Pytunnel package

Introduction

This repository contains scripts to run virtual corridors using Python 3. It is a sequential navigation task, where mice have to navigate through virtual corridor to reach landmarks in a specific order. For more details, check this Notion page

There is a separate repository for data analysis : 1DSequenceAnalysis

2023/05/17 This repository is for use in behavioral box. Cloned from Shohei Furutachi's project.

Installation

To retrieve the code, clone the repository using git:

git clone [email protected]:MasahiroNakano/1DSequenceTaskPy.git

To install dependencies, I recommend that you first create a virtual environment (with python or conda) and install dependencies inside using pip.

Using conda (from Anaconda Prompt), recommended for Windows users:

conda create -n py35 python=3.5
conda activate py35
pip install -r requirements.txt

Using python3 (from a terminal), recommended for Linux users: (I=masahiro haven't tested this)

python3.5 -m venv py35
source pytunnel_venv/bin/activate
pip install -r requirements.txt

Also, download texture files from Google drive folder and place them under examples/textures

Usage

To start a virtual corridor you need to:

• open a terminal
• activate the virtual environment
• run a tunnel script with the corresponding tunnel configuration file.

For example, to run one the example file for the basic tunnel:

conda activate py35
cd /to/this/folder
python src/pytunnel/main.py examples/yaml/desktop/desktop_no_daq.ymal'

All the configuration file (.yaml) can be found in examples/yaml. Each folder corresponds to the training protocols described in this Notion page

When testing codes, the most important thing is to specify nidaq_stub instead of nidaq in flip_tunnel/io_module, if your test computer is not connected to the experimental rigs.

flip_tunnel:
    io_module: nidaq_stub

You can give a manual reward by space bar if you set the following variable True.

flip_tunnel:
    manual_reward_with_space: True

In the desktop mode, pressing the space bar would mimic the lick. See if your codes as you expect.

To know more about the options of a tunnel script, use the --help/-h option, e.g.

python src/pytunnel/flip_tunnel.py --help

Task details that are relevant for the code

For more details, check this Notion page

Rules

sequence: Animal needs to visit one goal/landmark and lick to get a reward. Assist-reward may be used. run-auto: reward is given when the mouse has ran random length in the corridor. run-lick: reward is given when the mouse has licked after running more than a random length in the corridor.