# rgc-natstim

This repository contains the data and models underlying the analyses published in [Hoefling et al. 2024: A chromatic feature detector signals visual context changes] (https://www.biorxiv.org/content/10.1101/2022.11.30.518492v2). The code to run the analyses can be found [here](https://github.com/eulerlab/rgc-natstim-model).

## Repository structure
The repository contains two sub-directories holding the data and models, respectively, which are structured as outlined below.

#### Models
The **models** directory is further sub-divided into a directory hodling the LN and CNN models, respectively, which each hold folders for individual models identified by hashes. 

models/{linear, nonlinear}/{model_hashes}

#### Data
The **data** directory is sub-divided into four directories:
* base: holds the basic preprocessed data as .hf files
* movie: holds the movie stimulus files
* mei: holds intermediate analysis results (Figs. 3-5 in the paper)
* response_gradient: holds response gradient analysis results (Fig. 6 in the paper)

## How to use this data
Clone\* the contents of this repository to your machine using the gin CLI. 
Then follow instructions in the [GitHub repository](https://github.com/eulerlab/rgc-natstim-model) to re-run analyses or train models.

\*You may have to explicitly download the contents of larger files using the ``gin get-content`` command. An ``UnpicklingError`` when trying to open a pickle file is usually resolved by downloading the contents of the file.

## Authors and acknowledgements
See acknowledgments section in [Hoefling et al. 2024: A chromatic feature detector signals visual context changes] (https://www.biorxiv.org/content/10.1101/2022.11.30.518492v2).