/behavioral_dataframe_processing

Primary LanguageJupyter Notebook

Behavioral Dataframe Processing

Overview

Part 1. Reward Training Analysis with MED-Pc

  • This project helps extract and process MED-PC (https://www.med-associates.com/med-pc-v/) recording data. The overall goal is to train mice to associate a tone with a reward(sugary liquid). MED-PC records the times that the tones are played, when the mice enter/exit the port where the reward is dispensed, and the relevant metadata. The experiment is usually run with one subject for a recording that takes about a hour for multiple days. The data processing will calculate various metrics that displays or averages all the trials(each tone going off in a session and it's associated port entry/exit) for each recording session for all the subjects.

Part 2: ELO Score Calculation and Visualization

  • This project helps calculate the ELO score for different social competition assays. The data collected for these assays are Excel sheets that are used during the recording. The assays includes tube test, urine marking, home cage observation, and reward competition. ELO scores keep track of the overall performance for a given subject, which can be compared to other subjects within the same cage to approximate the "ranking". Because the assays have multiple recordings for each subject, the ELO scores will be calculated after each interaction.

Repository Organization

  • ./jupyter_notebooks
    • Directory that has the Jupyter Notebooks to extract/process the MED-PC and the accompanying dataframes/plots that are created from it.
  • ./results
    • Directory that has a record of previous iterations of the Jupyter Notebooks
  • ./src
    • Directory that has the Python source code used in the Jupyter Notebooks. All the original functions used in the notebooks will be imported from this. TODO: Add sections on conda environments and scripts

Steps To Take

Step 0: Learn how to use the Command Line

https://code.visualstudio.com/Download

  1. Before starting, you must pick a program for your Command-line interface.

1.1 When editing and reading code, it is helpful to use a code editor that has features specific to programing. We recommend downloading Visual Studio Code. - Download Link: https://code.visualstudio.com/Download

Step 1. Clone this repository

  1. Check if you have the Git program on your computer. Open up Command-line interface program. We recommend Command Prompt for Windows, and Terminal for Mac.

1.1 In the terminal, type git then enter. - Documentation of Git should be displayed if it is properly installed on your computer. If it says that command is not recognized or that the program doesn't exist, then install it with one of the following instructions: - Windows: https://garnet-rotate-01f.notion.site/Git-Installation-with-Windows-09e4f9f13c9f47c48c8d02a1f7647704 - Mac: https://garnet-rotate-01f.notion.site/Git-Installation-with-Mac-b5e51901e97b4c65a114bb25bc9f2dfa

  1. Once you have Git, download the repository in a folder that you want to save it in. To move to the desired folder, type the following command into your Command-line interface program: cd {/path/to/folder}

2.1 Download the Github repository with the command: git clone https://github.com/padillacoreanolab/behavioral_dataframe_processing - NOTE: Every time you use this repo after cloning it, check for updates with: - cd {./path/to}/behavioral_dataframe_processing - git pull origin main

Step 2. Create Conda Environment to Install Necessary Python libraries

  1. Check if you have Anaconda:

  2. Create the Conda environment by following the instructions in: ./conda_environment/environment_install.sh

    • NOTE: This step only needs to be done once. Every subsquent time you want to run the analysis, you'll turn on the Conda environment with:
      • conda deactivate
      • conda activate {./path/to}/behavioral_processing_env

Step 2. Turn on Jupyter Notebooks Program

  1. Type the command jupyter notebook --allow-root in your terminal, and then press inter

1.1 If a browser window doesn't pop up with Jupyter notebooks, open up a browser window. Then copy and paste the URL given in the command line to the address bar. It will be followed by To access the notebook, open this file in a browser: and Or copy and paste one of these URLs: - It should have a format similar to: http://localhost:8888/?token=gsOH0vx373NRzHUmDzwX9TFzBf5Wx6czWIr1irV7ERKBLmlK

  1. Click on ./jupyter_notebooks in the Jupyter Notebooks GUI (Should be in the web browser)

Step 3. Run each Data Processing/Analysis Jupyter Notebooks

  1. Follow the instructions to run the Jupyter Notebooks at ./jupyter_notebooks/README.MD
    • NOTE: The dataframes and the plots should be saved in subdirectories in: (./jupyter_notebooks/proc)[./jupyter_notebooks/proc]

Resources

MED-PC Trans

Python library to extract data from MED-PC Recording Files

ELO Score