NC-MCM-Visualizer

See on GitHub

A toolbox to visualize neuronal imaging data and apply the NC-MCM framework to it

This is a toolbox uses neuronal & behavioral data and visualizes it. NCMCM stands for Neuro-Cognitive Multilevel-Causel-Modeling, but also for:

• Numerous options to visualize neuronal data and create diagnostic plots
• Creating neural manifolds using sklearn dimensionality reduction algorithms or BunDLeNet
• Make interactive behavioral state diagrams using pyvis
• Cluster behavioral probability trajectories and test them for non-markovianity
• Movies of behavioral/neuronal trajectories saved as .gif-files

Getting Started (for end-users)

1. Installation: Open a terminal window in your Python project directory and run

   pip install ncmcm
   

2. Importing the package: In your Python script or notebook, import the package

   import ncmcm as nc
   

3. Usage: Now you can already start with creating a Database instance with your data, to start your analysis:

   db = nc.Database(neuron_traces=X, behavior=B, neuron_names=x_names, states=b_names, fps=fps)
   

   Or you can use one of the 5 the included C.elegans dataset to explore the package:

   loaded_data = Loader(data_set_no=1)
   nc.Database(*loaded_data.data)
   

4. Structure: the framework is divided into three main parts which can be imported separately:

   • Cognitive Graphs and 3D Visualizations: This includes the classes necessary for creating cognitive graphs and 3D visualizations. These are available under the module: ncmcm.ncmcm_classes
   • Helper Functions: These are functions that are used by some of the classes and can also be accessed independently. You can find them under: ncmcm.helpers
   • BunDLe-Net: This module is used to create Markovian 3D embeddings and is accessible via: ncmcm.bundlenet

5. Tutorial: Check out the demo-notebooks on the GitHub (here). They serve as a useful starting point to explore the functionalities of ncmcm.

Installation and usage information (for contributors)

If you're interested in contributing to this project or creating your own versions based on the existing code, follow these steps:

Installation

1. Clone the Repository: Clone this repository to your local machine using Git:

   git clone https://github.com/DriftKing1998/NC-MCM-Visualizer.git
   

2. Install Dependencies: Navigate to the project directory and install the required dependencies using pip:

   cd <project_directory>
   pip install -r requirements.txt
   

New Branches

1. Explore the Code:
   ncmcm.ncmcm_classes module contains the classes used by ncmcm:

   1. Loader can be used to load data from a certain MATLAB-file.
   2. Database is a container for data, which can be used to generate the behavioral probability maps by adding a sklearn-model. It also allows to create different plots and diagnostics.
   3. Visualizer is created by adding a mapping or a BundDLeNet (=default). It allows to create 3D plots and movies.
   4. CustomEnsembleModel is a model that creates an ensemble of models, specializing each model to detect a label.

   
   ncmcm.helpers module contains some auxiliary functions:

   1. Functions to prepare data
   2. Functions to test for stationarity & non-markovianity
   3. Plotting functions for the tests
   4. And many more...

   
   ncmcm.BundDLeNet module contains all the parts of BundDLeNet:

   1. Class for creating/training the model
   2. Functions to prep data and get loss

2. Make Changes: Make your desired changes or enhancements to the codebase. Feel free to add new features, fix bugs, or improve documentation.

3. Testing: Ensure that your changes are tested thoroughly. You can run existing tests or write new ones to validate your modifications.

4. Create a Branch: Create a new branch for your changes:

   git checkout -b <new_branch>
   

5. Commit Your Changes: Once you're satisfied with your changes, commit them to your branch:

   git add .
   git commit -m "description of changes"
   

6. Push Changes: Push your changes to your forked repository:

   git push origin <new_branch>
   

Add to the Codebase

Since this project is a first step, any additions are more than welcome.

1. Pull request: Go to the repository and open a pull request from your branch. Provide a clear description of your changes and why they're beneficial.

2. Review: Await feedback from me and address any requested changes. Once approved, your changes will be merged into the main branch.

This project was created as part of the masters project of Hofer Michael