RapheGIF

A Augmented generalized integrate-and-fire (aGIF) models add biophysical realism to GIF models. B aGIFs can be trained to imitate individual neurons. C Resampling a bank of aGIFs yields a population model with realistic heterogeneity. D, E Simulated 5-HT populations encode the derivative of their input.

Code for building experimentally-contstrained spiking neural network models of the dorsal raphe nucleus (DRN) used in Harkin et al., 2023.

Emerson F. Harkin, Michael B. Lynn, Alexandre Payeur, Jean-François Boucher, Léa Caya-Bissonnette, Dominic Cyr, Chloe Stewart, André Longtin, Richard Naud, and Jean-Claude Béïque. Temporal derivative computation in the dorsal raphe network revealed by an experimentally-driven augmented integrate-and-fire modeling framework. eLife, 2023. doi: 10.7554/eLife.72951

Are you an electrophysiologist interested in fitting spiking neuron models with a limited set of Hodgkin-Huxley currents to your data? See "A User's Guide to Generalized Integrate-and-Fire Models" (open access). For more detailed information about how the serotonin neuron model used here was developed, see A Simplified Serotonin Neuron Model. "Patch-clamp recordings from dorsal raphe neurons" used to contrain our models are freely available on the Dryad data repository.

Overview

The project is broken up into grr, a reusable library for fitting GIF neuron models forked from the excellent GIF Fitting Toolbox, analysis scripts for reproducing the models and simulations from our paper, and figs/scripts for reproducing the figures in our paper from the results of analysis.

.
├── analysis                        # Data analysis + simulation scripts
│   ├── GIF_pipeline                # Fit spiking neuron models
│   ├── GIFnet_pipeline             # Run network simulations
│   │   └── input_generators
│   ├── gaba_synapses
│   └── gating                      # Characterize I_A in 5-HT neurons
├── figs
│   └── scripts                     # Notebooks to reproduce figures
└── grr                             # Library for fitting + running models

To use the augmented GIF model in your own work, clone or download this repo and install grr using pip install . && pip install -r requirements.txt from inside the project.

To reproduce the results from our paper, follow these steps:

1. Clone or download this repo and install grr.
2. Get a copy of the raw data and extract it.
3. Set the environment variable DATA_PATH to the root directory of the raw data (recommended: data/raw) and the variable IMG_PATH to the location where you would like figures to be saved (recommended: figs/ims).
4. Run the scripts in analysis. GIF_pipeline must be run before GIFnet_pipeline.
5. Run the notebooks in figs/scripts.

Contributions

Christian Pozzorini wrote the GIF Fitting Toolbox that forms the foundation of grr. All remaining code was written by Emerson Harkin. Alexandre Payeur prototyped some of the models included in grr and provided valuable input on all aspects of the project along with Richard Naud and Michael Lynn.

License

Science thrives on openness. This work is released under the MIT license and is free to use for any purpose. If you find our work useful, please cite our paper!

@article{harkin2023temporal,
  title={Temporal derivative computation in the dorsal raphe network
    revealed by an experimentally-driven augmented integrate-and-fire
    modeling framework},
  author={Harkin, Emerson F and Lynn, Michael B and Payeur, Alexandre
    and Boucher, Jean-Fran{\c{c}}ois and Caya-Bissonnette, L{\'e}a
    and Cyr, Dominic and Stewart, Chloe and Longtin, Andr{\'e}
    and Naud, Richard and B{\'e}{\"\i}que, Jean-Claude},
  journal={eLife},
  volume={12},
  pages={e72951},
  year={2023},
  publisher={eLife Sciences Publications, Ltd},
  langid={english},
  doi={10.7554/eLife.72951},
}