphics to generate a movie of the simulation of mismatch detection.
• MODEL_FITTING/ - Code for fitting the model to extracted voltage-clamp peak and timing data. Also contains code for Supplementary figures 1, 2, and 3.
• SUPP/ - Code for plotting Extended figures 3 to 9. Also contains simulated data output for the Poisson train stimulus for the reference model, model without STP, and model without NMDA.
• SUPP_TABLES_PARAMS/ - Supplementary methods PDF, tabulated chemical-electrical model parameters, and model equations.
• polygonProtocols/ - Photostimulation protocol definition files used in the experiments.

In the directories for FIG5 to FIG9, any file of the form jgen*.py is for performing simulations to generate the output .h5 files. Most of these, with the exception of FIG5, require to be run on a moderate server with 24 or more threads.

DATA LOCATION AND SIZE

The various plotFig*.py scripts are for plotting the figures. They need simulation data and in many cases experimental data. These data files are in .h5 format. The paths to the data will need to be updated in the scripts, as per your system. The required experimental data files are in the zenodo repository. You'll need to run the jgen scripts to get the simulation data files. A few of the smaller .h5 simulation files are provided in this repository. For reference, here are some of the file sizes:

Zenodo repository: ~7 GB

FIG5: ~7 MB FIG6: ~100 MB FIG7: ~143 MB FIG8: ~781 MB FIG9: ~1181 MB